Spectrum-Based Pushover Analysis For Predicting Seismic Demands of Tall Buildings

Spectrum-based Pushover Analysis for Predicting
Seismic Demands of Tall Buildings
by
Yang Liu

A Thesis Submitted to
The Hong Kong University of Science and Technology
in Partial Fulfilment of the Requirements for
the Degree of Doctor of Philosophy
in Civil Engineering

July 2018, Hong Kong
HKUST Library
Reproduction is prohibited without the author’s prior written consent
Acknowledgements
My deepest gratitude and appreciation goes to my supervisor Prof. J. S. Kuang, for his
professional, enthusiastic but flexible guidance throughout my research and study. He is not
just an academic advisor. He is also a mentor. His inspiring recommendations and warm-
hearted encouragement helped me to overcome difficulties when facing problems or being
frustrated throughout my life in HKUST.
I wish to thank the instructors of courses I have taken during my graduate study, namely,
Profs J. S. Kuang, Zongjin Li, Christopher Leung, Xiaoping Wang, Gang Wang, and
Qingping Sun. I have learnt a lot from their course, which also helped a lot in my research.
Special appreciation is given to Prof. H. Sakata from Tokyo Institute of Technology, for his
kind support and host when I visited Tokyo Tech as an exchange Ph.D. student.
Sincere thanks are expressed to the examiners of my Ph.D. Qualify Exam, Prof. Lambros
Katafygiotis and Prof. Ilias Dimitrakopoulos and the Thesis Advisory Committee members,
Prof. Gang Wang, Prof. C. M. Chan and Prof. Xinzheng Lu, for their invaluable advice
throughout my research and thesis writing.
I also wish to express thanks to members of staff in our department, including Mandy Lam,
Anny Tsoi, Jack Iu, Richard Chan, Ka Chi Chan, Cheryl Tang, Queenie Tso and the others
not listed here, for their kind help during my graduate life at HKUST, especially when I was
teaching assistant coordinator of the Civil department.
My sincere thanks are expressed to my girlfriend Qingqing Gao, for her love, accompany
and encouragement in the wind and rain. I also wish to give my sincere appreciation to my
friends and colleagues in HKUST, including Dr. Nan Zhang, Roy Yuchen Huang, Daisy
Xueying Wang, Srinivas Mogili, Michele De Filippo, Pheobe Chan, Dr. Qing Zeng, Dr.
Hongyu Shao, Dr. Zhaofeng Li, Jing Cong, and other friends who are not listed here, for
their help and care on my study and life during the last three years. Finally, I would like to
express my sincere appreciation to my parents. Their love and support are the most
significant source of my motivation.
IV
Contents

Cover Pages ............................................................................................................................... I
Authorisation............................................................................................................................. II
Signature Page ......................................................................................................................... III
Acknowledgements ..................................................................................................................IV
Contents .................................................................................................................................... V
List of Figures ....................................................................................................................... VIII
List of Tables ......................................................................................................................... XII
Abstract ................................................................................................................................. XIII
Chapter 1 Introduction ............................................................................................................... 1
1.1 Background ...................................................................................................................... 2

1.2 Objectives and scopes of thesis........................................................................................ 6
Chapter 2 Literature Review ...................................................................................................... 7
2.1 Introduction ...................................................................................................................... 8

2.2 Theory of structure dynamics and earthquake engineering for two dimensional
problems ................................................................................................................................. 9
2.3 Current provisions in codes of practice ......................................................................... 10
2.3.1 Nonlinear static analysis method ............................................................................ 11

2.3.2 Nonlinear response time history analysis method .................................................. 20
2.4 Current advanced pushover procedures ......................................................................... 20

2.5 Summary ........................................................................................................................ 30
Chapter 3 Spectrum-based Pushover Analysis for Estimating Seismic Deformation of Frame

and Wall Structures .................................................................................................................. 31
3.1 Introduction .................................................................................................................... 32

3.2 Spectrum-based pushover analysis method ................................................................... 33
3.3 Case study ...................................................................................................................... 39
3.3.1 Details of ground motions ....................................................................................... 39

3.3.2 Prototype of structures ............................................................................................ 43
3.3.3 Analytical models ................................................................................................... 46
V
3.4 Results of the nonlinear time history analysis method .................................................. 49
3.5 Results of the nonlinear time history analysis method and pushover methods ............. 56
3.5.1 Frame structures ...................................................................................................... 56

3.5.2 Wall structures ........................................................................................................ 62
3.6 Concluding remarks ....................................................................................................... 65
Chapter 4 Modified Spectrum-based Pushover Analysis for Estimating Seismic Deformation

of Dual Wall-frame Structures ................................................................................................. 67
4.1 Introduction .................................................................................................................... 68

4.2 Modified spectrum-based pushover analysis method .................................................... 69
4.3 Case study ...................................................................................................................... 75
4.3.1 Ground motions selection ....................................................................................... 75

4.3.2 Prototype of structures ............................................................................................ 75
4.3.3 Analytical models ................................................................................................... 76
4.4 Results of the nonlinear time history analysis method .................................................. 79
4.5 Results of the nonlinear time history analysis method and pushover methods ............. 88
4.6 Concluding remarks ....................................................................................................... 98
Chapter 5 Spectrum-based Pushover Analysis for Estimating Seismic Deformation of Frame

Structures with Singly Symmetric Plan ................................................................................... 99
5.1 Introduction .................................................................................................................. 100

5.2 Spectrum-based pushover analysis method for buildings with singly symmetric
building plan ...................................................................................................................... 101
5.3 Case study .................................................................................................................... 105
5.3.1 Ground motions selection ..................................................................................... 105

5.3.2 Prototype of structures .......................................................................................... 105
5.3.3 Analytical models ................................................................................................. 107
5.4 Results of the nonlinear time history analysis method ................................................ 108
5.5 Results of the nonlinear time history analysis method and pushover methods ........... 133
5.6 Concluding remarks ..................................................................................................... 160
Chapter 6 Extended Spectrum-based Pushover Analysis for Estimating Seismic Forces of

Frame and Wall Structures..................................................................................................... 161
6.1 Introduction .................................................................................................................. 162

VI
6.2 Extended spectrum-based pushover analysis method .................................................. 163
6.3 Case study .................................................................................................................... 167
6.4 Result of nonlinear response time history analysis ...................................................... 167
6.5 Results of the nonlinear time history analysis method and pushover methods ........... 177
6.5.1 Comparison of results from ESPA and NLRHA .................................................. 177

6.5.2 Comparison of results from pushover analysis methods and NLRHA ................. 179
6.6 Concluding remarks ..................................................................................................... 190
Chapter 7 Conclusion ............................................................................................................. 191
7.1 Conclusions .................................................................................................................. 192

7.2 Future research ............................................................................................................. 195
References .............................................................................................................................. 196
VII
List of Figures
Figure 1.1 Deformation shape of three major types of structures (Kuang 2016). ..................... 5
Figure 2.1 Multiple capacity curves required modelling strength degradation. ...................... 12
Figure 2.2 Capacity curve with global strength degradation modelled. .................................. 13
Figure 2.3 Intersection point of Demand and Capacity Spectrum. .......................................... 14
Figure 2.4 Schematic illustrating the FEMA 356 process to estimate the target displacement
for a given response spectrum and effective period................................................................. 15
Figure 2.5 Determination of the idealized elasto-perfectly plastic force-displacement
relationship. .............................................................................................................................. 17
Figure 2.6 Determination of the target displacement for the equivalent SDOF System. ........ 19
Figure 3.1 Roof displacements of the first three modes of a building reached the peak at the
same time. ................................................................................................................................ 35
Figure 3.2 Flowchart of SPA procedure. ................................................................................. 39
Figure 3.3 Comparison of the acceleration spectra of the design spectra. ............................... 42
Figure 3.4 Design spectrum and the response spectrum of the corresponding scaled ground
motion: (a) motion Set 1; (b) motion Set 2; (c) motion Set 3. ................................................. 43
Figure 3.5 Details of the frame structures: (a) 9-storey frame; (b) 20-storey frame. .............. 45
Figure 3.6 Force-deformation relationship for hinges. ............................................................ 45
Figure 3.7 Elevation and plan views of W1. ............................................................................ 47
Figure 3.8 Elevation and plan views of W2. ............................................................................ 48
Figure 3.9 Inter-storey drift ratio of the frame structures from the nonlinear time history
analysis method: (a) F1; (b) F2. ............................................................................................... 51
Figure 3.10 Hinge plastic rotation of the frame structures from the nonlinear time history
analysis method: (a) F1; (b) F2. ............................................................................................... 53
Figure 3.11 Inter-storey drift ratio of the wall structures from the nonlinear time history
analysis method: (a) W1; (b) W2. ............................................................................................ 55
Figure 3.12 Comparison of Inter-storey drift ratio results of the frame structures: (a) F1; (b)
F2. ............................................................................................................................................ 59
Figure 3.13 Comparison of the hinge plastic rotation results of the frame structures: (a) F1;
(b) F2........................................................................................................................................ 61
Figure 3.14 Comparison of the inter-storey drift ratio results of the wall structures: (a) W1;
(b) W2. ..................................................................................................................................... 64
Figure 4.1 Modal shape distributed forces used in SPA method. ............................................ 69
VIII
Figure 4.2 Distribution of plastic hinges initiated under the first-mode shape forces. ............ 71
Figure 4.3 Flowchart of MSPA procedure. .............................................................................. 75
Figure 4.4 Plan view of the dual wall-frame structure............................................................. 77
Figure 4.5 Elevation of the 2-D dual wall-frame structure. ..................................................... 77
Figure 4.6 Inter-storey drift ratio of the dual wall frame structures from the nonlinear time
history analysis method: (a) WF1; (b) WF2; (c) WF3; (d) WF4. ............................................ 83
Figure 4.7 Hinge plastic rotation of the dual wall frame from the nonlinear time history
analysis method: (a) WF1; (b) WF2; (c) WF3; (d) WF4. ........................................................ 87
Figure 4.8 Comparison of the inter-storey drift ratio results of the dual wall frame structures:
(a) WF1; (b) WF2; (c) WF 3; (d) WF 4. .................................................................................. 93
Figure 4.9 Comparison of the hinge plastic rotation results of the dual wall frame structures:
(a) WF1; (b) WF2; (c) WF 3; (d) WF 4. .................................................................................. 97
Figure 5.1 Structural information of the frame structures with an asymmetric plan: (a) 9-
storey frame; (b) 20-storey frame .......................................................................................... 106
Figure 5.2 Plan arrangements of frame structures. ................................................................ 107
Figure 5.3 Inter-storey drift ratio of LS of the 9-storey frame structures with a singly
symmetric plan from the nonlinear time history analysis method: (a) F1-10; (b) F1-20; (c) F1-
30............................................................................................................................................ 111
Figure 5.4 Inter-storey drift ratio of LS of the 20-storey frame structures with a singly
30............................................................................................................................................ 114
Figure 5.5 Inter-storey drift ratio of HS of the 9-storey frame structures with a singly
30............................................................................................................................................ 117
Figure 5.6 Inter-storey drift ratio of HS of the 20-storey frame structures with a singly
30............................................................................................................................................ 120
Figure 5.7 Hinge plastic rotation of LS of the 9-storey frame structures with a singly
30............................................................................................................................................ 123
Figure 5.8 Hinge plastic rotation of LS of the 20-storey frame structures with a singly
30............................................................................................................................................ 126
IX
Figure 5.9 Hinge plastic rotation of HS of the 9-storey frame structures with a singly
30............................................................................................................................................ 129
Figure 5.10 Hinge plastic rotation of HS of the 20-storey frame structures with a singly
30............................................................................................................................................ 132
Figure 5.11 Comparison of inter-storey drift ratio results of LS of frame structures with
singly symmetric building plan: (a) F1-10; (b) F1-20; (c) F1-30. ......................................... 138
Figure 5.12 Comparison of inter-storey drift ratio results of LS of frame structures with
Figure 5.13 Comparison of inter-storey drift ratio results of HS of frame structures with
Figure 5.14 Comparison of inter-storey drift ratio results of HS of frame structures with
Figure 5.15 Comparison of hinge plastic rotation results of LS of frame structures with singly
symmetric building plan: (a) F1-10; (b) F1-20; (c) F1-30. .................................................... 150
Figure 5.17 Comparison of hinge plastic rotation results of HS of frame structures with singly
Figure 6.1 Base shear-roof displacement relations of two structures. ................................... 164
Figure 6.2 Flowchart of ESPA procedure. ............................................................................. 166
Figure 6.3 Shear forces of the frame structures from nonlinear time history analysis method:
(a) F1; (b) F2. ......................................................................................................................... 170
Figure 6.4 Bending moment of the frame structures from nonlinear time history analysis
method: (a) F1; (b) F2. ........................................................................................................... 172
Figure 6.5 Shear forces of the wall structures from nonlinear time history analysis method:
(a) W1; (b) W2. ...................................................................................................................... 174
Figure 6.6 Bending moment of the wall structures from nonlinear time history analysis
method: (a) W1; (b) W2. ........................................................................................................ 176
Figure 6.7 Comparison of earthquake-induced forces of ESPA and NLRHA for motion Set 2.
................................................................................................................................................ 179
X
Figure 6.8 Comparison of shear force results of the frame structures: (a) F1; (b) F2. .......... 182
Figure 6.9 Comparison of bending moment results of the frame structures: (a) F1; (b) F2. . 184
Figure 6.10 Comparison of shear force results of the wall structures: (a) W1; (b) W2......... 187
Figure 6.11 Comparison of bending moment results of the wall structures: (a) W1; (b) W2.
................................................................................................................................................ 189
XI
List of Tables
Table 3.1 Summary of the key parameters of the target spectrum .......................................... 40
Table 3.2 Details of ground motion Set 1 that matches with the target design spectrum No. 1
.................................................................................................................................................. 40
.................................................................................................................................................. 41
.................................................................................................................................................. 42
Table 3.5 Modal properties of the frame structures ................................................................. 44
Table 3.6 Modal properties of the wall structures ................................................................... 46
Table 4.1 Dimension of the main structural element for wall-frame structures ...................... 78
Table 4.2 Reinforcement details of the wall of wall frame structures. .................................... 78
Table 4.3 Design load for wall-frame structures. .................................................................... 78
Table 4.4 Modal property of dual structures and fraction of base shear taken by wall and
frame ........................................................................................................................................ 79
Table 5.1 Dynamic properties of frames used in the analysis ............................................... 107
XII
Abstract
Quick and accurate prediction of seismic demand of tall buildings is one of the toughest tasks
for engineers in civil industry. Nonlinear response time history analysis method (NLRHA) is
currently considered as the most accurate and rigorous method, but the computation
consumption is exceptionally high. The conventional pushover procedures, in which the
building is treated as a single degree of freedom system, are not capable of reasonably
estimating seismic demands of mid- to high rise-buildings, due to an over-simplification of the
multi-mode effect and the complexity of seismic performance of buildings.
In this study, an efficient, yet accurate, method of analysis, named spectrum-based pushover
analysis (SPA) method, is proposed for the quick estimate of seismic demands of tall buildings.
In the proposed SPA procedure, the very complex and complicated dynamic coupling effect of
different modes of the nonlinear seismic performance of the building is simplified and the
consecutive pushover technique is adopted to consider the simplified coupling effect. The SPA
procedure is then modified to include the impact of the wall-frame interaction in dual wall
frame structures for the fast prediction of the earthquake-induced deformations.
The SPA procedure is further applied for the quick estimate of the tall buildings with singly
symmetric building plans, by including the torsional effect. Finally, the SPA is extended to
predict the earthquake-induced forces of the structures. Comparison of the seismic demands
of tall buildings with different structural types from the NLRHA method, the modal pushover
analysis (MPA) method, the consecutive modal pushover analysis (CMP) method and the
proposed SPA procedures are conducted. It can be seen from the comparison of the results that
the SPA method is capable of predicting the seismic demand of buildings, such as inter-storey
drift ratio, hinge plastic rotation, shear force and bending moment accurately and the
predictions are very close to those obtained from NLRHA.
The SPA procedures are efficient and effective in estimating the seismic demands of building
structures and the computation procedure of SPA procedures is simple. Additonally, the roof
displacement calculation procedure is spectrum-based, which is user-friendly to engineers.
Owing the impressive advantages of the SPA method, it is believed that the proposed SPA
XIII
procedure can be one of the best alternatives of the NLRHA method in the fast estimate of
seismic demands of tall buildings.
XIV
Chapter 1
Introduction
1.1 Background
In the recent decades, the strong earthquakes, such as earthquakes in Taiwan (1997),
earthquakes in Wenchuan (2008), earthquakes in Nepal (2015) and more recently earthquakes
in Iran-Iraq border (2017), have resulted in catastrophe in those regions, with thousands of
peoples died or injured, and lots of building collapsed. According to the records of the United
States Geological Survey, just from the year 2010 to April 2018, there are 138 earthquakes
with magnitude larger than or equal to 7 all around the world. The frequent occurred strong
earthquakes and the disastrous consequences caused by these strong earthquakes have raised
people’s attention on the safety of buildings under the action of earthquakes. For regions with
high seismicities, such as California, New Zealand, and Japan, the buildings were designed
with the consideration of earthquakes. Unfortunately, in many countries and regions of low or
moderate seismicity, buildings were designed with non-seismic design philosophy. The use of
non-seismic design methods has resulted in the little overall seismic resilience of structures and
buildings, particularly, tall buildings. But, low or moderate seismicity does not indicate that
the chance for an earthquake is nil. Consequently, assessing those non-seismically designed
buildings that may have high seismic vulnerability is of critical importance for setting priority
criteria for the possible strengthening and retrofitting of those buildings. Hence, quick
prediction of the seismic demand of existing tall buildings that were designed and built without
seismic consideration is one of the most critical and tough challenges for the engineers and
researchers.
The current methods of analysis in seismic vulnerability evaluation can briefly be classified
into three groups according to their level of complexity, including level one ‘walk down’
evaluation (FEMA 2002), level two preliminary assessment methodologies (Yakut 2004) and
level three linear and nonlinear analyses of the buildings (ATC 1996; CEN 2004; Chinese
Standard 2008). Although level one, level two and the linear analysis are conceptually simple
and easy to conduct, they are mainly based on engineering judgement and linear elastic analysis
of buildings. However, except for some special high-performance buildings or buildings with
the particular requirement, such as hospitals, schools and power plants, it is not economical to
design and build a structure to remain fully elastic for ground motions in the level of the
maximum considered hazard (MCH) level. In other words, under the MCH level of
earthquakes, most buildings will yield and have nonlinear structural behaviour. Therefore,
2
these engineering judgement and linear elastic analysis-based evaluation procedures are not
suitable for analysing the complex structural nonlinear behaviour of buildings for the MCH
level of ground motions.
By careful selection and scaling of earthquake motion records and rigorous modelling
assumptions, the nonlinear response time history analysis (NLRHA) method is the most
accurate and reliable method to compute the seismic demand of buildings, and the NLRHA is
recommended to be used in most of the codes of practice for seismic design and analysis of
structures (ASCE 2013; CEN 2004; Chinese Standard 2008; IBC 2006; NZS 2004). However,
since the NLRHA needs to solve the coupled second order partial differential equation for
every time interval, the NLRHA method has extremely heavy computational demand and is
extremely time consuming, especially when the structure has a large number of degrees of
freedom and when the motion has a large volume of data. Therefore, the NLRHA method is
inapplicable for quick checking and evaluation of seismic performance of massive amount of
tall buildings. Alternatively, nonlinear static method, or pushover analysis method, where a
set of equivalent lateral forces are applied to the structures to obtain the strength and ductility
demand of the structure, is another effective and efficient method to evaluate the seismic
performance of buildings. Owing to its simplicity and reasonable accuracy, the pushover
analysis method has gained enormous popularity and is a standard tool for seismic assessment
in many codes of practice (ASCE 2013; CEN 2004; Chinese Standard 2008; IBC 2006; NZS
2004). Nevertheless, for tall buildings, where the high-mode effect is significant, the
conventional pushover method failed to estimate the seismic demand of tall buildings at a
reasonable accuracy range (Moghaddam and Hajirasouliha 2006; Kalkan and Kunnath 2007;
Huang and Kuang 2010).
To accomplish a more accurate estimation of the seismic performance of tall buildings, while
keeping the efficiency of the pushover method, a lot of multi-mode pushover analysis methods
were proposed. These methods include extend N2 methods (Kreslin and Fajfar 2011; Kreslin
and Fajfar 2012), envelope-based pushover analysis method (Brozovič and Dolšek 2014),
upper-bound pushover analysis (UBP) methods, generalised pushover analysis methods (GPA)
(Jan, Liu, and Kao 2004; Poursha and Samarin 2015), pushover analysis methods based on
energy balance concept (Leelataviwat, Saewon, and Goel 2009; Jiang, Li, and Yang 2010),
modal pushover analysis (MPA) methods (Chopra and Goel 2002; Chintanapakdee and Chopra
2003; Chopra and Goel 2004; Goel and Chopra 2005; Reyes and Chopra 2011; Chopra, Goel,
3
and Chintanapakdee 2004), consecutive modal pushover analysis (CMP) methods (Poursha,
Khoshnoudian, and Moghadam 2009; Poursha, Khoshnoudian, and Moghadam 2011;
Khoshnoudian and Kashani 2012; Khoshnoudian and Kiani 2012; Poursha, Khoshnoudian, and
Moghadam 2014; Zhang, Jiang, and Li 2017), adaptive modal pushover analysis (AMPA)
methods (Gupta and Kunnath 2000; Antoniou, Rovithakis, and Pinho 2002; Antoniou and
Pinho 2004b; Antoniou and Pinho 2004a; Papanikolaou and Elnashai 2005; Kalkan and
Kunnath 2006; Casarotti and Pinho 2007; Ferracuti et al. 2009; Shakeri, Shayanfar, and
Kabeyasawa 2010; Shakeri, Tarbali, and Mohebbi 2012; Abbasnia, Davoudi, and Maddah
2013; Tarbali and Shakeri 2014; Bhatt and Bento 2014) and other methods. Although there
are several advanced pushover analysis methods available to predict the seismic response of
structures and most of them can provide a reasonably good prediction of the seismic demand
of structures, there are some critical deficients of current procedures, which limit the
application of these methods. The limitations are:
1. Coupling of modes in the nonlinear stages

Many multi-mode pushover analysis methods, like extended N2 method, MPA methods,
consider the seismic performance of different modes in these methods is considered separately,
and neglect the coupling effect of different modes in the nonlinear stage. However, in the
nonlinear behaviour of the structure, the coupling effect of different modes is very significant
and will affect the deformation of the structure. The CMP methods, which do not consider the
high mode effect separately, do not deal with the coupling of modes properly. Neglecting or
improperly dealing with the coupling effect would induce significant errors.
2. Computational consumption
For the pushover analysis methods with adaptive computation procedure, to achieve high
accuracy, the modal properties and seismic inputs should be updated as frequently as possible,
making the computation consumption of the AMPA even comparable to the NLRHA method.
This means that the AMPA methods are much less efficient than other pushover methods.
3. Type of structures
In most of the pushover analysis methods, only frame structures with different heights are
studied. However, wall structure and wall-frame structure are other two types that are most
popular for the median or high-rise buildings (Kuang 2014). However, the frame, wall, and
wall-frame structures behave very differently from each other under the uniformly distributed
4
lateral load, which is shown in Figure 1.1. The overall deflected shapes of a shear wall structure
and a rigid frame structure have a flexural configuration and a shear configuration, respectively,
subjected separately to horizontal loading. While for dual wall-frame structures, when
subjected to seismic actions, it has a flexural profile in the lower part and a shear profile in the
upper part because of the wall-frame interaction. As a result, the seismic performance of the
wall and wall-frame structure are also quite different from that of frame structures. Currently,
there are insufficient studies on the pushover analysis method for estimating the seismic
responses of wall and wall frame structures.
Wall Frame Wall-frame

Figure 1.1 Deformation shape of three major types of structures (Kuang 2014).
4. Generic forces in the buildings

In the Performance-Based Earthquake Engineering (PBEE) philosophy, the deformation of the
structure is the primary concern for the seismic evaluation of buildings (Sucuoğlu and Akkar
2014; Lu and Guan 2016), and almost all the multi-mode pushover analysis methods focus on
the deformation of the structure. However, the generic forces, such as shear force and bending
moment are also of great importance when conducting the seismic design of structures (Ohtori
et al. 2004), but few of the multi-mode pushover analysis methods are practically used for
calculating earthquake-induced forces in tall buildings.
5
1.2 Objectives and scopes of the thesis
The primary objective of this research study is to propose a pushover-based method of analysis
that is effective, efficient, as well as user-friendly and can provide a good and reasonable
estimation of the seismic demand including deformation as well as the generic forces of
different types of structures.
This thesis has seven chapters, first of which is Chapter 1, where the introduction of the
background and discussion of the current seismic evaluation of structures is provided. At the
end of Chapter 1, the objectives of this research study, as well as the scope of this Ph.D. thesis,
is included. In Chapter 2, detailed review of previous pushover analysis methods, including
conventional pushover analysis methods that are included in the codes of practice and multi-
mode pushover analysis methods developed during the past few decades, is presented.
Discussions of this method are also conducted.
Chapter 3 presents a simplification of the mode coupling for efficiently considering the mode
coupling first. Based on the simplification, a spectrum-based pushover analysis (SPA) method
is proposed for the estimation of the seismically induced deformation of structures with
symmetric buildings plans. A comprehensive case study is presented in this chapter to verify
the accuracy of the SPA method. In Chapter 4, a modification of the SPA method is proposed
for the prediction of the earthquake-induced deformation of the dual wall-frame structures.
Four high-rise dual wall-frame structures are investigated to study the applicability of the
proposed modified SPA method. In Chapter 5, the torsional effect is incorporated in the SPA
method for the fast prediction of the seismically-induced deformation of structures with singly
symmetric building plans, whose accuracy has also been verified through investigation of
frames with different heights and mass eccentricity ratios.
In Chapter 6, the SPA method is further extended, by including a linear response spectrum
analysis procedure, to quickly estimate the earthquake-induced forces of structures with
symmetric building plans and a case study involving two frame structures and two shear wall
structures is conducted to verify the applicability of the extended SPA method. Finally,
Chapter 7 presents concluding remarks and possible future works.
6
Chapter 2
Literature Review
2.1 Introduction
In this chapter, the theory of structural dynamics and earthquake engineering, which is the basis
of all the analysis methods and procedures, will be presented first. Then the provisions of
conducting seismic evaluation and analysis of buildings in some of the most popular seismic
design codes of practice in the world, including, ATC-40, FEMA-356, ASCE 7-10, Eurocode
8, as well as the Chinese GB code, are discussed. After that, the current multi-mode nonlinear
static methods, such as adaptive modal pushover analysis (AMPA) methods, modal pushover
analysis (MPA) methods, extended N2 method, consecutive modal pushover analysis (CMP)
methods and other advanced pushover analysis methods will be reviewed. Finally, a summary
of the literature review is made.
8
2.2 Theory of structural dynamics and earthquake engineering for two-
dimensional problems
The elastic response of an N-degree of freedom system under the ground motion is governed
by the following equation (Chopra 2012)
  cu  ku  miug (t )
mu (2-1)
Where u , u , u
 , are the floor displacements, velocity and acceleration vector respectively; m ,
c and k are the mass, classical damping and stiffness vector of the system respectively; i is
the unit vector. The right-hand side of Equation (2-1) represents the effective earthquake force
and can be written as
peff (t )  miug (t )  sug (t ) (2-2)
where s represents the spatial distribution of the effective force over the height of the building,
and can be expanded as a summation of the modal inertia force distributions s n as follows:
N N
s   s n   n mn (2-3)
n 1 n 1
where n is the nth natural vibration mode of the structure, and:
Ln
n  , Ln  nT mi , M n  nT mn (2-4)
Mn
By the summation of the model response, the displacement of an N-degree of freedom system
can be expressed as
N
u(t )   n qn (t ) (2-5)
n 1
where the modal coordinates are governed by

qn  2nn qn  n2qn  nug (t ) (2-6)
The solution of Equation (2-6) is given by

qn (t )  n Dn (t ) (2-7)
where Dn (t ) is governed by the equation of motion for a single-degree-of-freedom (SDOF)
system subjected to ug (t ) :
  2  D   2 D  u (t )
D (2-8)
n n n n n n g
The floor displacement can then be represented as
9
N N
u(t )   u n (t )    nn Dn (t ) (2-9)
n 1 n 1
Since any response quantities r (t ) , such as inter-storey drifts, hinge rotations, internal element
forces can be expressed as
r (t )  rnst An (t ) (2-10)
An (t )  n2 Dn (t ) (2-11)
Where rnst is the modal static response, induced by the external force sn ; An (t ) is the pseudo-
acceleration response of the nth mode. The total response quantities of the system to the ground
motion excitation is
N N
r (t )   rn (t )   rnst An (t ) (2-12)
n 1 n 1
When the structure yields, the unloading and reloading curves will be different from the initial
loading branch. Thus, the lateral force is no longer independent of the loading history and
should be expressed as:
Fs  Fs (u, t ) (2-13)
Thus, the governing equation of motion for the nonlinear system should be represented as
  cu  f s (u, signu )  miug (t )
mu (2-14)
The resisting force depends on all model co-ordinate qn (t ) since the orthogonal property of
resisting force is not valid in the inelastic system due to the yielding of the system, the Equation
(2-5) cannot be used for the displacement. In other words, the modes are not independent, and
there is significant coupling effect between different modes.
2.3 Current provisions in codes of practice
In current codes of practice, both nonlinear static analysis method (known as pushover analysis
method) and nonlinear response time history analysis (NLRHA) method are provided for the
computation of the nonlinear seismic demand of structures.
10
2.3.1 Nonlinear static analysis method
Currently, most of the nonlinear static procedures recommended in the codes of practice
convert a structure, which is a multi-degree of freedom (MDOF) system, into an equivalent
single degree (ESDOF) of freedom structural models and the ground motion actions are
represented by response spectra. The top displacement is estimated by different ways and is
used to calculate the storey drift and component actions by the pushover or capacity curve that
were used to generate the ESDOF model. The main difference between different procedures
is the way to build ESDOF models and the technique to estimate the target displacement.
Following is the detail information of several conventional nonlinear static methods provided
in the codes of practice.
a. ATC-40 method
The nonlinear static analysis procedure in ATC-40 (ATC 1996) is based on an equivalent
linearization process. The primary assumption for the equivalent linearization method is that
the maximum inelastic deformation of a nonlinear SDOF system can be approximated from
the maximum deformation of a linear elastic SDOF system that has a longer period and a higher
damping ratio. In the Capacity-Spectrum Method of ATC-40, a force-deformation relationship
of the structure is generated first and then converted into acceleration displacement response
spectrum (ADRS) format. The force-deformation relationship in the ADRS format is also
known as the structure capacity curve. The ground motion records is also converted to ADRS
format which enables the capacity curve to be plotted on the same axes as the seismic demand.
In the ADRS plot, the period of structure can be represented as radial lines emanating from the
origin.
In the Capacity-Spectrum Method, the equivalent damping of the system is proportional to the
area enclosed by the capacity curve. The equivalent period, Teq , is assumed to be the secant
period at which the seismic ground motion demand, reduced for the equivalent damping,
intersects the capacity curve. Since the equivalent period and damping ratio are both a function
of the displacement, the solution to determine the maximum inelastic displacement (i.e.,
performance point) is iterative. The detailed procedure to calculate the structural capacity
curve is summarized in following:
11
1. develop a structure model;
2. Apply lateral forces to the structure, and the force is in proportion to the product of the
storey mass and fundamental mode shape;
3. Increase lateral loading until the structure reaches an ultimate limit;
4. Explicitly model global strength degradation. A new capacity curve is then created.
Create as many additional pushover curves as necessary to adequately define the overall
loss of strength as shown in Figure 2.1;
5. Plot the final capacity curve. Firstly, follow the first curve, then transit to the second
curve at the displacement corresponding to the initial strength degradation. Repeat this
for several times until to overall loss of strength is well captured. This procedure is
demonstrated in Figure 2.2.
Figure 2.1 Multiple capacity curves required modelling strength degradation.
12
Figure 2.2 Capacity curve with global strength degradation modelled.
After determining the capacity curve, the demand curves should also be constructed. This
procedure is mainly about finding the intersection point of the capacity spectrum and the
appropriate demand response spectrum, reduced for nonlinear effects. The equivalent damping
ratio is computed based the shape of the capacity curve, the estimated displacement demand,
and the resulting hysteresis loop:
1. Convert the capacity curve to the ADRS. The equations to transform are:
 N 
   wii1  / g 
1   iN1  (2-15)

  wii1 / g 
2

i 1
2
N 
   wii1  / g 
1  N  i 1  (2-16)
  N 
  wi / g     wii1  / g 
 i 1   i 1 
Vi / w
Sai  (2-17)
1
 roof
Sdi  (2-18)
PF1  1,roof
Where: 1 is modal participation factor for the first natural mode; 1 is modal mass coefficient
for the first natural mode; wi / g is the seismic mass of level i; i1 is the mode shape of mode
1 at level i; N is level which is the uppermost in the main portion of the structure; V is the base
13
shear; W is the total seismic mass of the building;  roof is the roof displacement; Sa is the
spectral acceleration; Sd is the spectral displacement;
2. Construct Bilinear Representation of the capacity spectrum;
3. Intersect Capacity Spectrum and Demand Spectrum as illustrated in Figure 2.3.
Figure 2.3 Intersection point of Demand and Capacity Spectrum.
b. FEMA 356 method
The nonlinear static procedure recommended in FEMA 356 (FEMA-356 2000) is a Coefficient
Method. In this approach, the maximum global displacement (elastic and inelastic), which is
called the target displacement, is determined by multiplying the linear elastic response of the
ESDOF system by a series of coefficients C0 through C3 . In the beginning, an idealized force-
deformation curve relating base shear to roof displacement is generated for the structure, as
depicted in Figure 2.4. An effective period, Teff , is obtained from the initial period, Ti , by a
graphical procedure considering the loss of stiffness in the transition from elastic to inelastic
behaviour.
The peak elastic spectral displacement is directly related to the spectral acceleration by:
14
Teff2
Sd  Sa (2-19)
4 2
And the roof peak inelastic displacement,  t , can be estimated as
Teff2
 t  C0C1C2C3 Sa (2-20)
4 2
The coefficient C0 is a shape factor that converts the spectral displacement to the displacement
at the roof and the value is usually taken as the participation factor of the first mode. The other
coefficients account for a separate inelastic effect respectively. The coefficient C1 is the ratio
of expected displacement (elastic plus inelastic) for a bilinear inelastic system to the
displacement for a linear system. This ratio depends on the strength of the system about the
response spectrum and the period of the SDOF system, Teff . The coefficient C2 accounts for
the effect of pinching in load-deformation relationships due to degradation in stiffness and

strength. Finally, the coefficient C3 adjusts for second-order geometric nonlinearity  P  
effects. The coefficients are empirical and derived primarily from statistical studies of the
nonlinear response-history analyses of SDOF oscillators and changed using engineering
judgment.
Figure 2.4 Schematic illustrating the FEMA 356 process to estimate the target displacement for a
given response spectrum and effective period.
15
c. Eurocode-8
The nonlinear static analysis presented in Eurocode-8 (CEN 2004) is a Coefficient Method too.
An idealized SDOF model with the elasto-perfectly-plastic force-displacement relationship is
generated. The target displacement is calculated from the  R    T  relation of the idealized
SDOF system.
In the pushover analysis, the forces Fi are proportional to the production of mi and specific
horizontal displacement pattern i :
Fi   mii (2-21)
According to EC- 8, pushover analysis should be performed using either uniform load pattern
or the modal load pattern of the fundamental mode.
The equivalent SDOF for pushover analysis is conducted using the N2 procedure given in
informative Annex B of EN 1998-1. The mass of the equivalent SDOF system m* is
m*   mii (2-22)
And the force F * and displacement d * of the equivalent system are

Fb
F*  (2-23)

dn
d*  (2-24)

m*
 (2-25)
 mii2
Where Fb and dn respectively, the base shear force and the control node displacement of the
Multi Degree of Freedom (MDOF) system.
16
Figure 2.5 Determination of the idealized elasto-perfectly plastic force-displacement relationship.
The yield force 𝑭𝒚∗ , which is also the ultimate strength of the idealized system, is taken as the
base shear force when plastic mechanism forms. The initial stiffness of the idealized system is
determined to ensure that the areas under the actual and the idealized force-deformation curves
are equal, which is denoted in Figure 2.5. Based on this assumption, the yield displacement of
the idealised SDOF system 𝒅𝒚∗ computed as
 * Em* 
d  2  dm  * 
*
(2-26)
y  Fy 

Where Em* is the actual deformation energy up to the formation of the plastic mechanism.
Then, the period of the idealized equivalent SDOF system can be computed as
m*d y*
T *  2 (2-27)
Fy*
The target displacement of the structure with period T * and linear elastic behaviour is given
by:
2
 T* 
d  Se (T )  
* *
(2-28)
 2 
et
Where Se (T * ) is the elastic acceleration response spectrum at the period T * .
Then the target displacement of the structure with period T * and elasto-perfectly plastic
17
behaviour dt* can be calculated from d et* . For structures in the short-period range and
structures in the medium and long-period ranges, different expressions should be used as
indicated below. The corner period between the short- and medium-period range is TC .
a) T *  TC (Short period range)
If Fy / m  Se (T ) , the response is elastic and thus

* * *
dt*  d et* (2-29)
If Fy / m  Se (T ) , the response is nonlinear and

* * *
det*  TC 
dt*  1   qu  1 T *   det
*
(2-30)
qu  
Where qu is the ratio between the acceleration in the structure with unlimited elastic behaviour
Se (T * ) and the structure with limited strength Fy* / m* .
b) T *  TC (medium and long period range)
dt*  d et* (2-31)
The relation between different quantities can be visualized in Figure 2.6.
18
(a) Short period range
(b) Medium and long period range

Figure 2.6 Determination of the target displacement for the equivalent SDOF System.
The target displacement corresponds to the control node can then be calculated as
dt  dt* (2-32)
d. Chinese Standard 2008

The simple method to estimate the nonlinear inter-storey drift of the structures recommended
in Chinese standard (Chinese Standard 2008) is also a coefficient method. The inelastic inter-
storey drift can be expressed as
p
u p   p ue  u y  u (2-33)
y y
Although there are many procedures to determine the roof displacement, the processes are quite
complicated. What is more, since, in all the codes of practice, the buildings are simulated by
the equivalent single degree of freedom systems, and just the fundamental mode is considered.
19
However, for tall buildings, where the contribution of high modes is significant, using single-
mode pushover analysis is not capable of estimating the seismic demand of tall buildings with
high accuracy.
2.3.2 Nonlinear response time history analysis method
In most the of codes of practice, nonlinear response time history analysis method is also
recommended for conducting the seismic evaluation of buildings. However, since the
frequency contents of the earthquakes varies significantly for different motions, the seismic
response of the buildings under different input ground motions is also completely different,
even if the motions are scaled to have similar response spectra. Consequently, in all the codes
of practice, multi-records requirements are set to minimize the uncertainty caused by the
motion selection and to ensure the safety of buildings. The multi-records requirements are very
similar in different codes of practice:
 Take the average response quantities if at least seven ground motions are analysed
 Take the most unfavourable response quantities if at least three ground motions are
investigated
Although analysing the response of buildings under more motions will reduce the uncertainty
induced by the frequency contents of motions, it will increase the computation demand of the
analysis dramatically, and there are still variations for a different selection of motions. Thus,
it is still not very suitable for quick evaluation of the seismic response of tall buildings.
2.4 Current advanced pushover procedures
In high seismicity regions, such as Chile and Japan, most buildings were designed with
symmetric plans to avoid the serious torsional effect, which will cause unexpected damage to
buildings under seismic actions. During the past few decades, considerable research efforts
have been devoted to developing multi-mode pushover analysis methods for the fast yet
accurate estimate of the seismic demand of tall buildings with symmetric building plans
(Antoniou, Rovithakis, and Pinho 2002; Antoniou and Pinho 2004a; Antoniou and Pinho 2004b;
Brozovič and Dolšek 2014; Chintanapakdee and Chopra 2003; Chopra and Goel 1999; Chopra
and Goel 2002; Chopra, Goel, and Chintanapakdee 2004; Fajfar and Fischinger 1988; Fajfar
and Gašperšič 1996; Fajfar 1999; Fajfar 2000; Ferracuti et al. 2009; Fragiadakis and
20
Vamvatsikos 2010; Goel and Chopra 2004; Goel and Chopra 2005; Jan, Liu, and Kao 2004;
Jiang, Li, and Yang 2010; Kalkan and Kunnath 2006; Kalkan and Kunnath 2007; Kreslin and
Fajfar 2011; Leelataviwat, Saewon, and Goel 2009; Lin and Chang 2003; Liu and Kuang 2017;
Liu, Kuang, and Huang 2018; Miao et al. 2011; Miranda 1999; Moghaddam and Hajirasouliha
2006; Papanikolaou and Elnashai 2005; Poursha, Khoshnoudian, and Moghadam 2009;
Poursha and Samarin 2015; Sucuoğlu and Günay 2011; Zhang, Jiang, and Li 2017; Abbasnia,
Davoudi, and Maddah 2013). However, in regions with moderate and low seismic activity,
where non-seismic design philosophy is adopted in the design of structures, buildings with
asymmetric plans are very popular to achieve attractive architectural appearance. Buildings
with an asymmetric plan are more vulnerable to earthquake actions, compared to the buildings
with a symmetric-plan arrangement, since the torsion effect caused by the asymmetricity will
induce higher seismic demand and may cause unexpected damage to the buildings. During the
past two decades, impressive progress has been achieved in the development of the fast
prediction for the seismic demand of buildings with asymmetric plan arrangements (Baros and
Anagnostopoulos 2017; Bhatt and Bento 2011; Bosco et al. 2013; Chopra and Goel 2004;
Khoshnoudian and Kiani 2012; Kilar and Fajfar 1997; Kreslin and Fajfar 2012; Manoukas,
Athanatopoulou, and Avramidis 2012; Poursha, Khoshnoudian, and Moghadam 2011; Poursha
and Samarin 2015; Reyes and Chopra 2011a; Reyes and Chopra 2011b; Shakeri, Tarbali, and
Mohebbi 2012; Tarbali and Shakeri 2014; Kaats𝚤z and Sucuoğlu 2014).
Extended N2 method
The original N2 method was developed by Fajfar (Fajfar and Fischinger 1988; Fajfar and
Gašperšič 1996), which just considers the nonlinear seismic response of the fundamental mode.
Thus the original N2 method is not capable of predicting the seismic response of median- and
high-rise buildings, where the high mode effect is significant. Then Kreslin et al. extended the
N2 method (Kreslin and Fajfar 2011), based on the assumption that the structure remains in the
elastic range when vibrating in higher modes. The seismic demand regarding displacements
and storey drift can be obtained by combining the results of fundamental pushover analysis and
those of standard elastic modal analysis. In the extended N2 method, a conventional N2
prcoedure is conducted first to determine the nonlinear seismic demand of buildings of the
fundamental mode and the target roof displacement. Then the standard response spectrum
analysis (RSA) method is conducted to get the elastic response of the buildings. The elastic
response from RSA is scaled so that the roof displacement of the RSA equals to that of the
21
basic N2 method. Finally, the final response of the buildings is taken as the envelope of the
results of the basic N2 method and the scaled results of RSA method. The extended N2
procedure is straightforward, and after considering the high mode effect, it can predict the
seismic response of tall buildings well. However, since the coupling of the mode of vibration
is ignored, there is still a significant gap between the response quantities from NLRHA and
extended N2 method, especially for the localized responses such as inter-storey drift ratio and
hinge plastic rotation.
Similar to the extended N2 method for buildings with symmetric building plans, Kreslin et al.
proposed an extended N2 method for the quick estimation of the seismic demands of structures
with asymmetric building plans (Kreslin and Fajfar 2012), by including the torsional mode
effect and considering separately the seismic demand of different vibration modes. However,
It is noticed that although the extended N2 method includes the torsional effect and the multi-
mode effect, it still ignored the coupling of the modes in the nonlinear vibration of structures.
Thus, the extended N2 method cannot predict the seismic deformation of the buildings with
asymmetric building plans very well.
Envelope-based Pushover Analysis Method
The envelope-based pushover analysis method was developed by Brozovic et al. (Brozovič and
Dolšek 2014). In the envelope-based pushover analysis method, the authors studied the
damage modes of the structures caused by lateral forces with different mode distributions
separately and proposed failure-based SDOF model for the structures. Nonlinear response time
history analysis procedure is performed with the failure-based SDOF model separately to
obtain the seismic input for the pushover analysis procedure. Modal pushover analysis
procedures, where a lateral force with the distribution of the mode shape, are conducted with
the seismic input computed from the NLRHA of the failure-based SDOFs. The final seismic
demand of the structure is the envelope of the seismic demand of all the pushover analysis
procedures.
Although the envelope-based pushover analysis procedure considers different damage modes
of the structures and can provide a good prediction of the seismic demand of structures, it still
includes the effect of varying vibration mode separately and cannot offer a satisfactory estimate
of the seismic demand at the middle part of the structures.
22
Upper-bound pushover analysis procedure
The upper-bound pushover analysis (UPA) procedure proposed by Jan et al. (Jan, Liu, and Kao
2004) includes a single run pushover analysis procedure, where the lateral force adopted has a
distribution that consists of the first mode shape and the corrected second mode shape. The
target roof displacement of the pushover analysis procedure is a factored displacement that is
calculated by the response spectrum analysis of the equivalent SDOF systems. However, since
the force vector in the UPA procedure indicates that the first mode and second mode will affect
the structural behaviour at the same pace, which is not reasonable, the UPA procedure does not
consider the multimode effect appropriately. It is found the upper-bound pushover analysis
procedure cannot provide a satisfactory prediction of the seismic demand of structures (Kalkan
and Kunnath 2007).
A modification of the UPA procedure was proposed by Poursha et al. (Poursha and Samarin
2015), which supplements a conventional pushover analysis that is performed with a uniformly
distributed load or an inverted triangular distributed load in the upper-bound pushover analysis
procedure. It is found that the modified UPA (MUPA) procedure can have a much better
prediction of the seismic demand of structures as compared with the UPA procedure. However,
since a uniformly distributed load or an inverted triangular distributed load cannot well
simulate the effect of higher modes, the MUPA procedure still cannot provide a satisfactory
prediction of the seismically induced deformation of the structures.
An extended UPA (EUPA) method was developed by Poursha et al. (Poursha and Samarin
2015) for the estimate of the seismic demand of structures with asymmetric building plans. In
the EUPA method, lateral forces with distributions that consist of the first mode shape and the
corrected higher modes shapes are applied to the structures separately to include the torsional
effect in the calculation the seismic demand of structures with asymmetric building plans. The
final demand of structures with asymmetric building plan is the envelope of the maximum
demands of the pushover analysis procedures using the forces corrected multimode shapes and
those of the conventional pushover analysis method. Similar to EUPA procedure, which is
initially developed for buildings with symmetric plans, the extended upper bound pushover
analysis method can provide a reasonable estimation of the seismic deformation of the
structures with asymmetric building plans. Meanwhile, since the interaction of higher modes
23
is not well dealt with, the seismic demand from the extended upper bound pushover analysis
method satisfactorily.
Modal Pushover Analysis Method
The modal pushover analysis method (MPA) developed by Chopra et al. (Chopra and Goel
2002) is another prevalent method of assessing the seismic response buildings. In MPA, the
interaction of modes in the nonlinear stage is neglected, and pushover analyses are conducted
independently for each vibration mode. Modal response values for each mode are determined
when the target roof displacement is reached. The final response quantities of buildings are
obtained from the combination of the results of all the modal pushover procedures, by using
the SRSS method. In the MPA method, the mode shapes and lateral force profiles are assumed
to be invariant in this analysis procedure. Target displacement values may be computed by
applying displacement modification or equivalent linearization procedures to an elastic
spectrum for an equivalent SDOF system representative of each mode to be considered.
Goel et al. further modified the MPA method (Goel and Chopra 2004), in which the P-∆ effects
in all modes are considered, and only the nonlinearity of the first mode is considered, by
assuming the higher mode will remain elastic. The modified MPA can increase the
computational efficiency of the method and provide a conservative estimation of the seismic
response. However, since both the MPA method and modified MPA method ignore the
coupling of the modes in the nonlinear stage, the MPA procedures cannot provide an accurate
estimation of the localized seismic responses especially the hinge plastic rotation.
By including the torsional moment in modal forces and computing the seismic demand of
vibration modes separately, the modal pushover analysis (MPA) method was extended to three
dimensional cases to conduct seismic evaluation of asymmetric buildings and a three
dimensional MPA was also developed for the estimated the seismic demand of structures with
doubly asymmetric building plan under two horizontal ground motions (Chopra and Goel 2004;
Reyes and Chopra 2011a; Reyes and Chopra 2011b). Even though it is found that the MPA,
as well as the three dimensional MPA, can reasonably predict the seismic demand of the
structures with asymmetric building plans, the MPA did not consider the mode coupling effect,
and thus it still cannot provide a satisfactory prediction of the seismic demand.
24
Pushover analysis method based on energy balance concept
Energy balance concept was first applied in developing a nonlinear static method by
Leelataviwat et al. (Leelataviwat, Saewon, and Goel 2009). In their pushover model, the
energy balance concept, which was initially proposed for single-degree-of-freedom (SDOF)
system, was extended to multi-degree-of-freedom (MDOF) systems for analysing the seismic
performance of structures. The energy balance concept-based pushover analysis method firstly
computes the structural energy capacity curve by calculating the area of the pushover curve of
equivalent SDOF system of the structures. Then, the structural energy capacity curve with the
seismic energy demand curve, which is obtained from the response pseudo-velocity spectrum
of a ground motion or design level is compared. The seismic input for the pushover analysis
procedure is the target roof displacement, which can be obtained from the intersection of the
energy demand and capacity curves. After getting the target roof displacement, pushover
procedure is conducted using lateral force with first mode-shape distributions to load the
structure until the target roof displacement is reached and then obtain the maximum demand
of the computed seismic demand. However, since in Leelataviwat’s model, the high-mode
effect is not considered, the energy balance concept-based pushover analysis method can just
be applied to the seismic analysis of low-rise buildings.
To incorporate the multi-mode effect, a modified approach of energy balance concept based
multimode pushover analysis was proposed by Jiang et al. (Jiang, Li, and Yang 2010). In the
modified approach of energy balance concept based multimode pushover analysis, pushover
analysis procedures were conducted separately with forces with different mode shape
distributions to generate the structural energy capacity curves were produced for all the modes.
The target displacements for various modes can then be calculated by comparing the
corresponding structural energy capacity curve with the seismic energy demand curve. Then,
lateral forces with different mode-shape distributions are applied to the structure separately to
load to structure to the target roof displacement to obtain the maximum demand. The final
structural demand is the combination of the maximum demands of a different mode.
In the pushover analysis methods based on energy balance concept, the seismic demand of the
structures can be better estimated, as compared with the conventional pushover analysis
procedures. But since the methods consider the multi-mode effect separately and neglect the
mode coupling effect, the energy balance concept-based pushover analysis method and
25
modified energy balance concept-based pushover analysis method still cannot well estimate
the seismically induced deformations, such inter-storey drift ratio.
Generalised pushover analysis procedure
In the generalised pushover analysis (GPA) procedure, developed by Sucuoğlu et al. (Sucuoğlu
and Günay 2011), the target seismic input for the pushover analysis procedure is the maximum
inter-story drift ratios of each story that are calculated by the response spectrum analysis
procedure. A series of generalised force vectors, whose distributions are separately calculated
based on the maximum inter-story drift ratios of each story, is used to perform pushover
analysis procedures individually to load the structure to reach the corresponding target seismic
inputs. The final seismic demand is the envelope of the maximum seismic demands of all the
pushover analysis procedure. It is found that the GPA procedure can provide a good prediction
of the seismic demand. For the quick estimate of the seismic demand of structures with
asymmetric building plans, Kaatsız et al. (Kaats𝚤z and Sucuoğlu 2014) include the torsional
effect in the calculation of the GPA procedure. It is found that the GPA can also well predict
the seismic demand of structures with asymmetric building plans.
However, since the inter-storey drift ratio is one of the crucial structural performance index
and essential seismic input for the pushover analysis procedure, it is of great significance to
have an accurate prediction of the inter-storey drift ratio. In the GPA procedure, the inter-
storey drift ratio is calculated based on the response spectrum analysis procedure, which is
proved to not able to predict the inter-storey drift well. Thus, the inter-storey drift ratio results
from the GPA procedure is not satisfactory and may also affect the results of the pushover
analysis procedure. What is more, although it is stated that the GPA procedure does not
consider the vibration modes separately, the generalised pushover analysis method assumes
that the vibration modes affect the vibration at the same pace, which is not reasonable, due to
the complicated frequency contents of the ground motion and the significant differences
between the vibration periods of different modes.
Consecutive modal pushover analysis methods
Consecutive modal pushover analysis (CMP) method (Poursha, Khoshnoudian, and

Moghadam 2009) and modified consecutive modal pushover (MCMP) method (Khoshnoudian
26
and Kashani 2012) proposed by Poursha and Khoshnoudian, consist of multistage and single-
stage pushover analyses. In the multi-stage pushover analysis, a few modal pushover analysis
procedures were conducted consecutively to reach the target roof displacement. The
consecutive pushover procedure adopted are that when one pushover analysis procedure
completed, the next pushover procedure started with the initial structural state (like internal
stress, strain, and displacement) the same as the structural state at the end of last pushover
analysis procedure. The target roof displacement of each mode in the CMP procedure is
proportion to the modal mass participation factors, and, in the CMP method, the total roof
displacement used is the mean roof displacement. While. the contribution of different vibration
modes in the MCMP procedure is proportion to the elastic modal participating factor. More
recently, an improved CMP (ICMP) method (Zhang, Jiang, and Li 2017) was developed by
Zhang et al., where a single run pushover analysis performed with a lateral load having a
distribution of the combination of different mode shapes is added in the analysis procedure of
the MCMP method. The additional pushover procedure considers the interaction of different
modes in the vibration. It is found that all the CMP methods can provide good prediction of the
seismic demands, especially the hinge plastic rotation.
Poursha et al. (Poursha, Khoshnoudian, and Moghadam 2011; Poursha, Khoshnoudian, and
Moghadam 2014) and Khoshnoudian et al. (Khoshnoudian and Kiani 2012) included the
torsional effect in the CMP and MCMP methods respectively, for the fast prediction of the
seismic demand of structures with asymmetric building plans. It is found that the CMP, as well
as MCMP, can provide reasonably estimation of the seismic demand of structures with
asymmetric building plans when compared with the results of nonlinear response time history
analysis.
However, the mass proportional pushover procedure of the CMP method will overestimate the
effect of higher modes, resulting in improper dealing with the coupling effect of modes. Thus,
the accuracy of the CMP method is affected, especially in the upper part of the buildings. While
in the MCMP method, the contribution of different vibration modes is proportion to the elastic
modal participating factor, which is more reasonable. However, as the modal property will
change due to the yield of the structure, especially the fundamental mode, using the elastic
modal property to calculate the modal contribution will also induce errors in the computation
of the seismic demand of structures. In the ICMP method, the additional pushover procedure
assumes that the interaction of modes affects the deformation of the structures at the same pace,
27
which is not reasonable as the frequency contents of the ground motion are complicated and
the natural periods of different mode are different. Additionally, in all the research articles
about the CMP methods, the authors always used the mean roof displacement of the NLRHA
method as the overall target roof displacement. There is no specific method used in the
publications to determine the total target roof displacement. It is meaningless to conduct the
pushover analysis when the NLRHA method is already performed for the structures. Besides,
even with accurate mean roof displacement, there are still significant disparities between the
results of CMP methods and NLRHA method.
Adaptive Pushover Analysis Method
The first version of the adaptive pushover analysis method was developed by Gupta et al.
(Gupta and Kunnath 2000), which accounts for high mode effects by combining the response
of individual modal pushover analyses and considers the progressive change of the modal
property under the applied loads by using the adaptive analysis algorithm. After that, an
adaptive modal combination procedure is developed by Kalkan et al. (Kalkan and Kunnath
2006), where the energy-based formulation of the ESDOF system is adopted to compute the
displacement increment for each step by employing an inelastic response spectrum plotted in
AD format. Thus, the adaptive modal combination procedure includes the graphical feature of
the capacity spectrum method as well. The final response of the structures is a combination of
the response of each pushover procedure. Meanwhile, Casarotti et al. (Casarotti and Pinho
2007) proposed an adaptive capacity spectrum method, using displacement-based adaptive
pushover analysis to simulate an equally adaptive SDOF capacity curve in an AD format. The
equally adaptive SDOF capacity curve is then used to compute the performance index by
intersecting an appropriate overdamped elastic spectrum. This procedure considers both the
effects of higher modes and varying dynamic characteristics through its adaptive feature in
pushover analysis as well as in the definition of the SDOF capacity curve. Additionally, it also
eliminates the dependency of the capacity curve on the choice of displacement control node,
are the SDOF capacity curve is generated at each pushover step using the actual deformation
pattern of the original structure. The final response of the structures is a combination of the
response of each pushover procedure.
To accounts for the interaction of modes in the nonlinear stage, a storey shear-based adaptive
modal pushover procedure was proposed by Shakeri et al. (Shakeri, Shayanfar, and
28
Kabeyasawa 2010). The method is based on the storey shears, which takes into account the
reversal of sign in the higher modes. In each step, the applied load pattern is obtained from the
instantaneous combined storey shear profile of the modes considered. The sign of the applied
loads in consecutive steps are changed, and the structure is simultaneously pushed and pulled
in different storey levels. At each analysis step, an assumed fundamental mode shape is derived
from the load profile. Based on this adaptive primary mode shape and the energy concept, the
multi-degree of freedom system is converted into a single degree of freedom system. An
effective modal mass combination rule was proposed Abbasnia et al. (Abbasnia, Davoudi, and
Maddah 2013). In This combination rule, modification factors associated with the modes of
interest are computed and applied to the corresponding load vector. The modified modal load
vectors are directly superposed and subtracted, resulting in a range of load pattern. These load
patterns are independently applied to the structure within an adaptive framework, and the
envelope of demand values is considered. These modification factors are updated
proportionally to the instantaneous dynamic characteristic of structure in each step. By
including the torsional effect in the adaptive pushover analysis procedures (Shakeri, Tarbali,
and Mohebbi 2012; Tarbali and Shakeri 2014; Bhatt and Bento 2014), the adaptive pushover
analysis procedures can be further utilised for the seismic evaluation of structures with
asymmetric building plans.
Although the adaptive pushover analysis methods consider the progressive damage in the
structure and the continuous change of the modal properties, and some of the advanced adaptive
pushover analysis procedures even consider the interaction of modes in the nonlinear stags, the
method requires that the modal properties should be updated as frequently as possible to ensure
the accuracy of the estimation. This makes the computation cost of the adaptive modal
pushover method comparable to that of the NLRHA method, even though there are still some
variations between the results of the adaptive modal pushover method. Consequently, the
adaptive modal pushover analysis procedures are much less efficient and not capable of quickly
estimating the seismic demand of buildings, especially tall buildings.
29
2.5 Summary
In this chapter, the primary structural dynamics were presented followed by the seismic
evaluation procedures required by codes of practice. Then a detailed review of the advanced
multi-mode pushover analysis methods that were proposed in past decade is presented. Based
on the study of currently developed pushover analysis methods, the following defects can be
found in current multi-mode pushover analysis procedures:
1. Although all the newly developed pushover analysis methods can provide a good prediction
of the seismic demand of buildings, none of them incorporates the mode coupling effect
very efficiently and reasonably.
2. All the pushover methods mentioned above just developed for estimating the seismic
response of frame structures. Acutely few publications assessed the applicability of the
pushover analysis method for predicting the seismic response of other types of structure
like shear wall structure and wall-frame structure.
3. The deformation response of the structure is of great importance to evaluate the safety of
buildings, the generic force induced by the earthquake is another crucial quantity revealing
the required forces for the structures to take under the attack of earthquakes. There are very
few publications compared the generic forces computed by the pushover analysis method
and that from NLRHA method.
30
Chapter 3
Spectrum-based Pushover Analysis for
Estimating Seismic Deformation of
Frame and Wall Structures
3.1 Introduction
This chapter firstly presents a simplification of the coupling effect of modes in the nonlinear
state. Based on the simplification, the spectrum-based pushover analysis (SPA) method is
proposed and the second part of this chapter presents the details of the SPA method.
Accordingly, the assessment of current pushover procedures and the proposed SPA method for
seismic evaluation of different types of buildings with a symmetric building plan will be
presented. Overall four building structures with different heights were investigated to consider
the effect of different structural behaviours and high mode effect. It is found from the
assessment that the SPA method can provide a very good estimation of the deformations for
both frame and wall structures.
32
3.2 Spectrum-based pushover analysis method
By conducting the modal analysis, the solution of the equations of motion of a linear elastic
multi-degree of freedom system can be expressed as
N N
u(t )   u n (t )    nn Dn (t ) (3-1)
n 1 n 1
The peak roof displacement can be expressed as a superposition of the target roof displacement
of the important modes of vibration, such as
N
ur max   urn 0 (3-2)
n 1
The equation (3-2) can also be interpreted as that all the important modes of vibration force the
structure to have the target roof displacement at the same time to induce the peak roof
displacement. Assuming that the natural periods of different modes of the building are well
separated, which means that 𝑡 𝑚𝑡 and 𝑚 ≫ 1. As a result, to make sure that different
modes have the target roof displacements at the same time instant, all the modes of vibration
except the fundamental mode start the vibration much later than the previous modes. Thus, the
case all the modes of vibration reach the target roof displacement at the same time can even
be equivalented as the case that when the 𝑖th mode of vibration starts, the roof displacements
of all the previous modes, i.e. the 𝑖 1 th, 𝑖 2 th, …, and 1st mode of vibration have
already reached their corresponding target values and the structural states, like storey
displacements, inter-storey drift, internal stress and strain of 𝑖th mode, keep unchanged after
the roof displacement of 𝑖th mode reaches the target value. When the structure stays in the
elastic range, this assumption would cause no error, as all the modes of vibration are
independent and the linear superposition of the roof displacement of different modes of
vibration is correct.
However, when the structure yields and steps into the inelastic range, the governing equation
of motion is changed as shown in Equation (3-3). The solution of Equation (3-3) cannot be
expressed by Equation (3-1) as in nonlinear stage, different modes are not independent and
33
there are coupling effects for different modes. Thus, the applicability of this assumption should
be reconsidered.
mu  cu  f s (u, signu )  miug (t ) (3-3)
Although the overall roof displacement of the structure cannot be expressed as the linear
superposition of the target roof displacement of the important modes, the peak roof
displacement of a nonlinear structure should also be the combination of the target displacement
of the modes. Consequently, it is still assumed that the peak roof displacement of the structure
is achieved by the case that the roof displacement caused be different modes of vibrations
reached the corresponding target values at the same time. For simplicity, the following
assumptions are used. It is assumed that the input motion data has three frequency contents
with frequencies identical with the first three modes of the building. The elongation of the
natural periods of the building due to the yield of the structure is ignored, as all the natural
periods will be elongated and the well separated natural periods are still well separated. Thus,
the delaying phenomenon between different modes of vibration will not be affected too much.
Figure 3.1 illustrates the normalised displacement time history of the case that the roof
displacements of the first three modes of a building reach the target values at the same time.
In Figure 3.1, t1, t2 and t3 are the natural period of the first three modes of the structure
respectively and have a relation as t1=3×t2=9×t3. The relationship of the natural period of the
first three modes reveals that the building has well separated natural periods. 𝑢 stands for
the target roof displacement of ith mode. It can be seen from Figure 3.1 that when the structure
starts to have the 2nd mode vibration, the roof displacement of the structure under the vibration
of the 1st mode has already reached more than 85% of the target roof displacement of first
mode, and when the 3rd mode of vibration starts, the roof displacementof the strutcure under
1st and 2nd mode has already reached more than 98% and 83% of the target values respectively.
In other words, when all the modes of vibration except the fundamental mode start, the building
already has large deformation under previous modes. Then the assumption that when the 𝑖th
mode of vibration starts, the roof displacement of all the previous modes (the 𝑖 1 th, 𝑖
2 th, …, and 1st mode) of vibration have already reached the target value and the structural
states like roof displacement, inter-storey drift, internal stress and strain of 𝑖 th mode of
vibration keep unchanged after the roof displacement of 𝑖th mode of vibration reaches the
target value is still rational for the nonlinear case. With this assumption, the very complex and
complicated coupling effect of different modes in the nonlinear seismic performance of the
34
building is simplified and the modes of vibration are related by the initial structural state (like
internal stress, strain and displacement).
Figure 3.1 Roof displacements of the first three modes of a building reached the peak at the same
time.
It should be noted that as the frequency contents of the ground motion are very complicated
and unpredictable, the case where different modes reach the target roof displacement at the
same time cannot be achieved in any specific ground motion. Consequently, the simplification
of mode coupling effect cannot be utilised in the seismic evaluation of buildings under a
particular ground motion. In fact, the case where different modes reach the target roof
displacement at the same time represents a general trend of the seismic performance of a
structure under the action several ground motions that have similar response spectra. Thus, the
proposed simplification of mode coupling is only applicable for simplifying the coupling effect
of modes when determining the mean or median seismic response quantities of a structure
under a series of ground motions that have similar response spectra. Additionally, it is also
obvious that the more separated the natural periods of the structure are, the more effective and
accurate the proposed simplification is in the simplification of mode coupling.
Based on the simplification of coupling effect of modes, it is possible to reproduce the mean
quantities of the structural states such as inter-storey drift ratio, hinge rotation, stress and strain
of the building under a series of ground motions that have similar response spectra, provided
that the final roof displacement and roof displacement of each mode can be reasonably
estimated. To achieve this, a spectrum-based pushover analysis (SPA) procedure (Liu and
35
Kuang 2017) is proposed. In the proposed SPA method, the consecutive pushover technique is
adopted to reproduce structural states of the building. The main idea of the consecutive
pushover technique, developed by Poursha et al. (Poursha, Khoshnoudian, and Moghadam
2009), is to consecutively implement a few modal pushover analysis procedures to reach the
target roof displacement. The details of consecutive pushover procedure adopted are that when
one pushover analysis procedure completes, the next pushover procedure starts with the initial
structural state (like internal stress, strain, and displacement) the same as the structural state at
the end of last pushover analysis procedure. The final seismic response of the building is the
envelope of the peak response of every pushover stage. The force vectors used for the SPA
method are mode-shape distributions that are obtained from eigenvalue analysis of linear
structures. The changes of modal properties and mode shapes at the nonlinear stage are not
considered. The order of the modal pushover analysis should follow the order of modes, which
is from the first mode to higher modes. For Frame structures, all the modes that have an
effective modal participating mass ratio 𝜆 larger than 1%, should be included in the pushover
procedure. The effective modal participating mass ratio is expressed as
meff 1 L2n
n  N
 N
(3-4)
m m
Mn
j j
i 1 i 1
where ∑ 𝑚 is the total lamped mass of the structure. While for wall structures, the first three
modes should be included.
The target roof displacement for ith modal pushover procedure is obtained as
uir  i  ur 0 (3-5)
where 𝑢 is the estimated roof displacement of the structure; 𝛼 represents the contribution
factor of 𝑖th mode to the roof displacement. To determine the roof displacement 𝑢 , response
spectrum analysis method was used. SRSS combination rule was used combine the roof
displacements of different modes. Equal displacement assumption that the inelastic spectrum
displacement is equal to elastic spectrum displacement was adopted. It is noted that this
assumption underestimates the roof displacement of modes with a short natural period, while
the estimation of the roof displacement of modes with a long natural period is not affected
(Surmeli and Yuksel 2015; CEN 2004). The demarcation period, larger than which the
nonlinearity of the structure would not cause difference between the elastic spectrum
displacement and inelastic spectrum displacement, is called corner point and can be obtained
36
from the code of practice. For typical mid- and high-rise buildings, the natural periods of the
first three modes are generally larger than the corner period, meaning that the equal
displacement assumption has very a slight effect on the determination of the roof displacement.
Thus, the estimated roof displacement 𝑢 can be computed as:
0.5
 N 2
ur 0     uri 0   (3-6)
 i 1 
where, 𝑢 can be expressed as:
uri 0  i i ri Di (3-7)
β is the reduction factor for the roof displacement and the expression of β is:
 1.0, i  2,3,..., N

ur 0        (3-8)
cos(  t  )  cos 

     , i  1
 2  N 1  
1 N 1 1
 
where, 𝜔 (rad/sec) is the natural frequency of 1st mode, and 𝜔 is the natural frequency of
𝑁 1 th mode. 𝑡 is the time instant when 𝑁 1 th mode starts considering that all the
modes have the maximum roof displacements at the same time. Then the final target roof
displacement can be expressed as:
0.5
 N 2
ur 0     i i ri Di   (3-9)
 i 1 
It can be seen from Equation (3-7) that the responses of each mode contain three parts: the
modal participation part 𝛤 𝜙 ), dynamic input part (𝐷 ) and reduction factor. As a result,
when considering the contribution of different modes to the seismic performance of the
buildings, both modal participation part and dynamic input part should be considered. Thus,
the contribution factor of ith mode to the roof displacement is expressed as
uri 0 i i ri Di
i  N
 N (3-10)
 uri0
i 1
 i i ri Di
i 1
The details of the proposed SPA procedure are summarized as:
1) Apply the gravity loads to building and conduct eigenvalue analysis of structure.
Compute the natural frequencies, 𝜔 , 𝛼 , β , 𝜆 and the mode-shapes, 𝝓𝒏 . The mode-
shapes are normalized so that the roof component of 𝜙 equal to unity (𝜙 1).
37
2) Compute the force vector of different modes. 𝒔𝒏∗ 𝒎𝝓𝒏 , 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) Obtain the design spectrum for the sites of interest or the mean spectrum of a series
ground motion data and compute the target roof displacements of the structure for
different mode 𝑢 , the total roof displacement 𝑢 and contribution using Equations
(3-9) and (3-5).
4) Carry out the spectrum-based pushover analysis procedure according to the following
sub-steps:
(a) In the first stage, perform the nonlinear static pushover analysis, using the
incremental lateral forces 𝒔𝟏∗ 𝒎𝝓𝟏 , until the displacement increment at the roof
reaches 𝑢 𝛼 𝑢 . Obtain the peak value to the response such as displacements,
storey drifts, and hinge plastic rotations for this stage, the response is noted by 𝑟 ;
(b) Then implement the second stage of analysis using the incremental lateral forces
𝒔𝟐∗ 𝒎𝝓𝟐 until the displacement increment at the roof equals 𝑢 𝛼 𝑢 . It is
noted that the initial condition in this stage pushover analysis is the same as the state
at the last step of analysis in the last stage. Obtain the peak value to the response 𝑟
for this stage;
(c) Repeat step 4 (b) for the rest of the modes considered, with the target displacement
of the 𝑖th step being 𝑢 𝛼𝑢 and the initial condition in this stage pushover
analysis being the same as the state at the last step of analysis in the last stage.
5) Envelope the peak response for each step:

r  max r1 , r2 ,..., rn  (3-11)
The value of r shows the seismic responses estimated by the SPA procedure. A flowchart
showing the procedure of the SPA method is presented in Figure 3.2.
38
Figure 3.2 Flowchart of SPA procedure.
3.3 Case study
3.3.1 Details of ground motions
In this study, three typical ASCE 7-10 design spectra (ASCE 2013), representing different
levels of the motion intensity were selected as the target response spectra. The critical
parameters, Sds, Sd1, and TL, of the target design spectra, are summarised in Table 3.1. The
design acceleration spectra are plotted in Figure 3.3. Three sets of ground motions were
selected and scaled to match the acceleration spectra of the motions with target response spectra
respectively. There were 20 motion records in each set, and all the motion data were obtained
from the strong ground motion database of the Pacific Earthquake Engineering Research
Centre (PEER). The motion selection and scaling method were the minimising MSE method,
which is an abbreviation for minimising the computed weighted mean squared error of record,
and suite average, with regards to the target spectrum. The weight value was 1.0 for periods
of 0.1s, 1s, 4s and 10s to ensure that the response spectra of the scaled selected records fit the
target spectrum well in a broad period range. The chosen strong ground motions have moment
magnitudes ranging from 6.5 to 9.0. The motion records were far-field records, and the
corresponding distance between the recording stations and the epicentre was at least 12 km.
39
The soil type for the site is NEHRP site class C, and no pulse like the record was considered.
Details of the selected records are listed in Tables 3.2-3.4. The pseudo-acceleration spectra are
shown in Figure 3.4.
Table 3.1 Summary of the key parameters of the target spectrum

Spectrum Number Sds (g) Sd1 (g) TL (s)
1 1.00 0.75 12
2 1.20 0.85 10
3 1.52 0.82 12
Record Scale Component

Earthquake Name Year Station Name Magnitude
ID Factor (deg)
Fremont - Mission
1 4.764 Loma Prieta 1989 6.93 0
San Jose
2 4.582 Loma Prieta 1989 Gilroy Array #6 6.93 0
Palo Alto - SLAC
3 2.303 Loma Prieta 1989 6.93 270
Lab
Fortuna - Fortuna
4 3.548 Cape Mendocino 1992 7.01 0
Blvd
5 4.805 Chi-Chi_Taiwan 1999 HWA009 7.62 E
9 4.332 Chi-Chi_Taiwan 1999 HWA055 7.62 N
10 2.098 Chi-Chi_Taiwan 1999 TCU042 7.62 E
13 2.326 Hector Mine 1999 Amboy 7.13 90
14 3.832 Hector Mine 1999 Joshua Tree 7.13 90
Ferndale Fire
15 1.133 Cape Mendocino 1992 7.01 70
Station
"South Bay Union
16 3.021 Cape Mendocino 1992 7.01 270
School"
North Palm Springs
17 3.642 Landers 1992 7.28 90
Fire Sta #36
Tokamachi
18 4.190 Chuetsu-oki_Japan 2007 6.80 NS
Matsunoyama
Sawa Mizuguti
Tokamachi
Darfield_New
20 3.756 2010 WAKC 7.00 E
Zealand
40
ID Factor (deg)
1 4.106 Kern County 1952 Taft Lincoln School 7.36 21
2 3.828 Irpinia_Italy-01 1980 "Calitri 6.9 0
3 4.319 Loma Prieta 1989 APEEL 7 - Pulgas 6.93 0
Coyote Lake Dam -
4 3.975 Loma Prieta 1989 6.93 195
Southwest Abutment
Fortuna - Fortuna
5 4.168 Cape Mendocino 1992 7.01 0
Blvd
LA - Temple &
6 4.620 Northridge-01 1994 6.69 90
Hope
8 4.179 Chi-Chi_Taiwan 1999 HWA051 7.62 N
12 4.536 Hector Mine 1999 Joshua Tree 7.13 90
Ferndale Fire
13 1.404 Cape Mendocino 1992 7.01 270
Station
14 2.127 Cape Mendocino 1992 Loleta Fire Station 7.01 270
South Bay Union
15 3.357 Cape Mendocino 1992 7.01 270
School
North Palm Springs
16 4.136 Landers 1992 7.28 90
Fire Sta #36
Tani Kozima
Nagaoka
Sawa Mizuguti
Tokamachi
19 2.729 Iwate_Japan 2008 Matsuyama City 6.90 NS
20 2.862 Iwate_Japan 2008 Yuzawa Town 6.90 NS
41
ID Factor (deg)
1 4.765 Kern County 1952 Taft Lincoln School 7.36 21
2 4.211 Imperial Valley-06 1979 Cerro Prieto 6.53 147
3 4.565 Irpinia_Italy-01 1980 "Calitri 6.9 0
Anderson Dam
4 3.225 Loma Prieta 1989 6.93 250
(Downstream)
Coyote Lake Dam -
5 4.376 Loma Prieta 1989 6.93 195
Southwest Abutment
Castaic - Old Ridge
6 1.479 Northridge-01 1994 6.69 90
Route
LA - Brentwood VA
7 4.095 Northridge-01 1994 6.69 195
Hospital
8 4.106 Chi-Chi_Taiwan 1999 CHY010 7.62 N
9 4.569 Chi-Chi_Taiwan 1999 CHY046 7.62 E
Ferndale Fire
12 1.749 Cape Mendocino 1992 7.01 270
Station
13 2.675 Cape Mendocino 1992 Loleta Fire Station 7.01 270
South Bay Union
14 4.153 Cape Mendocino 1992 7.01 270
School
Joetsu Yasuzukaku
Yasuzuka
Matsushiro
Tokamachi
Yoshikawaku Joetsu
City
Tokamachi
Chitosecho
Yoitamachi Yoita
Nagaoka
20 3.196 Iwate_Japan 2008 Yuzawa Town 6.90 NS
Figure 3.3 Comparison of the acceleration spectra of the design spectra.
42
(a) (b)
(c)
Figure 3.4 Design spectrum and the response spectrum of the corresponding scaled ground motion: (a)
motion Set 1; (b) motion Set 2; (c) motion Set 3.
3.3.2 Prototype of structures
Frame structures
Two steel moment-resisting frames (SMRFs) with two different heights of 9 and 20 storeys are
used for the case study. The SMRFs, representing typical medium and high-rise buildings
designed for the Los Angeles, California, are designed for the SAC Phase II Steel project and
meet seismic code requirement of the 1994 UBC. The characteristics, natural periods and
modal participating mass ratio are summarised in Table 3.5. Details of the member size,
43
seismic mass and materials of the two frame structures are given by Ohtori et al. (Ohtori et al.
2004) and are summarised in Figure 3.5.
The nonlinear behaviour of the structures is assumed to be concentrated at the discretised

hinges that located at the ends of the frame members. For columns, the hinges are modelled
based on the interaction of axial forces and bending moment, while for beams the hinges are
under the action of bending moment. Figure 3.6 shows a typical generalised force-deformation
relationship used for modelling the hinges. In Figure 3.6, 𝑄 is the generalised force; 𝑄
represents the yield strength of the component; ∆ or 𝜃 is the deformation and, a, b and c are
hinge modelling parameters. The properties of the hinges and modelling parameters are
obtained according to FEMA 365 (FEMA-356 2000).
Table 3.5 Modal properties of the frame structures with a symmetric building plan
Mode
NO. Storey 1 2 3
T (s) 𝜆 T(s) 𝜆 T(s) 𝜆
F1 9 2.268 0.75 0.856 0.1 0.496 0.15
F2 20 3.832 0.77 1.325 0.1 0.767 0.12
(a)
44
(b)
Figure 3.5 Details of the frame structures: (a) 9-storey frame; (b) 20-storey frame.
Figure 3.6 Force-deformation relationship for hinges.
Reinforced concrete wall structures
The RC shear wall structures studied are an eight-storey RC shear wall structure and a twenty-
storey RC shear wall structure that were both modified from the eight-storey RC shear wall
building used in the document FEMA-440/ATC-55 (FEMA 2005). The reinforcement steel
45
used in the model is Grade 60 steel, and the yield strength of the steel is 413.67 MPa (60 ksi),
and the specified compressive strength of concrete is 27.58 MPa (4000 psi). The elevation and
plan views of the two RC shear wall models are shown in Figure 3.7 and Figure 3.8 respectively.
The assumed gravity load, which is a summation of dead load and 25 % live load of the eight-
storey wall structure, is 1468 kN (330 kips) for each floor, while the gravity load for every
storey of the twenty-storey wall structure is 2348.48 kN (528 kips). In all the wall model, only
flexural inelasticity of the shear wall was modelled, and the shear deformation of the wall is
set to be elastic since it was assumed that the wall would have sufficient shear strength. The
modal properties of first three modes of the wall structures presented in Table 3.6.
Table 3.6 Modal properties of the wall structures

Mode
NO. Storey 1 2 3
T (s) 𝜆 T(s) 𝜆 T(s) 𝜆
W1 8 0.625 0.680 0.127 0.200 0.061 0.045
W2 20 1.830 0.632 0.341 0.199 0.145 0.056
3.3.3 Analytical models
To verify the accuracy and applicability of the proposed SPA method, a comparison of the
results from SPA, CMP, MPA as well as the NLRHA are made. In this study, NLRHA and
MPA were carried out for all the selected ground motions, the mean values of the responses
were used for comparison. The numerical implicit Wilson–θ time integration method, in which
the parameter θ was assumed to have a value of 1.4, was adopted to perform the NLRHA
procedure. A 5% damping ratio was set for the first and third mode to build to damping matrix
for analysis. For MPA method, three modes were considered for the F1 and W1, while five
modes were included for F2 and W2. In CMP method, three-stage analysis procedure is used
for both structure, and the target roof displacements for the CMP procedure were obtained from
mean roof displacement of NLRHA. While for the SPA method, the target displacements, as
well as the contribution of different modes, were calculated based on the response spectrum
analysis using the selected ASCE design spectrums. P-∆ effects are included in all the methods,
the nonlinear version of SAP 2000 software (CSI 2011) is used for all the analysis procedures.
46
Figure 3.7 Elevation and plan views of W1.
47
Figure 3.8 Elevation and plan views of W2.
48
3.4 Results of the nonlinear time history analysis method
In most of the seismic codes of practice, it is required that when conducting the nonlinear time
history analysis of buildings (ASCE 2013; CEN 2004; Chinese Standard 2008; IBC 2006; NZS
2004), at least seven ground motions should be analysed, and the mean demand of these
nonlinear time history analysis is treated as the design demands. However, as the frequency
contents of the motion vary significantly among the different ground motions, seismic
responses of buildings under different ground motions are also entirely different. Taking the
mean response quantities of a certain number of ground motions will reduce the variations
considerably and make the results more reliable, while even the mean response quantities still
vary with the different selection of ground motion groups. This can be shown by a comparison
of the mean demands of different motion selections conducted for all the motion sets. For each
ground motion set, five groups of the motion are used for the comparison.
In each group, seven ground motions are randomly selected from the motion set, and the mean
demands of buildings under the seven motions are computed and treated as the demand of the
group. Figure 3.9 to Figure 3.11 show the comparison of the seismically induced deformations
of all the structures under the selected groups for all the motion sets. The mean demands, as
well as the mean demand quantities plus and minus σ of the whole set of motions, are also
plotted as a reference. It is clear from the comparison that there are drastic variations of the
mean responses for different groups of ground motion, and the deviation is more noticeable
with the increase of input motion intensity. Also, despite the significant variations of the mean
responses, it is also noted from Figure 3.9 to Figure 3.11 that almost all the responses from all
the motion groups are in a range of (Mean - σ, Mean + σ) of the whole set. This implies that
the range of (Mean - σ, Mean + σ) is contained in the value range of the responses that are
obtained from the mean responses of any seven nonlinear time history analyses of the ground
motion data from that set. As a result, the range of (Mean - σ, Mean + σ), or one standard
deviation range, is a very good tool to judge the reasonability of the seismic demands estimated
by the simplified methods. If the demands from the simplified methods are within the one
deviation range, the results can be recognised reasonable, since the results located in a value
range of the demands that are obtained from the mean demands of any seven nonlinear time
history analyses of the ground motion data from that set.
49
Storey
Storey
Inter-storey drift ratio (%) Inter-storey drift ratio (%)
Set 1 Set 2
Storey
Inter-storey drift ratio (%)

Set 3
(a)
50
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
Figure 3.9 Inter-storey drift ratio of the frame structures from the nonlinear time history analysis
method: (a) F1; (b) F2.
51
Storey
Storey
Hinge plastic rotation (rad) Hinge plastic rotation (rad)
Set 1 Set 2
Storey
Hinge plastic rotation (rad)

Set 3
(a)
52
Storey
Storey

Set 1 Set 2
Storey

Set 3
(b)
Figure 3.10 Hinge plastic rotation of the frame structures from the nonlinear time history analysis
method: (a) F1; (b) F2.
53
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
54
Storey
Inter-storey drift ratio (%) Storey Inter-storey drift ratio (%)
Set 1 Set 2
Storey
Set 3
(b)
Figure 3.11 Inter-storey drift ratio of the wall structures from the nonlinear time history analysis
method: (a) W1; (b) W2.
55
3.5 Results of the nonlinear time history analysis method and pushover methods
3.5.1 Frame structures
The height-wise variations of the inter-storey drift ratio and hinge plastic rotation of F1 and F2,
calculated by NLRHA, MPA, CMP, and SPA methods, are shown in Figure 3.12 and Figure
3.13, in which the results of MPA and NLRHA in the figures are the mean values. In Figure
3.12 to Figure 3.13, the mean demands plus the standard deviation, σ, of the demand quantities
of the sets, denoted as (Mean + σ), and the mean response minus σ, indicated as (Mean – σ) are
also plotted as a reference.
Figure 3.12 (a) shows the comparison of the drift ratio of F1. From the comparison of the inter-
storey drift ratio of F1 under ground motion Set 1, the MPA method tended to overestimate the
drift ratio at the lower part of the building and to underestimate the drift ratio at the upper floors
of the building. The CMP method overestimated the inter-storey drift ratio at the lower and
upper storeys of the building and underestimated the ratio in the middle part. The SPA method
was prone to underestimate the inter-storey drift ratio along the height of the building, but the
results are still very close to the NLRHA mean value. For the ground motion Set 2, where the
intensity of the input motions is higher, the trend of the estimation from MPA is the same as
that of ground motion Set 1, with the drift ratio which was overestimated at the lower part and
underestimated at the upper portion. The CMP method exaggerated the drift ratio for all the
height levels of the building, especially the upper part, while the SPA procedure gave a very
close estimation of the inter-storey drift ratio along the height of the building. For the ground
motion Set 3, where the input motions have the highest intensity, the drift from all the three
simplified methods have the same trend with that of the ground motion Set 2. It is found that
the drift ratio predictions of SPA are much better than the other two methods.
The comparison of the inter-storey drift ratio of F2 in Figure 3.12 (b) reveals that the trends of
the estimate of the inter-storey drift ratio from the three pushover methods do not change with
the variation of the ground motion sets. The MPA method overestimated the lower part inter-
storey drift ratio and underestimated the upper part inter-storey drift ratio. While for the CMP
method, the inter-storey drift ratios at the upper and lower storeys are overestimated, but those
at the middle part are underestimated. The SPA method tended to overestimate the drift ratio
at the lower portion and had a very close prediction at the upper part. It can also be found from
56
the comparison of inter-storey drift ratio of both F1 and F2 that among the three pushover
methods, the proposed SPA method provided the best estimation of the storey drift ratios, and
all the drift ratios predicted from the SPA method lie in a range of (Mean - σ, Mean + σ).
Figure 3.13 presents the maximum hinge plastic rotations in each storey. It is observed that all
the methods give a reasonable estimation of the hinge plastic rotation, as almost all the
predicted hinge rotations are close to the NLRHA mean values and vary between the range of
(Mean - σ, Mean + σ). From the comparison of the hinge plastic rotation of F1, it is found that
regardless of the ground motion sets, the MPA method tended to overestimate the rotation at
the lower part and underestimate the rotation at the upper portion, while the CMP method
overestimated the hinge plastic rotation along the height of the building. The SPA method
provided a very close prediction of the hinge plastic rotation for all three sets of ground motion.
For the comparison of the hinge plastic rotation of F2, the trend is similar to that of inter-storey
drift ratio of F2. The MPA overestimated the lower levels’ hinge plastic rotation and
underestimated the upper levels’ hinge plastic rotation, while CMP overestimated the inter-
storey drift ratio at upper and lower part. The SPA method overestimated the drift ratio at
lower storeys but predicts the hinge plastic rotation very close to that of the NLRHA mean
value at upper storeys. It can also be seen that the SPA gave the best prediction of the hinge
plastic rotation among all the simplified methods, and almost all the hinge plastic rotation
predicted from the SPA method lied in the one standard deviation range.
57
Storey
Storey
Set 1 Set 2
Storey

Set 3
(a)
58
Storey
Storey

Set 1 Set 2
Storey

Set 3
(b)
Figure 3.12 Comparison of Inter-storey drift ratio results of the frame structures: (a) F1; (b) F2.
59
Storey
Storey
Set 1 Set 2
Storey

Set 3
(a)
60
Storey
Storey

Set 1 Set 2
Storey

Set 3
(b)
Figure 3.13 Comparison of the hinge plastic rotation results of the frame structures: (a) F1; (b) F2.
61
3.5.2 Wall structures
Figures 3.14 presents the comparison of the inter-storey drift ratio of the wall structures,
namely W1 and W2 respectively. In Figures 3.14, the one standard deviation range denoted as
Mean + σ and Mean - σ respectively are also plotted. It is noticeable that both MPA and SPA
method can well predict the trend of the drift ratio along the height of the wall, but the CMP
method failed to predict the distribution of the drift ratio well for both walls and all the motion
intensities. More specifically, the CMP method under- and overestimated the inter-storey drift
ratios at the lower and upper floors of W1 respectively, regardless of the variation of the input
motion intensities. When W1 is subjected to motions from motion Set 1 and Set 2, both the
MPA method and the SPA method overpredicted the drift ratio, but the results from the SPA
is better since the drift ratio is close to that of the NLRHA method. When analysing W1 using
motion set 3, it can be found that the SPA method provided an excellent estimation of the drift
ratio, although the MPA method was still prone to overestimating the drift ratio.
When comparing the drift ratio of W2, which is shown in Figure 3.14 (b), it can be found that
all the advanced pushover analysis methods tended to provide a broader estimation of the inter-
storey drift ratio. The CMP predicted the drift ratio of W2 the worst, especially at upper levels,
where the drift ratio is overpredicted significantly. The MPA and the SPA method can well
estimate the drift ratio regardless the change of motion intensities, but the results from the SPA
method match better with the those of the NLRHA method. Additionally, it is also noticed that
for both wall structures, the drift ratio of the MPA and SPA method are all within the one
standard deviation range, but some drift ratio of the CMP method is beyond the one standard
deviation range, indicating that the CMP method cannot reasonably predict the drift ratio of
wall. As a result of the comparison of the drift ratio of the wall structures, it can be concluded
that the SPA method can predict the inter-storey drift ratio of the walls best and the predictions
are very reasonable.
62
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
63
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
Figure 3.14 Comparison of the inter-storey drift ratio results of the wall structures: (a) W1; (b) W2.
64
3.6 Concluding remarks
In this chapter, a spectrum-based pushover analysis (SPA) procedure is proposed for quick and
effective estimation of the seismic demand of tall buildings with symmetric building plans,
where simplification of the very complicated coupling effect of different vibration modes in
the nonlinear vibration of the structure under the seismic action is introduced. The
simplification of coupling of mode is that that when the 𝒊th mode of vibration starts, the roof
displacement of all the previous modes (the 𝑖 1 th, 𝑖 2 th, …, and 1st mode) of vibration
have already reached the target value and the structural states like roof displacement, inter-
storey drift, internal stress and strain of 𝑖th mode of vibration keep unchanged after the roof
displacement of 𝒊th mode of vibration reaches the target value. The proposed SPA method
includes several modal pushover procedures that are performed consecutively. The target roof
displacements and contributions of different vibration modes are calculated from the response
spectrum analysis. The final seismic demand is obtained by enveloping the demands of each
pushover process. A comparative study is conducted to verify the applicability and accuracy
of the SPA procedure on analysing frame and wall structures with different heights. The
variation of ground motion intensities is also included in the case study for a better and more
comprehensive study of the applicability of the pushover analysis methods. The following
conclusions can be drawn.
1. All the simplified methods, including the MPA method, the CMP method, and the
proposed SPA method, can provide a reasonably good estimate of the seismic demand of
frame and wall structures with a symmetric building plan, regardless the change of the
input ground motion intensities. The majority of the results of these methods show good
agreement with the mean response obtained from NLRHA.
2. Among all the simplified methods, the proposed SPA method provides the best
predictions of both the inter-storey drift ratio and the hinge plastic rotation of frames and
drift ratio of walls. Most of the seismic responses predicted by SPA are located between
the mean responses plus and minus the standard deviation of the responses of the sets,
i.e. one standard deviation range. Thus, the accuracy and applicability of the SPA method
are verified.
65
3. Owing to its efficiency, accuracy and the design spectrum-based calculation procedure,
the proposed SPA method is considered as one of the most promising tools for quick
seismic assessment of frame and wall structures with a symmetric building plan.
66
Chapter 4
Modified Spectrum-based Pushover
Analysis for Estimating Seismic
Deformation of Dual Wall-frame
Structures
4.1 Introduction
This chapter presents a modified spectrum-based pushover analysis (MSPA) method for quick
evaluation of seismic demand of dual wall-frame structural systems. The SPA method is
modified where the forces vectors are changed to be used for pushover analysis, and
formulations for determining the roof displacement to consider the wall-frame interaction and
unique damage modes of the wall-frame structure. The applicability and accuracy of the
proposed MSPA in predicting the seismic demand of dual wall-frame structures are
investigated through a case study of four 25-storey reinforced concrete wall-frame structures,
which are subjected to different levels of input ground motion. It is shown from the comparison
of the results from nonlinear response time history analysis (NLRHA), the advanced pushover-
based models including MPA and CMP methods, and the proposed MSPA that only MSPA
can predict the trend of the distribution of seismic demand along the height of buildings well,
and the anticipated seismic demand from MSPA shows very good agreement with that from
NLRHA.
68
4.2 Modified spectrum-based pushover analysis method
In the SPA method, the consecutive pushover technique, which is when one pushover
procedure is completed, the next pushover procedure will be started with the initial structural
state, including internal stress, strain, and displacement, the same as that in end of the
immediate past pushover procedure, is used to reproduce structural states of building structures.
Forces with mode shape distribution (𝒎𝛟𝒊 ) are used to conduct the pushover analysis, and the
order of the consecutive pushover procedure should follow the order of modes, which is from
the 1st mode to higher modes. The force distributions for the first three pushover analysis
stages in SPA for F2 are presented in Figure 4.1(a).
First mode Second mode Third mode

(a) Frame structure
First mode Second mode Third mode

(b) Dual wall-frame structure
Figure 4.1 Modal shape distributed forces used in SPA method.
69
The location of the plastic hinges initiated in the first pushover stage for the frame structure is
shown in Figure 4.2(a), where the forces with a first mode shape distribution, 𝒎𝛟𝟏 , were used.
It is seen that the plastic hinges caused by first-mode shape distributed forces are mainly
concentrated in the lower part of the frame structure. With the increase in the roof displacement,
the plastic deformation at the bottom will become much larger than other parts. After
experiencing larger plastic deformation, the frame members at the bottom should take smaller
forces in the second stage of pushover analysis. Hence, when the first stage of the pushover
analysis is completed and the force with the second-mode shape distribution, 𝒎𝛟𝟐 , is applied,
the external force at the bottom of the structure decreases, while the external force in the upper
part keeps increasing and plastic hinges are progressively forming from the lower part of the
frame to higher levels. Similarly, after the second stage pushover is completed and the force
of the third stage with a distribution of 𝒎𝛟𝟑 is applied to the structure, the force at the levels
where the plastic hinges are formed and the frame members have very large plastic deformation
in the second stage reverses; while the force on those levels where the frame members
experience much lighter plastic deformations increased, as compared with that when the
structure is subjected to forces with second mode distributions. As a result, plastic hinges are
progressively formed, and plastic deformation is gradually increased from the lower part of the
structure to the top through the SPA procedure. From the result of NLRHA, it is found that
plastic hinges on the frame may most likely occur in the lower storeys first and then
progressively propagate to the upper storeys.
There are different definitions of dual wall-frame structures in different seismic design codes
of practice. In Eurocode 8 (CEN 2004), the dual wall-frame structure is a structure where the
rigid frame takes 35%-65% of the base shear; and in ASCE 7-10 (ASCE 2013), it is required
that the rigid frame should take at least 25% of the seismic forces, while in Chinese seismic
design code (Chinese Standard 2008) the frame should take 20% of the fraction of the base
shear. Despite different requirements for the base shear forces carried by the frame and the
wall, the percentage of the base shear taken by both of them is similar in magnitude, but the
interaction force between wall and frame structure is significant. The interaction force between
the wall and the frame will change the damage mode of the dual structure.
Figure 4.1(b) shows the distribution of forces in the first three pushover procedures when
adopting SPA method to estimate the seismic demand of a dual-wall frame structure, whose
70
details will be presented later. It can be seen from Figure 4.1(b) that although the seismic
behaviour of the dual wall-frame structure and frame structure is different, they have similar
vibration mode shapes. Thus, the force vectors used in the analysis of frame structures and
wall-frame structures should have a similar distribution.
(a) Frame structure (b) Wall-frame structure

Figure 4.2 Distribution of plastic hinges initiated under the first-mode shape forces.
Figure 4.2 (b) shows the location of the plastic hinges of a wall-frame structure under the force
with first-mode shape distribution, 𝒎𝛟𝟏 . It can be seen from Figure 4.2(b) that plastic hinges
occur at upper storeys first, and the plastic deformation of frame members at the top is the
largest. This indicates that the wall-frame interaction changes the damage mode of the structure
as compared with the bare frame structure. Since the frame takes more than 20% of the base
shear and plays an important role in the lateral-load resistance system, the yielding of the frame
will affect the overall structural performance considerably. The force vectors used for
subsequent pushover stages in SPA need to be changed to account for the change of the damage
mode.
Because the hinges are mainly initiated in the upper storeys, and the plastic deformations in the
upper storeys are more significant, the force vector distributions for the second pushover stage
71
should be changed from 𝒎𝛟𝟐 to 𝒎𝛟𝟐 , indicating that the forces of the second pushover
stage have a reverse distribution as compared with that plotted in Figure 4.1(b). Otherwise, the
deformation in the upper and lower storeys will be significantly overestimated and
underestimated, respectively. Since there is a significant wall-frame interaction, and higher
modes have much less contribution to the overall roof displacement compared with the first
two modes, both force directions (𝒎𝛟𝒏 and 𝒎𝛟𝒏 ) should be considered. Additionally, a set
of uniformly distributed forces should also be used to consider the case that the hinges may be
progressively formed from the bottom to the top, which may also occur when the frame takes
more seismic design forces. Therefore, for a better and more reasonable estimate of seismic
demand of the dual wall-frame structure, modifications need to be made to the SPA method,
where the force vectors are changed for pushover analysis and a uniformly distributed load
case to consider the different damage modes of the wall-frame structures should be included
(Liu, Kuang, and Huang 2018).
In the modified spectrum-based pushover analysis (MSPA), the first three vibration modes are
considered. The target roof displacement for the ith mode in the consecutive pushover
procedure can be expressed as
uir  i  ur 0 (4-1)
where αi is the friction of the roof displacement contributed by of ith mode; ur0 is the
estimated total roof displacement of the structure. The contribution factor of ith mode to the
roof displacement is expressed as
uri 0  i i ri Di
i  N
 N
(4-2)
u
i 1
ri 0 
i 1
i i ri Di
In which i is the reduction factor of the roof displacement and can be calculated by
    
cos(1  t N 1 )  cos 1    , i  1
i     2  N 1  (4-3)

1.0, i  2,3,..., N
72
where 1 is the natural frequency of the first mode, and N1 is the natural frequency of (N-1)th
mode. tN-1 is the time instant when (N-1)th mode starts considering that all the modes have the
maximum roof displacements at the same time instant.
To determine the roof displacement of the dual structures, a simple estimate of the total roof
displacement was proposed by Miranda et al. (Miranda and Reyes 2002),
ur 0  1D1 (4-4)
where 𝐷 is the displacement spectrum value corresponding to the fundamental period of the
structure. However, using Equation (4-4), the inelastic roof displacement of the structures is
generally overestimated by about 30%. Hence, a modified expression is proposed for a better
estimate of the total roof displacement,
1 D1
ur 0  (4-5)
1.3
Steps of the modified spectrum-based pushover analysis can be summarised as follows:
1. Apply seismic gravity loads to the structure and perform eigenvalue analysis to obtain the
dynamic elastic properties of the structure, including the natural frequencies, 𝜔 , 𝛽 and
the mode-shapes, 𝝓𝒏 . The mode-shapes are normalised so that the roof component of 𝜙
is equal to unity (𝜙 1).
2. Compute the force vector of different modes: 𝒔𝒏 𝒎𝝓𝒏 , where 𝐬𝒏 is the distribution of
incremental lateral force over the height of the structure for the nth mode.
3. Determine the target roof displacements of the structure for a different mode 𝑢 , and the
total roof displacement 𝑢 , using Equations (4-1) and (4-5) respectively, where the
spectrum displacement value, 𝐷 , is obtained from the elastic design spectrum of the site
of interest.
73
4. Carry out the consecutive nonlinear pushover process to the structures under the seismic
gravity load according to the following sub-steps:
(a) Conduct the pushover analysis using the incremental lateral force 𝒔𝟏∗ 𝒎𝝓𝟏 , until the
displacement at the roof reaches 𝑢 𝛼 𝑢 . Calculate the peak values to the
response, such as displacement, storey drift, and hinge plastic rotation for this stage,
where the response is noted by 𝑟 .
(b) Implement pushover analysis using the lateral force 𝒔𝟐∗ 𝒎𝝓𝟐 , until the
displacements at the roof equals 𝑢 𝛼 𝑢 . It is noted that the initial condition at
this stage pushover analysis is the same as the state at the last step of analysis in Step
4(a). Obtain the peak values to the response 𝑟 .
(c) With the initial structural state the same with that at the end of Step 4(b), perform
pushover analysis with lateral force 𝒔𝟑∗ 𝒎𝝓𝟑 to make the roof displacement
reaches 𝑢 𝛼 𝑢 , and obtain the peak response 𝑟 .
(d) Perform pushover analysis with lateral force 𝒔∗𝟑 𝒎𝝓𝟑 with the initial structural
condition the same with that at the end of Step 4(b). This stage shall stop when
displacement reaches 𝑢 𝛼 𝑢 , and obtain the peak response 𝑟 .
5. Perform pushover analysis to buildings with seismic gravity load applied using the
uniformly distributed forces until the roof displacement equals to 𝑢 and get the peak
response quantities 𝑟 .
6. Envelope the peak response in all steps above,
r  max r1 , r2 , r31 , r32 , rU  (4-6)
The value of r shows the seismic responses estimated by the MSPA procedure. A flowchart
summarizing the procedure of the MSPA method is presented in Figure 4.3.
74
Figure 4.3 Flowchart of MSPA procedure.
4.3 Case study
4.3.1 Ground motions selection
In this study, the three sets of motions, namely, Set 1, Set 2 and Set 3, whose details are
presented in section 3.3.1 were selected, and nonlinear response time history analysis procedure
is conducted for all the chosen motions and pushover analysis methods performed for all the
selection motion sets.
Four reinforced concrete (RC) wall-frame structures (Huang 2009) are modified and used in
this case study. The wall-frame structures are 25-storey tube-in-tube reinforced concrete wall-
frame structures with storey height of 3.5 m. The floor plan is shown in Figure 4.4. Dimensions
of the wall and frame vary to represent different combinations of wall and frame. Details of
geometric properties of the RC shear wall and the rigid frame are presented in Table 4.1. The
materials used for the concrete is C40, with a specified compressive strength of 26.8 MPa, and
for the steel bars, HRB 500 steel was used, and the yield strength of the steel bar was 500 MPa.
75
As for the design of the wall-frame structures, the reinforcement ratio of the wall structure is
modified from and tabulated in Table 4.2, while the reinforcement of the frame system is
designed to ACI 318-14 (ACI (American Concrete Institute) 2014) by the software SAP 2000
and the design load is shown in Table 4.3. The seismic mass is assumed to be the summation
of dead load and 50% live load.
Since the structure is symmetric, the three-dimensional system can be modelled as a two-
dimensional one. The two-dimensional model of the structure is presented in Figure 4.5. The
modal properties, including the first three natural periods and the corresponding modal mass
participation factor of the three wall-frame systems, were summarised in Table 4.4. Linear
elastic pushover analysis, where the building was subjected to force with first mode
distributions, was conducted, and the fraction of the elastic seismic base shear was also
tabulated in Table 4.4. According to EC8, WF2, WF3 and WF4 can be classified as a wall
system, wall-equivalent dual system, and frame-equivalent dual system, respectively; while
considering the definition of a wall-frame dual system in the ASCE 7-10, WF1 and WF2 can
also be classified as dual wall-frame structures.
For the nonlinear property of the structures, discretised hinges located at the ends of frame
members are used to model the nonlinear behaviour of the frame structures. For columns, the
interaction of axial force and bending moment is considered when modelling the hinges, while
for hinges at beams, only bending moments are considered. The properties of hinges and
modelling parameters are calculated following the FEMA 365. For the wall model, only the
flexural inelasticity of shear wall is modelled, while the shear deformation of the wall is set to
be elastic since the wall is assumed to have sufficient shear strength.
procedure.
76
Figure 4.4 Plan view of the dual wall-frame structure.
Figure 4.5 Elevation of the 2-D dual wall-frame structure.
77
A 5% damping ratio was expected for the first and third mode to build to damping matrix for
analysis. For MPA method, five modes were included. In CMP method, three-stage analysis
procedure is used for both structure, and the target roof displacements for the CMP procedure
were obtained from mean roof displacement of NLRHA. While for the SPA method, the target
displacements, as well as the contribution of different modes, were calculated based on the
response spectrum analysis using the selected ASCE design spectrums. P-∆ effects were
included in all the methods, the nonlinear version of SAP 2000 software was used for all the
analysis procedures.
Table 4.1 Dimension of the main structural element for wall-frame structures
Structure NO. WF1 WF2 WF3 WF4
Length 6,000 5,000 5,000 5,000
(LC: mm)
Wall
Thickness 350 350 350 250
(b: mm)
Column (b × h mm) 1,0001,000 1,0001,000 1,2001,200 1,2001,200
Beam (b × h mm) 400900 400900 5001,000 6001,100
Table 4.2 Reinforcement details of the wall of wall frame structures.

Structure NO. floor 1-2 3-4 5-6 7-8 9-25
2
Area (mm /m) 11000 7000 4600 2400 1700
WF1
Reinforcement ratio (%) 3.143 2 1.314 0.686 0.486
2
Area (mm /m) 11000 7000 4600 2400 1700
WF2
2
Area (mm /m) 11000 7000 4600 2400 1700
WF3
2
WF4 Area (mm /m) 7800 5000 3280 1720 1220
Table 4.3 Design load for wall-frame structures.

Loading Value
Gravity Load Dead Load 5 kN/𝑚
Live Load 3 kN/𝑚
Wind Load 300 kN/storey
78
Table 4.4 Modal property of dual structures and fraction of base shear taken by wall and
frame
Structure mode Natural period (s) Modal mass seismic base shear seismic base shear
participation ratio taken by frame (%) taken by wall (%)
WF1 1 2.85 0.73 27 73
2 0.84 0.12
3 0.41 0.05
WF2 1 2.96 0.75 33 67
2 0.90 0.11
3 0.46 0.04
WF3 1 2.31 0.75 46 54
2 0.72 0.11
3 0.38 0.04
WF4 1 2.00 0.76 55 45
2 0.63 0.11
3 0.34 0.04
In Section 3.4, It is seen that the earthquake-induced deformations of wall and frame structures
with symmetric building plans, computed by the NLRHA procedure following the requirement
of codes of practice vary within the mean demands plus and minus one standard deviation (σ)
of the whole set of motions, or just one standard deviation range. Earthquake-induced
deformation of the dual wall-frame structures that is computed following the codes of practice
is also presented for all the motion sets. The grouping of the motions is the same as that in
Section 3.4. Figures 4.6 and 4.7 show the height-wise distribution of the inter-storey drift ratio
and hinge plastic rotation of the dual wall-frame structures of the selected groups, where the
mean deformation, and the deformation plus and minus σ of the whole motion set, are plotted.
It can be seen that the seismically induced deformation of the structures varies significantly
under different selections of ground motions and the difference is more noticeable when the
intensity of ground motion is higher. It is also seen that the earthquake-induced deformation
of different groups from the same set is within the one standard deviation range. This indicates
that the one standard deviation is part of the value range of the deformation of the dual wall-
frame structures computed following the NLRHA procedure of codes. Thus, one standard
79
deviation range is used as a benchmark to judge the reasonability of the seismically induced
deformation of dual wall-frame structures computed by the pushover analysis methods.
Storey
Storey

Set 1 Set 2
Storey

Set 3
(a)
80
Storey
Set 1 Set 2
Storey

Set 3
(b)
81
Storey
Set 1 Set 2
Storey

Set 3
(c)
82
Storey
Set 1 Set 2
Storey
Set 3
(d)
Figure 4.6 Inter-storey drift ratio of the dual wall frame structures from the nonlinear time history
analysis method: (a) WF1; (b) WF2; (c) WF3; (d) WF4.
83
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
84
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
85
Storey
Hinge plastic rotation (rad) Storey Hinge plastic rotation (rad)
Set 1 Set 2
Storey

Set 3
(c)
86
Storey
Set 1 Set 2
Storey
Set 3
(d)
Figure 4.7 Hinge plastic rotation of the dual wall frame from the nonlinear time history analysis
method: (a) WF1; (b) WF2; (c) WF3; (d) WF4.
87
Figures 4.8 to 4.9 present the comparisons of variation of the peak inter-storey drift ratio and
peak plastic rotation of frame members at each floor, computed by NLRHA, MPA, CMP and
the proposed MSPA method. The mean peak demand quantities plus and minus one standard
deviation of the peak demand quantities of all motion sets calculated from the NLRHA method
are also plotted and denoted as (Mean + σ) and (Mean – σ) respectively. It can be seen from
Figure 4.8 and Figure 4.9 that the mean inter-storey drift ratio and the mean peak hinge plastic
rotation of NLRHA share the same trend, regardless of the change of structural combination
and ground motion intensity. The seismic demand increases rapidly at 1st to 4th floors and
then grows slowly until the floor level reaches the 15th floor. After that, the seismic
deformation decreases gradually as the floor level increases.
From the comparison of the inter-storey drift ratio shown in Figures 4.8, it is seen that only the
proposed MSPA method provided a similar prediction of the drift-ratio distribution over the
height of the structure, while CMP and MPA did not capture the change of the drift ratio along
the structural height. More specifically, for all input motion levels, the drift ratios of WF1
predicted by MPA and CMP are smaller and more extensive than that NLRHA in the lower
and upper floors respectively, while the results from CMP agree with those from NLRHA better.
It is seen that although the MSPA overestimated slightly the drift ratio over the height of the
building for motion Set 3, the results from MSPA show the best agreement with those
computed by NLRHA for all the motion intensities.
For the wall-frame WF2, whose inter-storey drift ratio was plotted in Figure 4.8 (b), all three
simplified methods overestimated the inter-storey drift ratios in the upper part of the structure
but underestimated them in the lower part. Among the results from the three pushover analysis
methods, the drift ratio of MSPA shows the best agreement with those of NLRHA over the
height of the structure. For WF3, it is seen from Figure 4.8 (c) that the drift ratios from CMP
and MPA are larger than those of NLRHA in the upper part, while at lower storeys, the drift
ratios of both methods are smaller than those of NLRHA. Even though the inter-storey drift
ratios predicted by MSPA are larger than those from NLRHA along the height of the structure
for motion Set 2, they show the best agreement with those predicted by NLRHA.
88
When comparing the drift ratios for WF4, as shown in Figure 4.8 (d), it is still the case that
MPA significantly overestimated and underestimated the drift ratios in the upper and lower
parts of the building, respectively, while both CMP and the proposed MSPA provided a very
good prediction of the drift ratio at lower storeys. The CMP method did not predict the drift
ratio at upper storeys reasonably well, but the results of MSPA still agree well with those of
NLRHA. Also, it can be seen from Figure 4.8 that most of the predictions of MPA is not within
the one standard deviation range and most of the upper part drift ratio from CMP is not located
in the one standard deviation range. Only the drift ratio predicted by the MSPA method is in
the one standard deviation range. It is indicated that the proposed MSPA method estimated the
drift ratio very well and reasonable.
The peak hinge plastic rotations at each level along the height of the buildings are shown in
Figure 4.9. Only the predicted distribution of the hinge plastic rotation by the proposed MSPA
method has a similar trend to that predicted by NLRHA. It is also seen that MSPA provides
the best prediction of the hinge plastic rotations all along the heights of buildings for WF2 and
WF3. As for WF 1, it is the CMP and MSPA methods that predicted the plastic rotation best
at lower and upper storeys respectively. While for WF4, hinge plastic rotations predicted by
CMP and MSPA are similar in lower part of the building, but only the MSPA method can
predict the hinge plastic rotations well in the upper part, as compared to those from NLRHA.
Moreover, only MSPA can provide almost all the predictions of hinge plastic rotation within
the one standard deviation range, indicating that the MSPA method can estimate the hinge
plastic rotations very well and reasonable as compared with those of NLRHA.
Among the three pushover methods, MPA cannot predict the drift and hinge plastic rotation
very well, and the trend of the variation of the response quantities along the height of the
structure deviates that given by NLRHA method, while the trend of the distribution of the
seismic response quantities predicted by CMP is much better than that of MPA. From the
comparison of the results of CMP and MSPA, in which the dynamic coupling of modes of
vibration and unique damage modes of the wall frame structure, it is seen that MSPA can
predict the seismic demand much better than CMP. Since the procedure of the roof
displacement prediction is simple and spectrum-based, and the basis of the analysis method of
MSPA is nonlinear static analysis, the computation procedure of the MSPA method is more
efficient and easy to carry out.
89
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
90
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
91
Storey
Storey
Set 1 Set 2
Storey
Inter-storey drift
Inter-storey drift ratio
ratio (%)
(%)
Set 3
(c)
92
Storey
Storey
Set 1 Set 2
Storey

Set 3
(d)
Figure 4.8 Comparison of the inter-storey drift ratio results of the dual wall frame structures: (a) WF1;
(b) WF2; (c) WF 3; (d) WF 4.
93
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
94
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
95
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
96
Storey
Storey
Set 1 Set 2
Storey

Set 3
(d)
Figure 4.9 Comparison of the hinge plastic rotation results of the dual wall frame structures: (a) WF1;
(b) WF2; (c) WF 3; (d) WF 4.
97
In this chapter, the spectrum-based pushover analysis (SPA) is modified for estimating the
seismic deformation of dual wall-frame structures. By changing the modal-shape distributed
forces in SPA for analysis for predicting the seismic demand of dual wall-frame structures, the
dynamic interaction of wall and frame and the damage mode of the structure can be considered
in the pushover analysis. A case study of four 25-storey wall-frame structures with different
wall and frame combinations under different input intensities was conducted. From the
comparison of the results from the NLRHA method, MPA method, CMP method, and the
proposed MSPA method, the following conclusions can be drawn.
1. Since there is a strong structural interaction between wall and frame, damage mode of a
frame structure is different from that of a dual wall-frame structure. The applicability of
a pushover-based method, which is initially developed for frame structures, for
estimating the seismic demand of dual wall-frame structures should be verified.
2. All three pushover methods of analysis, including the MPA, the CMP, and the proposed
MSPA method can provide a reasonably good prediction of the seismic demand.
However, since there is no consideration of the different damage modes of wall-frame
structures in MPA and CMP, both the methods cannot capture well the trend of the
distribution of the seismic demand along the height of the dual system.
3. Among the three pushover methods, only the proposed MSPA method can well predict
the trend of the deformation distribution along the height of the structure. Also, the
seismic demand estimated by MSPA shows the best agreement with that predicted by
nonlinear response time history analysis (NLRHA). Almost all the results from MSPA
are within one standard deviation range. It is indicated that that MSPA can predict the
seismic demand of dual wall-frame structures very well and the results are also shown
very reasonable.
4. Since the proposed MPSA method is spectrum-based and has simple computation process,
it is considered as a quick and efficient, yet accurate, pushover-based tool of analysing
the seismic demand of dual wall-frame structures in tall buildings.
98
Chapter 5
Spectrum-based Pushover Analysis for
Estimating Seismic Deformation of
Frame Structures with Singly
symmetric Plan
5.1 Introduction
The primary objective of this chapter is to develop a spectrum-based pushover analysis (SPA)
procedure for quickly predicting the seismic demand of singly symmetric buildings. The
simplification of the mode coupling, which is proved to be effective in calculating the seismic
demand of buildings with a symmetric plan is adopted in the SPA procedure. Two series of
steel moment-resisting frames (SMRFs), which have various heights and mass eccentricities
and were subjected to different input motion levels, were studied to verify the feasibility of the
SPA method in estimating the seismic demand of one-way asymmetric-plan buildings. By
comparing the seismic demands predicted by different pushover analysis methods and the
nonlinear response time history analysis (NLRHA), it is found that the SPA method is capable
of predicting the seismic demand of one-way asymmetric buildings efficiently and accurately,
especially on the heavy side of the building, where the seismic demand is more significant.
100
5.2 Spectrum-based pushover analysis method for buildings with singly
symmetric building plan
Consider a one-way asymmetric building with a symmetric plan along the x-axis and
asymmetric plan along the y-axis. When the structure is subjected to horizontal ground motion
excitations in the y-direction, ugy (t ) , the equation of motion can be expressed by
 y   k yy
 m 0  u k yθ  u y   m 0  1  m1
 0 I  u +    =     ugy (t )     ugy (t )   sugy (t ) (5-1)
 
0 θ   k θy k θθ   u θ   0 I 0  0   0 
where m is the diagonal mass matrix, with 𝒎𝒋𝒋 𝒎𝒋 , where 𝒎𝒋 is the lumped mass on the jth
floor; I0 is the diagonal polar moment of inertia matrix with 𝑰𝒋𝒋 𝑰𝑶𝒋 , where 𝑰𝑶𝒋 is the polar
moment of inertia of jth floor diaphragm about the vertical axis through the mass centre (CM);
 are are the y-lateral and torsional floor accelerations vectors, respectively; u y and
 y and u
u
u are the displacement vectors, whose elements are the y-lateral and torsional floor
displacements, respectively; kyy, kyθ, kθy and kθθ are sub-matrices of the overall stiffness matrix
of the structure; sugy (t ) is the effective earthquake forces; s is the height-wise distribution of
the effective force, and can be expressed as the sum of the modal inertia force, sn, given by
m1 2 N 2N
 m 
s      sn   n  2 yn  (5-2)
 0  n 1 n 1 r m n 
where  yn and  n are the lateral and rotational components of the nth natural vibration mode
n respectively, and the modal participating factor, Γn, can be calculated by
Ln
n  (5-3)
Mn
where
m1
 
N
Ln  yn
T
Tn    ynT m1   m j jyn (5-4)
0 j 1
101
and
m 0   yn 
M n   yn Tn  
N N
T
    T
m    T
I 
n 0 n   m 2
  I 0 j j2 n (5-5)
0   n 
yn yn jyn
 0 I j 1 j 1
Since all the modes of vibration are independent in the elastic vibration of structures, the floor
displacements of nth mode can be expressed as
un  n qn (t )  n n Dn (t ) (5-6)
Substituting Equation (5-6) into Equation (5-1), the equation of motion for a multi-degree-of-
freedom system is converted into equations of motion of a series single-degree of freedom
system, which can be solved independently. When the structure yields, the lateral force is
dependent on the loading history, which can be expressed by
Fs  Fs (u, t ) (5-7)
Thus, the governing equation of motion for the nonlinear system can be written as
  cu  fs (u,signu )  miug (t)

mu (5-8)
In the problem of one-way asymmetric structures, the first five vibration modes that have a
non-zero effective modal participating mass ratio in the direction of the input motions are
included. The effective modal participating mass ratio is computed by
meff 1 L2n
n  N
 N
(5-9)
m m
Mn
j j
i 1 i 1
where ∑𝐍𝐢 𝟏 𝐦𝐣 is the total mass of the building.
102
The overall roof displacement in the direction of the input motion u yr 0 and the target roof
displacement in the y-direction u yir for each pushover analysis can be obtained as
0.5
 N 2
u yr 0     i i  yir Di   (5-10)
 i 1 
where
u yir   yi  u yr 0 (5-11)
and Di is the spectrum displacement values corresponding to the period of ith mode; β is the
roof displacement reduction factor; α stands for the contribution factor of the roof
displacement of the ith mode in the y-direction. The reduction factor β can be calculated by
 1.0 (i  2,3,..., N ) 
 
i       
cos(1  tN 1 )  cos 1  2    (i  1)  (5-12)
   N 1   
where ω and 𝜔 are the vibration frequency of the first mode and (N-1)th mode
respectively; tN-1 is the time point where (N-1)th mode should start to ensure that all the modes
have the target roof displacements at the same time. The ith mode roof displacement
contribution factor can be computed by
i i  yir Di
 yi  N (5-13)

i 1
i i  yir Di
Detailed steps of the SPA procedure for one-way asymmetric structures are summarised as
follows.
1. Apply the seismic gravity loads to the structure.
103
2. Carry out eigenvalue analysis:
a. Calculate the dynamic properties of the structure.
b. Mode-shapes of the included modes are normalised to ensure that the component at
roof ϕyir is equal to 1.
3. Compute the total roof displacement uyr0 and the target roof displacements of different
modes uyir. The displacements are computed using Equations (5-10) and (5-11), where
the spectrum displacement values can be obtained from the elastic design spectrum of
the studied sites or the mean response spectrum of a set of ground motion data.
4. Conduct the spectrum-based pushover analysis:

a. Perform the pushover analysis with lateral force with the distribution of 𝒔𝟏∗ 𝒎𝝓𝟏 ,
and the structure is loaded until the roof displacement reaches uy1r; then obtain the
peak value of the demand r1.
b. With the initial structural condition which is the same as that at the last step of the
previous analysis procedure, carry out pushover analysis using the lateral force with
the distribution of 𝒔𝟐∗ 𝒎𝝓𝟐 , until the displacement increment at the roof is equal
to uy2r. Then obtain the peak value of the demand r2.
c. Repeat Step 4b for all the modes considered with the corresponding force
distribution 𝒔𝒊∗ 𝒎𝝓𝒊 , where the ith analysis procedure shall stop when the roof
displacement of that procedure reaches the target displacement uyir; then obtain the
peak demand ri.
5. Get the maximum demand of the peak demands of all the modes considered as the
demands estimated by the SPA method,
r  max r1 , r2 ,..., rn  (5-14)
104
5.3 Case study
5.3.1 Ground motions selection
In this study, the three sets of motions, Set 1, Set 2, Set 3, whose details are presented in section
3.3.1 were selected, and nonlinear response time history analysis procedure is conducted for
all the chosen motions and pushover analysis methods performed for all the selection motion
sets.
Two series of the steel moment-resisting frame structures with different heights of 9 and 20
storeys were studied. The SMRFs were modified from the frame structures that were initially
designed for the phase II of SAC project. Details of the structure members are shown in Figure
5.1. To involve the torsional effects, the buildings with symmetric plan arrangements were
modified by keeping the stiffness property of the structure unchanged and shifting the centre
of mass (CM) of each storey along the x-direction. In each frame series, there were three frames,
and the eccentricity between CM and centre of stiffness (CS) for the three frames is set to be
10%, 20% and 30% of the plan dimension respectively. The moment of inertia of floors to
mass ratio of storeys of the one-way asymmetric structures was assumed to be equal to that of
the structures with a symmetric plan. The modified plan arrangements of the 3D frame
structures are plotted in Figure 5.2. The modal properties of the six frames are summarised in
Table 5.1. Discretised hinges located at the ends of frame members, which are modelled
following FEMA 365, are used to model the nonlinear behaviour.
105
(a)
(b)
Figure 5.1 Structural information of the frame structures with an asymmetric plan: (a) 9-storey frame;
(b) 20-storey frame
106
Figure 5.2 Plan arrangements of frame structures.
Table 5.1 Dynamic properties of frames used in the analysis

Frame Storey Eccentricity Period (s)
T1 T2 T3 T4 T5
F1-10 9 10% 2.292 1.796 0.853 0.674 0.489
F1-20 9 20% 2.446 1.686 0.911 0.632 0.524
F1-30 9 30% 2.635 1.570 0.982 0.588 0.564
F2-10 20 10% 3.394 2.589 1.178 0.920 0.693
F2-20 20 20% 3.599 2.443 1.255 0.865 0.741
F2-30 20 30% 3.860 2.284 1.349 0.808 0.798
procedure. A 5% damping ratio was expected for the first and third mode to build to damping
matrix for analysis. For MPA method, three modes are considered for the 9-storey frame
structure, and five modes were included for the study of the 20-storey frame structure. In CMP
method, three-stage analysis procedure is used for both structure, and the target roof
107
displacements for the CMP procedure were obtained from mean roof displacement of NLRHA.
While for the SPA method, the target displacements, as well as the contribution of different
modes, were calculated based on the response spectrum analysis using the selected ASCE
design spectrums. P-∆ effects were included in all the methods, the nonlinear version of SAP
2000 software was used for all the analysis procedures.
In Section 3.4, It is seen that the earthquake-induced deformations of wall and frame structures
with symmetric building plans, computed by the NLRHA procedure following the requirement
of codes of practice vary within the mean demands plus and minus one standard deviation (σ)
of the whole set of motions, or just one standard deviation range. Earthquake-induced
deformation of the frame structures with singly symmetric building plans that is computed
following the codes of practice is also presented for all the motion sets. The grouping of the
motions is the same as that in Section 3.4. For the structures with an asymmetric plan, two
sides of the frames will undergo different deformations. The side that closer to CM will take
more gravity loads than the side that is far from CM. In this study, the side that is close to CM
is denoted as the heave side (HS) and the other as the light side (LS). The relations between
CM, LS and HS are given Figure 5.2. The peak inter-storey drift ratios and hinge plastic
rotations on both LS and HS of all the buildings, which are subjected to different levels of input
motions, are presented in Figures 5.3 - 5.10. It can be seen that the seismically induced
deformation of the structures varies significantly under different selections of ground motions
and the difference is more noticeable when the intensity of ground motion is higher. It is also
seen that the earthquake-induced deformation of different groups from the same set is within
the one standard deviation range. This indicates that the one standard deviation is part of the
value range of the deformation of the frame structures with singly symmetric building plan
computed following the NLRHA procedure of codes. Thus, one standard deviation range is
used as a benchmark to judge the reasonability of the seismically induced deformation of frame
structures with singly symmetric building plan computed by the pushover analysis methods.
108
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
109
Storey

Set 1 Set 2
Storey

Set 3
(b)
110
Storey
Storey

Set 1 Set 2
Storey

Set 3
(c)
Figure 5.3 Inter-storey drift ratio of LS of the 9-storey frame structures with a singly symmetric plan
from the nonlinear time history analysis method: (a) F1-10; (b) F1-20; (c) F1-30.
111
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
112
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
113
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.4 Inter-storey drift ratio of LS of the 20-storey frame structures with a singly symmetric plan
114
Storey
Storey

Set 1 Set 2
Storey

Set 3
(a)
115
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
116
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.5 Inter-storey drift ratio of HS of the 9-storey frame structures with a singly symmetric plan
117
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
118
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
119
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.6 Inter-storey drift ratio of HS of the 20-storey frame structures with a singly symmetric plan
120
Storey

Set 1 Set 2
Storey
Set 3
(a)
121
Storey

Set 1 Set 2
Storey
Set 3
(b)
122
Storey

Set 1 Set 2
Storey

Set 3
(c)
Figure 5.7 Hinge plastic rotation of LS of the 9-storey frame structures with a singly symmetric plan
123
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
124
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
125
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.8 Hinge plastic rotation of LS of the 20-storey frame structures with a singly symmetric plan
126
Storey
Storey
Set 1 Set 2
Storey

Set 3
(a)
127
Storey
Set 1 Set 2
Storey

Set 3
(b)
128
Storey

Set 1 Set 2
Storey

Set 3
(c)
Figure 5.9 Hinge plastic rotation of HS of the 9-storey frame structures with a singly symmetric plan
129
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
130
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
131
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.10 Hinge plastic rotation of HS of the 20-storey frame structures with a singly symmetric
plan from the nonlinear time history analysis method: (a) F2-10; (b) F2-20; (c) F2-30.
132
The peak inter-storey drift ratios and hinge plastic rotations on both LS and HS of all the
buildings, which are subjected to different levels of input motions, are presented in Figure 5.11
to Figure 5.18. The mean demand plus and minus σ is also plotted and denoted as Mean+σ and
Mean-σ, as shown in Figure 5.11 to Figure 5.18. Figure 5.11 presents the height-wise
distribution of the inter-storey drift ratio of LS of the 9-storey frames. It is seen that the results
of MPA and SPA are good and reasonable, as most of them are within the range of one standard
deviation; while CMP significantly overestimated the drift ratio at lower storeys. Compared
with NLRHA, SPA gave a better estimate of the drift ratio on LS from the bottom to 6th storey,
and conservatively estimated the drift ratio at upper storeys, regardless the change of the input
intensities and mass eccentricity ratio. The inter-storey drift ratio on LS of the 20-storey
structures with different mass eccentricity ratios and input intensities is shown in Figure 5.12.
It is seen that MPA predicted the drift ratio best since all the results are within the range of one
standard deviation and the prediction matches the mean drift ratio of NLRHA very well. SPA
gave the best predictions for the drift ratio until the 10th level, then the drift ratios predicted by
SPA become slightly conservative as compared with those of NLRHA. The drift ratio
predicted by CMP method at lower storeys is considerably larger than that of the mean drift
ratio of NLRHA.
Inter-storey drift ratios on HS of the frame structures with a singly symmetric building plan are
plotted in Figures 5.13 and 5.14. It can be seen from Figure 5.13 that for all the input intensities
and mass center eccentricities, SPA was capable of predicting the drift ratio on HS of the 9-
storey frame structure best, since the drift ratio of SPA fits the best with the mean result of
NLRHA and all the results are within the range of one standard deviation. Although most of
the drift ratios of CMP are within the one standard deviation range, CMP method was prone to
overestimating the drift ratios at 1-5 and 7-9 storeys. The MPA method predicted the drift ratio
well at upper levels of the 9-storey frames but exaggerated the drift ratio at lower levels
considerably. As for the inter-storey drift ratios on HS of the 20-storey frames, which is plotted
in Figure 5.14, it is still the SPA procedure that gave the best predictions of the mean drift ratio,
and all the drift ratios are within the range of one standard deviation. CMP overestimated the
drift ratios at both lower and upper storeys and underestimated the drift ratios at the middle
part of the structure for all cases, where frames with different mass eccentricities and under
133
different input motion intensities. MPA estimated the drift ratio at upper storeys satisfactory
but provided conservative prediction at the lower storeys.
From the comparison of the drift ratios, it can be seen that MPA was capable of estimating the
drift ratios on LS of the structure best, but cannot provide reasonable predictions at the lower
part on HS. On the other hand, CMP overpredicted the drift ratio at bottom storeys on LS
considerably but provided a reasonably good estimation of the drift ratio on HS. Despite the
conservative prediction of drift ratios at the upper levels on LS, it is only the SPA method that
can predict the drift ratios very well on both LS and HS, as the results anticipated by SPA show
good agreement with the mean results of NLRHA and most of them are within the range of one
standard deviation.
The height-wise distribution of the peak hinge plastic rotation on LS of the structures subjected
to different levels of ground motion intensity is plotted in Figures 5.15 and 5.16. It is seen
from the hinge plastic rotations on both LS and HS, computed using NLRHA, that with the
increase of the eccentricity ratio and decrease of the seismic mass taken by LS of the frames,
the hinge plastic rotation of the frame members decreases, especially for LS of 20-storey frames,
where the plastic rotation approaches gradually to zero. The minimal hinge plastic rotation
indicates that the plastic deformation of frame members on LS is minor. By comparing the
hinge plastic rotations computed by the pushover methods, SPA and MPA had a reasonable
estimate of the plastic rotation on LS, although the results from SPA are slightly conservative
at upper floors of the 9-storey frames, while CMP overestimated the rotation significantly at
the lower storeys.
The distributions of hinge plastic rotations on HS of the 9-storey and 20-storey frame structures
are shown in Figures 5.17 and 5.18, respectively. The hinge plastic rotations vary along the
height of the structure, and the magnitude of the plastic deformation on HS is much more
significant than that on LS. For comparing the hinge plastic rotations predicted by the different
pushover methods and NLRHA, it is seen that SPA estimated the plastic rotation of hinges best
from the bottom to the middle part of the structures, while MPA and CMP over-predicted the
hinge plastic rotation considerably. When the hinge plastic rotations of the upper part of the
structure occur, there are distinct variations between the results from the pushover methods and
those from NLRHA. This is because both SPA and CMP over-predicted the plastic rotation of
hinges, while MPA provided a reasonable prediction for the hinge plastic rotation. However,
134
most of the hinge plastic rotations predicted by MPA and CMP are not within the one standard
deviation range, while most results of SPA are within the scope of one standard deviation.
Thus, the SPA method can estimate the hinge plastic rotations well, especially at the lower part
of the singly symmetric structure, and most of these predictions are very reasonable.
135
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
136
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
137
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.11 Comparison of inter-storey drift ratio results of LS of frame structures with singly
symmetric building plan: (a) F1-10; (b) F1-20; (c) F1-30.
138
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
139
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
140
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.12 Comparison of inter-storey drift ratio results of LS of frame structures with singly
141
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
142
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
143
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.13 Comparison of inter-storey drift ratio results of HS of frame structures with singly
144
Storey
Storey
Set 1 Set 2
Storey

Set 3
(a)
145
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
146
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
Figure 5.14 Comparison of inter-storey drift ratio results of HS of frame structures with singly
147
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
148
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
149
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
150
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
151
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
152
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
153
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
154
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
155
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
156
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
157
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
158
Storey
Storey
Set 1 Set 2
Storey
Set 3
(c)
159
In this chapter, the spectrum-based pushover analysis (SPA) method for estimating the seismic
demand of one-way asymmetric frame structure is presented. The simplification of the mode
coupling in the nonlinear vibration of structures proposed in the SPA method for a symmetric
structure is adopted in the analysis. Using the forces with mode shape distributions, including
the translation and rotation components, the torsional issues caused by one-way structural
asymmetric is considered. A case study on two series of steel moment-resisting frames, which
have different structural heights and mass centre eccentricities under varying levels of input
motions, was conducted. The NLRHA method, the MPA method, CMP method and the SPA
method were used to perform the analysis and compute the seismic demands of the asymmetric
frame structures. Based on the comparison of the seismic demands obtained from NLRHA and
the advanced pushover analysis methods, the following conclusions can be drawn.
1. For the seismic demands on the heavy side (HS) of the structures, which is more
important for the cases in connection to the torsional effect of buildings, the proposed
SPA method estimated both inter-storey drift ratios and hinge plastic rotations the best
among all the advanced pushover analysis methods. The results of the seismic demands
show very good agreement with those predicted by NLRHA. Also, most of the results
of SPA fall into the range of one standard deviation, while CMP and MPA cannot
anticipate the seismic demands at the lower part of the structures well.
2. On the light side (LS) of the structures, which is farther from the centre of mass, the SPA
method can provide the best predictions of the seismic demands at lower levels of the
structures, while those at upper levels are conservative. The MPA method estimated the
seismic demands, including the inter-storey drift ratios and hinge plastic rotation very
well, and the CMP method significantly overestimated the demands in the lower part of
the structures but has good results in the upper part.
3. Owing to the nature of spectrum-based computations and its efficiency, yet accuracy, the
proposed SPA method can be considered as one of the most promising tools for quickly
estimating the seismic demands of singly symmetric-plan building structures.
160
Chapter 6
Extended Spectrum-based Pushover
Analysis for Estimating Seismic
Forces of Frame and Wall Structures
6.1 Introduction
This chapter presents a method of analysis, named as the extended spectrum-based pushover
analysis (ESPA), based on the SPA procedure for fast and accurate estimation of the
earthquake-induced forces in tall buildings. In ESPA, a linear response spectrum analysis
procedure is added into the SPA procedure, counting the combination of modal forces in the
linear stage. Two special steel moment-resisting frames (SSMRF) and two reinforced concrete
shear wall structures were studied under different ground-motion intensities to investigate the
applicability and accuracy of the ESPA method. A comparative study of the results of the
NLRHA and different methods of advanced pushover analysis, including MPA, CMP and the
proposed ESPA method was conducted. It is shown from the comparative study that the ESPA
method can estimate the earthquake-induced forces very well, and the accuracy of the ESPA
method is comparable with that of the code required NLRHA method required by codes of
practice.
162
6.2 Extended spectrum-based pushover analysis method
In the SPA procedure, forces with first mode-shape distributions are applied to the structure
first, and after the roof displacement of the first stage reached the target value, forces with
second mode distribution will be applied to the structure. In general, the target roof
displacement of the first stage is the largest, and its value is considerably larger than those of
higher modes. The structure will experience extensive deformation and probably yield at the
first stage of the pushover analysis. When the structure yields at the first stage, maximum shear
forces and bending moments of yielded parts calculated by SPA will be close to the shear and
moment capacities of these parts under the first-mode forces. In other words, for the SPA
procedure, the strength of the structure under the first mode distributed forces determines the
shear forces and moments calculated in the nonlinear analysis.
In Figure 6.1, the base shear-roof displacement relations of two frame structures are presented,
and details of the two frame structures can be found later in Section 3.3.2. It is seen from
Figure 6.1 that for both frame structures, shear strengths of higher modes are comparable or
even larger than that of the first mode, and the shear stiffness of higher modes is much higher
than the that of the first mode. Hence, the resultant shear force due to the combination of
different modes in the elastic stage can easily exceed the shear strength of the first mode. This
indicates that the earthquake-induced shear forces of the structures will be underestimated by
SPA since the computation procedure does not consider the combination of forces in elastic
vibration.
163
(a) 9-storey frame structure
(b) 20-storey frame structure

Figure 6.1 Base shear-roof displacement relations of two structures.
To make the SPA method more general and applicable in providing a reasonable estimation of
both earthquake-induced deformations and forces, an extended SPA procedure (ESPA) is
developed for a better estimate of the generic forces. In the ESPA, a linear modal response
spectrum analysis procedure is incorporated into the SPA procedure by taking into account the
mode combination at the linear stage. Details of the proposed ESPA procedure with the linear
modal response spectrum analysis procedure are summarised as follows.
164
(1) Exert seismic gravity loads to building and carry out eigenvalue analysis of structure to
calculate the dynamic properties of the structure, and normalise the mode-shapes to
make the deformation at top equal to one (𝜙 1).
(2) With the design spectrum of a particular region, compute the target roof displacements
of the structure for different modes, u , and the total estimated roof displacement ur 0
using Equations (3-9) and (3-5), respectively. The spectrum displacement D can be
obtained from the elastic design spectrum.
(3) Carry out the nonlinear consecutive pushover procedures:

a. Apply the lateral force with first mode-shape distribution, 𝐬𝟏∗ 𝐦𝛟𝟏 , to the
structure. The first pushover procedure will end until the roof displacement equals
to 𝑢 . Acquire the peak demand of the structure r , including deformations and
forces.
b. Apply the lateral force with second mode-shape distribution, 𝐬𝟐∗ 𝐦𝛟𝟐 , to the
deformed structure, whose structural state is the same as the end of first pushover
procedure. Get the maximum demand 𝑟 of this analysis stage when the increment
of roof displacement of this stage reaches 𝑢 .
c. Repeat the analysis procedures in Step 3(b) for the rest of the modes included to
obtain the peak demand 𝑟 . In each procedure, the pushover procedure will end when
the roof displacement reaches 𝑢 and the structure shall have a structural condition
the same as that at the end of previous pushover analysis.
(4) Carry out linear elastic modal response analysis to the structure under the seismic
gravity loading according to the following sub-steps.
a. Perform the pushover analysis with lateral force with first mode-shape distributions
𝒔𝟏∗ 𝒎𝝓𝟏 , until the roof displacement reaches 𝑢 𝛤 𝜙 𝐷 . Plot the base-
shear and roof-displacement curve and check whether the structure yields. If the
structure yields, take the structural demand when the structure yield as the 𝑟 of this
pushover stage. Otherwise, take the demand at the last step of the pushover analysis
as r .
165
b. Repeat Step 4(a) for the rest of modes, by applying the lateral force 𝒔𝒊∗ 𝒎𝝓𝒊 to
the structure until the roof displacement reached 𝑢 𝛤 𝜙 𝐷 and obtain the
demand value r .
c. Combine the internal force of all the modes considered by the SRSS method:
 
0.5
re  re21  re22  ...  ren2 (6-1)
(5) Envelope the peak demand of the steps above and obtain the seismic demand of the
ESPA procedure:
r  max r1, r2 ,..., rn , re  (6-2)
A flowchart of the ESPA procedure is shown in Figure 6.2.
Figure 6.2 Flowchart of ESPA procedure.
166
6.3 Case study
In this case study, the three sets of ground motions that are given in Section 3.3.1, namely Set
1, Set 2, Set 3, are used for the analysis. As for the structure models, the frame and wall
structures with symmetric building plans, F1, F2, W1, and W2, whose prototypes and member
dimensions are presented in section 3.3.2, are studied.
Earthquake-induced forces, which were calculated by the methods of ESPA, MPA and CMP
and the NLRHA method, were compared. NLRHA, which is commonly recognised as the
most rigorous and accurate method for computing the seismic demand, was performed for all
the ground motions. The mean demand of NLRHA of each set was used as a benchmark for
the comparison. The NLRHA method was implemented with the Wilson-θ time integration
method, where the parameter θ was set to be 1.4 for the convergence of the numerical
integration. The damping ratio of the first and third modes was assumed to be 5% to construct
the damping matrix for analysis. It is stated in that including the first two or three modes is
enough for the MPA procedure. Therefore, when performing the MPA procedure, the first
three modes are included for the 9-storey frame, and first five modes are included for the 20-
storey frame. The mean demand of the set was taken as the demand calculated by MPA for
that motion set. On the other hand, three-stage analysis procedure was used in CMP. The
average roof displacement from NLRHA was used as the total estimated roof displacements in
the CMP procedure. In the SPA method, the analysis procedure follows that presented in
Section 3. All the analytical methods were performed with the nonlinear version of SAP2000,
and P-∆ effects were considered.
6.4 Result of nonlinear response time history analysis
In most of the design codes of practice, the mean demand of the structure under at least seven
ground motions is taken as the design requirement when conducting the NLRHA of the
structure. It is seen that the earthquake-induced deformations computed by the NLRHA
procedure following the provision of codes of practice vary within the mean demands plus and
minus one standard deviation (σ) of the whole set of motions, or just one standard deviation
range. Earthquake-induced forces of the frame structures that are computed following the
167
codes of practice are also presented for all the motion sets. The grouping of the motions is the
same as that in Section 3.4.
Figures 6.3 and 6.4 show the height-wise distribution of the shear forces and bending moments
of the selected groups, where the mean forces, and the forces plus and minus σ of the whole
motion set, are plotted. It can be seen that the forces vary significantly under different
selections of ground motions and the difference is more apparent when the intensity of ground
motion is higher. It is also seen that the forces of different groups from the same set are within
the one standard deviation range. This indicates that the one standard deviation is part of the
value range of the forces computed following the NLRHA procedure of codes. Thus, one
standard deviation range is used as a benchmark to judge the reasonability of the forces
calculated by the pushover analysis methods.
168
Storey
Storey
Shear force (kN) Shear force (kN)
Set 1 Set 2
Storey
Shear force (kN)

Set 3
(a)
169
Storey
Storey

Set 1 Set 2
Storey
Shear force (kN)
Set 3
(b)
Figure 6.3 Shear forces of the frame structures from nonlinear time history analysis method: (a) F1;
(b) F2.
170
Storey
Storey
Bending moment (kNm) Bending moment (kNm)
Set 1 Set 2
Storey
Bending moment (kNm)

Set 3
(a)
171
Storey
Storey

Set 1 Set 2
Storey

Set 3
(b)
Figure 6.4 Bending moment of the frame structures from nonlinear time history analysis method: (a)
F1; (b) F2.
172
Storey
Shear force (kN) Storey Shear force (kN)

Set 1 Set 2
Storey
Shear force (kN)
Set 3
(a)
173
Storey
Shear force (kN) Storey Shear force (kN)

Set 1 Set 2
Storey
Shear force (kN)

Set 3
(b)
Figure 6.5 Shear forces of the wall structures from nonlinear time history analysis method: (a) W1; (b)
W2.
174
Storey
Bending moment (kNm) Storey Bending moment (kNm)
Set 1 Set 2
Storey

Set 3
(a)
175
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
Figure 6.6 Bending moment of the wall structures from nonlinear time history analysis method: (a)
W1; (b) W2.
176
6.5.1 Comparison of results from ESPA and NLRHA
Figure 6.7 shows the comparison of the shear forces and bending moment of motion Set 2. In
the figures, the earthquake-induced forces calculated by the linear and nonlinear procedure in
the proposed ESPA method are plotted. It is seen from Figure 6.7 (a), where the shear forces
of the frame are shown, that the linear combination of the modal forces of the 9-storey frames
are larger than that of the nonlinear procedures and are closer to the shear forces of the NLRHA
method. For the shear forces of the 20-storey frame structure, below the 8th floor, the shear
forces from the linear procedure are more significant than that from the nonlinear process and
are much closer to that of NLRHA while the trend reverses for the shear forces above the 8th
floor. The bending moments of ground motion Set 2 calculated by the linear and nonlinear
procedures in ESPA and NLRHA are plotted in Figure 6.7 (b). It is seen that the bending
moments from the nonlinear procedure are higher than that of the linear process and is closer
to that of the NLRHA method for both structures. From the comparison of shear forces and
bending moments calculated by the linear and nonlinear procedures of ESPA and NLRHA, it
is noticeable that the nonlinear process of the ESPA method is capable of capturing the
earthquake-induced moments well, while when computing the earthquake-induced shear forces,
both linear and nonlinear modal combinations should be considered.
Storey
Shear force (kN)

F1
177
Storey
Shear force (kN)

F2
(a) Shear forces
Storey
F1
178
Storey

F2
(b) Bending moment
Figure 6.7 Comparison of earthquake-induced forces of ESPA and NLRHA for motion Set 2.
6.5.2 Comparison of results from pushover analysis methods and NLRHA
The comparison of shear forces of the 9-storey and 20-storey frame structures calculated by
the proposed ESPA, MPA, CMP methods and the most rigorous NLRHA method is shown in
Figure 6.8. It is seen in Figure 6.8 (a) that all other pushover methods provided a very good
estimation of the earthquake-induced shear forces of the 9-storey frame, despite the different
input intensities, since all the results are close to those from NLRHA and are within the one
standard deviation range. Precisely, the MPA method predicted the shear forces at the lower
part of the building better, while the CMP and the ESPA methods predicted the shear forces at
the upper portion better.
From Figure 6.8 (b), where the shear forces results of the 20-storey frame structure are depicted,
it still can be seen that the shear forces from the MPA method are closer to the NLRHA method
in the lower part, while shear forces from ESPA and CMP methods are more accurate at the
upper portion. It should be noted that although the CMP method predicted the shear forces
well at the upper part of the 20-storey frame structure, the lower part shear forces from CMP
is much smaller than (MEAN – σ) values, indicating that the CMP method cannot provide
satisfactory estimation of shear forces at the lower part of the 20-storey frame structure.
179
However, for ESPA and MPA, all the predictions are within the one standard deviation range,
and the results are satisfactory, despite some variations from the mean forces of the NLRHA
method. It is also seen that the maximum differences between the computed shear forces of
the 9-storey and 20-storey frames from ESPA and those from NLRHA are 18% and 16%,
respectively; while the maximum differences in the results from MPA and CMP, as compared
with the results of the NLRHA method, are all larger than 26%, showing that the ESPA method
can give better predictions of the shear forces than MPA and CMP.
Figure 6.9 shows the comparison of earthquake-induced bending moments of both structures
calculated by the pushover methods and the NLRHA method. For the CMP, MPA and ESPA
methods, similar predictions of the bending moments of the 9-storey frame are obtained, and
all the predictions agree with the results of NLRHA very well, with a maximum difference of
about 11% between the computed bending moments of the multi-mode pushover analysis
methods and the NLRHA method. It is also noticed that all the results of the multi-mode
pushover analysis methods are falling within the one standard deviation range. However, for
the 20-storey frame structure, the MPA method underestimated the bending moments over the
height of the 20-storey frame structure. Especially for ground motions Set 1, where the input
intensity is lowest, most of the bending moments of the MPA method are even smaller than the
(MEAN – σ) value. While the ESPA and CMP methods predicted the bending moments of the
20-storey frame much better than the MPA method, with all the results match well with the
mean of NLRHA and all results are fell within the one standard range. Additionally, ESPA can
predict the bending moments well with a maximum difference of 17% when compared with
the results from NLRHA, while both the MPA and CMP resulted in a maximum difference of
25% in bending moment prediction as compared with NLRHA method.
180
Storey
Storey
Set 1 Set 2
Storey
Shear force (kN)

Set 3
(a)
181
Storey
Storey
Set 1 Set 2
Storey
Shear force (kN)

Set 3
(b)
Figure 6.8 Comparison of shear force results of the frame structures: (a) F1; (b) F2.
182
Storey
Storey

Set 1 Set 2
Storey

Set 3
(a)
183
Storey
Storey
Set 1 Set 2
Storey

Set 3
(b)
Figure 6.9 Comparison of bending moment results of the frame structures: (a) F1; (b) F2.
184
Figure 6.10 presents the distribution of shear forces along the height of the wall structures with
a symmetric building plan. For W1, all the pushover analysis methods have a similar prediction
of the shear forces at the lower part of the structure and all the shear forces computed by the
pushover analysis methods are smaller than that of the NLRHA method. While for the upper
part of the W1 structure, both MPA and ESPA method can provide a very good prediction of
the shear forces there, but the CMP method tended to overestimate the shear force significantly.
In Figure 6.10 (b), where the shear forces of W2 are presented, it is very clear that the CMP
method cannot reasonably predict the seismic induced shear force of W2, since the results from
the CMP method are considerably smaller and larger than that of the NLRHA method at the
lower and upper part of the structure respectively. As for MPA and ESPA method, except for
ground motion Set 3, they can well estimate the seismic shear force of the structure and the
results are within one standard deviation range. When W2 is subjected to ground motions in
motion Set 3, both the MPA and ESPA methods overpredicted the shear forces at upper storeys.
As a consequence of the results comparison, both MPA and ESPA method can well predict the
shear forces, as the results are close to that of the NLRHA method, and almost all the results
are in the one standard deviation range. However, it is also seen that the CMP method failed
in reasonably predicting the shear force of the wall structures.
The plot in Figure 6.11 is the comparison of the heightwise distribution of the bending moment
of the wall structures. By comparison, it is seen that despite the variation of the input ground
motion levels, the CMP method cannot well predict the bending moment of the shear walls as
it tended to overvalue the bending moment at the upper levels significantly. The results of the
ESPA and MPA method are very close, and almost all of them approximate the NLRHA results
very well and are located in the one standard deviation range. Thus, it can be concluded that
when computing the earthquake induced forces of the wall structures, both ESPA and MPA
method can provide a reasonably good estimation of the shear forces and bending moment, but
the CMP method is not able to reasonably predict the forces.
185
Storey
Storey
Set 1 Set 2
Storey
Shear force (kN)
Set 3
(a)
186
Storey
Storey
Set 1 Set 2
Storey
Shear force (kN)

Set 3
(b)
Figure 6.10 Comparison of shear force results of the wall structures: (a) W1; (b) W2.
187
Storey
Storey
Set 1 Set 2
Storey
Set 3
(a)
188
Storey
Storey
Set 1 Set 2
Storey
Set 3
(b)
Figure 6.11 Comparison of bending moment results of the wall structures: (a) W1; (b) W2.
189
This chapter presents a method of analysis named as the extended spectrum-based pushover
analysis (ESPA) for estimating the earthquake-induced forces in tall buildings. In ESPA, a
linear response spectrum analysis procedure is incorporated into the SPA method to consider
the linear combination effect of vibration modes in the seismic response of buildings, especially
the earthquake-induced forces. A case study on two moment resisting frames and two
reinforced concrete shear wall structures with symmetric building plan are conducted. The
results from the proposed ESPA method with those of the NLRHA method, and the advanced
pushover methods, including the MPA method, and the CMP method are compared to verify
the accuracy of the proposed ESPA method. Based on this study, the following conclusions
are drawn.
1. The one standard deviation range is a part of the value range of the earthquake-induced
forces that are computed following the requirements of the codes of practice, so the one
standard deviation range can be a good measure for judging the reasonability of the
results of pushover analysis methods.
2. For frame structures, the MPA method can only provide a good prediction of shear forces,
but not bending moments, since some of the results from MPA are smaller than the
(MEAN – σ) value of NLRHA method. The CMP method predicts the bending moment
very well, but cannot reasonably anticipate the shear forces, as the CMP method
significantly underestimated the shear forces at the lower part of the example frame
structures. While for wall structures, the MPA method can well estimate both shear forces
and bending moment, but the CMP method failed to reasonably predict the shear forces
and overturning moment.
3. The proposed ESPA predicted both shear forces and bending moments best among the
pushover analysis methods for both structure types, as all the results from the ESPA are
close to the NLRHA mean and fall within the one standard deviation range. Since the
ESPA is efficient, accurate and easy to perform, it is believed that the ESPA is one of the
most promising tools for the fast predict of the earthquake-induced forces.
190
Chapter 7
Conclusion
7.1 Conclusions
The principal objective of this research on the development of an effective, efficient, accurate
yet user-friendly method of analysis for the fast estimation of the seismic demand of tall
buildings with different structural types have been attained. By theoretical studies and
numerical studies, the following conclusions can be drawn.
1. In most of the design codes of practice, it is required to take the mean demand of seven
nonlinear response time history analyses as the seismic design demand. However, as the
frequency contents of the motions are very different, it is found from that the computed
seismic demand varies significantly, and the demands of any seven motions from a set
of motions are seen to most likely located in the mean demands of the set plus and minus
the one standard deviation range (one standard deviation range) of the demands of the
motion set. As a result, the one standard deviation range is used to verify the
reasonability of the pushover analysis method. If the results from the pushover analysis
methods are within the one standard deviation range, the results can be recognized as
reasonable.
2. To consider the coupling effect of modes of vibration while keeping the efficiency of the
pushover analysis method, a simplification of the mode coupling is adopted, where the
modes of vibration are directly connected with the structural state (internal stress, strain
and displacement) when the vibration of the mode is initiated. The simplification is that
when the ith mode of vibration starts, the roof displacements of all previous modes (the
(i−1) th, (i−2)th, …, and 1st mode) of vibration have already reached the target values,
and the structural conditions, such as roof displacement, storey drift, internal stress and
strain of the ith mode keep unchanged after the roof displacement of ith mode reaches
the target value.
3. Based on the simplification of the mode coupling, a spectrum-based pushover analysis

method (SPA) is proposed for the fast prediction of the seismically induced deformation
of structures with symmetric building plan and singly structural form, such as frame
structures and wall structures. The analysis is performed with the consecutive pushover
technique, and the earthquake input is generated from the response spectrum. From the
comparison of the results from the proposed SPA method, two famous advanced
192
pushover analysis methods and the most rigorous nonlinear response time history
analysis (NLRHA) method, it is found that the seismic induced deformation calculated
by the SPA method, including inter-storey drift ratio and hinge plastic rotation,
approximate that of the NLRHA method very well, and almost all the results are within
one standard deviation range. This means that the SPA method is capable of providing
very good and reasonable estimation of the seismically induced deformation of buildings
with symmetric building plan and singly structural form.
4. Dual wall-frame structures, which consist of a shear wall system and a frame system,
have very different structural behaviour and damage mode under the action of the
earthquake effect, as compared with frame and wall structures. Consequently, when
predicting the seismic demand of the dual wall-frame structures, the wall-frame
interaction and the particular damage mode of the wall-frame structures should be
considered. Based on the damage mode of the dual-wall frame structures, a modification
of the SPA method is made, and a modified SPA method is developed for the fast
prediction of the seismically induced deformation of dual wall-frame structures with a
symmetric building plan. It is found from the comparison of the results computed by the
NLRHA methods, the proposed MSPA method as well as two advanced pushover
analysis methods that only the MSPA can well predict the trend of the distribution of the
seismic deformation of the structures well, and the seismic deformation of the MSPA
method is very good and reasonable.
5. The SPA method is applied to predict the seismic deformation of frame structures with a
singly symmetric building plan, by including the torsional effect into consideration.
From a comprehensive case study involving structures with different mass eccentricity
ratios under varying levels of ground motions, it is found that the SPA method can well
and reasonably predict the seismic deformation of the frame structures with a singly
symmetric building plan.
6. As the analysis procedure of the SPA method only considers the combination of modal
forces in the nonlinear vibration procedures, the calculated forces of some part equal to
yield strength when the structures are subjected to forces with first mode distributions,
which will induce the underestimation of the seismic induced force. Thus, a linear
response spectrum analysis is added to the original version of SPA method, and the SPA
193
method is extended to predict the seismic induced force. Comparing the seismic induced
force computed from the extended SPA method, the NLRHA method, the CMP method
and the MPA method, it is found that the ESPA can also well estimate the seismic induced
force and the result is reasonable.
7. Owing to the accuracy of the results, the efficiency of the calculation procedure as well
as the spectrum-based analysis procedure, the proposed spectrum-based pushover
analysis methods, including the SPA method, the MSPA method and the ESPA method,
are believed to be one of the most promising tools for the fast prediction of the seismic
demand of a large number of buildings, such as regional seismic assessment.
194
7.2 Future research
This study proposed a series of spectrum-based pushover analysis procedures for quickly and
reasonably estimating the seismic demand of tall buildings with different structural forms,
including frame structures, shear wall structures and dual wall-frame structures. There are
several lines of research arising from the study may be pursued in the future.
1. The SPA procedures are proved to be effective in predicting seismic demands of tall
buildings with three of the most common structural types, but there are still other
structural forms for tall buildings, such as braced frame structure, coupled shear wall
structure and structures with transfer storey. The applicability of the spectrum-based
pushover analysis procedures on the prediction of seismic demands of other structure
forms of tall buildings will also need to be verified in the future.
2. Additional versions of the SPA methods may also be developed for the quick assessment
of other kinds of structures, such steel-concrete composite and timber structures, bridges,
dams, and so on.
3. Since the different selection of motions will induce significant variations, and the SPA
method can get a very good estimation of the seismic demand, it may be possible to
develop a pushover-based motion selection technique to quickly select a group of
motions with which the computed seismic demand is very close to that of the whole
ground motion set.
195
References
Abbasnia R, Davoudi AT, Maddah MM (2013). “An adaptive pushover procedure based on
effective modal mass combination rule.” Engineering Structures. 52, 654-666.
American Concrete Institute (ACI). (2014). “Building code requirements for reinforced
concrete.” ACI 318-14
Antoniou S, Pinho R (2004a). “Advantages and limitations of adaptive and non-adaptive force-
based pushover procedures.” Journal of Earthquake Engineering. 8, 497-522.
Antoniou S, Pinho R (2004b). “Development and verification of a displacement-based adaptive
pushover procedure.” Journal of Earthquake Engineering. 8, 643-661.
Antoniou S, Rovithakis A Pinho R (2002). “Development and verification of a fully adaptive
pushover procedure.” in Proceedings of the Twelfth European Conference on Earthquake
Engineering. London, UK.
American Society of Civil Engineers (ASCE). (2013). “Minimum Design Loads for Buildings
and Other Structures” ASCE/SEI 7-10. American Society of Civil Engineering, Reston,
Virginia.
Applied Technology Council (ATC). (1996). “Seismic evaluation and retrofit of concrete
buildings.” ATC-40. Applied Technology Council, report Redwood City, California.
Baros DK, Anagnostopoulos SA (2017). “A new approach for 3-D pushover-based analysis of
asymmetric buildings: development and initial evaluation.” Earthquake and Structures. 12,
543-557.
Bhatt C, Bento R (2011). “Extension of the CSM‐FEMA440 to plan‐asymmetric real building
structures.” Earthquake Engineering & Structural Dynamics. 40, 1263-1282.
Bhatt C, Bento R (2014). “The extended adaptive capacity spectrum method for the seismic
assessment of plan-asymmetric buildings.” Earthquake Spectra. 30, 683-703.
Bosco M, Ghersi A, Marino EM, Rossi PP (2013). “Comparison of nonlinear static methods
for the assessment of asymmetric buildings.” Bulletin of Earthquake Engineering. 11, 2287-
2308.
Brozovič M, Dolšek M (2014). “Envelope-based pushover analysis procedure for the
approximate seismic response analysis of buildings.” Earthquake Engineering & Structural
Dynamics. 43, 77-96.
Casarotti C, Pinho R (2007). “An adaptive capacity spectrum method for assessment of bridges
subjected to earthquake action.” Bulletin of Earthquake Engineering. 5, 377-390.
CEN (2004). Eurocode 8: Design of structures for earthquake resistance—Part 1: general
rules, seismic actions and rules for buildings, EN 1991-1. European Committee for
Standardization, Brussels
196
Chinese Standard (2008). Code for Seismic Design of Buildings, GB 50011-2010. Chinese
Building Press, Beijing, China.
Chintanapakdee C, Chopra AK (2003). “Evaluation of modal pushover analysis using generic
frames.” Earthquake Engineering & Structural Dynamics. 32, 417-442.
Chopra AK (2012). “Dynamics of structures: Theory and Applications to Earthquake
Engineering.” 4th edition. Prentice Hall, Hong Kong.
Chopra AK, Goel RK (2004). “A modal pushover analysis procedure to estimate seismic
demands for unsymmetric-plan buildings.” Earthquake Engineering & Structural Dynamics.
33, 903-927.
Chopra AK, Goel RK (2002). “A modal pushover analysis procedure for estimating seismic
demands for buildings.” Earthquake Engineering & Structural Dynamics. 31, 561-582.
Chopra AK, Goel RK (1999). “Capacity-demand-diagram methods based on inelastic design
spectrum.” Earthquake Spectra. 15, 637-656.
Chopra AK, Goel RK, Chintanapakdee C (2004). “Evaluation of a modified MPA procedure
assuming higher modes as elastic to estimate seismic demands.” Earthquake Spectra 20,
757-778.
CSI SV. (2011). “15: Integrated finite element analysis and design of structures basic analysis
reference manual.” Berkeley, California (USA). Computers and Structures Inc.
Fajfar P (2000). “A nonlinear analysis method for performance-based seismic design.”
Earthquake Spectra 16, 573-592.
Fajfar P (1999). “Capacity spectrum method based on inelastic demand spectra.” Earthquake
Engineering & Structural Dynamics. 28, 979-994.
Fajfar P, Fischinger M (1988). “N2-A method for non-linear seismic analysis of regular
buildings.” in Proceedings of the Ninth World Conference in Earthquake Engineering,
Tokyo-Kyoto, Japan.
Fajfar P, Gašperšič P (1996). “The N2 method for the seismic damage analysis of RC buildings.”
Earthquake Engineering & Structural Dynamics. 25, 31-46.
FEMA (2000). “Prestandard and Commentary for the Seismic Rehabilitation of Buildings.”
FEMA-356, Federal Emergency Management Agency, Washington, D.C.
FEMA (2002). “Rapid Visual Screening of Buildings for Potential Seismic Hazards, A
Handbook.” FEMA 154, Federal Emergency Management Agency, Washington, D.C.
FEMA (2005). “Improvement of nonlinear static seismic analysis procedures.” FEMA-440,
Federal Emergency Management Agency, Washington, D.C.
Ferracuti B, Pinho R, Savoia M, Francia R (2009). “Verification of displacement-based
adaptive pushover through multi-ground motion incremental dynamic analyses.”
Engineering Structures. 31, 1789-1799.
Fragiadakis M, Vamvatsikos D (2010). “Fast performance uncertainty estimation via pushover
and approximate IDA.” Earthquake Engineering & Structural Dynamics. 39, 683-703.
197
Goel RK, Chopra AK (2005) “Extension of modal pushover analysis to compute member
forces.” Earthquake Spectra. 21, 125-139.
Goel RK, Chopra AK (2004). “Evaluation of modal and FEMA pushover analyses: SAC
buildings.” Earthquake Spectra. 20, 225-254.
Gupta B, Kunnath SK (2000). “Adaptive spectra-based pushover procedure for seismic
evaluation of structures.” Earthquake Spectra. 16, 367-392.
Huang K (2009). “Continuum MDOF model for seismic analysis of wall-frame structures.”
Ph.D. Thesis. Hong Kong University of Science and Technology, Hong Kong.
Huang K, Kuang JS (2010). “On the applicability of pushover analysis for seismic evaluation
of medium-and high-rise buildings.” The Structural Design of Tall and Special Buildings.
19, 573-588.
IBC I. (2006). “International building code.” International Code Council, Inc.(formerly BOCA,
ICBO and SBCCI) 4051, 60478-65795.
Jan TS, Liu MW, Kao YC (2004). “An upper-bound pushover analysis procedure for
estimating the seismic demands of high-rise buildings.” Engineering Structures. 26, 117-
128.
Jiang Y, Li G, Yang D (2010). “A modified approach of energy balance concept based
multimode pushover analysis to estimate seismic demands for buildings.” Engineering
Structures. 32, 1272-1283.
Kaats𝚤z K, Sucuoğlu H (2014). “Generalized force vectors for multi-mode pushover analysis
of torsionally coupled systems.” Earthquake Engineering & Structural Dynamics. 43, 2015-
2033.
Kalkan E, Kunnath SK (2007). “Assessment of current nonlinear static procedures for seismic
evaluation of buildings.” Engineering Structures. 29, 305-316.
Kalkan E, Kunnath SK (2006). “Adaptive modal combination procedure for nonlinear static
analysis of building structures.” Journal of Structural Engineering. 132, 1721-1731.
Khoshnoudian F, Kashani MMB (2012). “Assessment of modified consecutive modal
pushover analysis for estimating the seismic demands of tall buildings with dual system
considering steel concentrically braced frames.” Journal of Constructional Steel Research.
72, 155-167.
Khoshnoudian F, Kiani M (2012a). “Modified consecutive modal pushover procedure for
seismic investigation of one-way asymmetric-plan tall buildings.” Earthquake Engineering
and Engineering Vibration. 11, 221-232.
Kilar V, Fajfar P (1997). “Simple push-over analysis of asymmetric buildings.” Earthquake
Kreslin M, Fajfar P (2012). “The extended N2 method considering higher mode effects in both
plan and elevation.” Bulletin of Earthquake Engineering. 10, 695-715.
198
Kreslin M, Fajfar P (2011). “The extended N2 method taking into account higher mode effects
in elevation.” Earthquake Engineering & Structural Dynamics. 40, 1571-1589.
Kuang JS (2014). “CIVL 5310 Course Notes: Structural Analysis and Design of Tall Buildings.”
the Hong Kong University of Science and Technology, Hong Kong.
Leelataviwat S, Saewon W, Goel SC (2009). “Application of energy balance concept in seismic
evaluation of structures.” Journal of Structural Engineering. 135, 113-121.
Lin Y, Chang K (2003). “An improved capacity spectrum method for ATC‐40.” Earthquake
Liu Y, Kuang JS (2017). “Spectrum-based pushover analysis for estimating seismic demand
of tall buildings.” Bulletin of Earthquake Engineering. 15, 4193-4214.
Liu Y, Kuang JS, Huang Q (2018). “Modified spectrum-based pushover analysis for estimating
seismic demand of dual wall-frame systems.” Engineering Structures. 165, 302-314.
Lu X, Guan H. (2016) “Earthquake Disaster Simulation of Civil Infrastructures.” Springer,
Beijing, China.
Manoukas G, Athanatopoulou A, Avramidis I (2012). “Multimode pushover analysis for
asymmetric buildings under biaxial seismic excitation based on a new concept of the
equivalent single degree of freedom system.” Soil Dynamics and Earthquake Engineering.
38, 88-96.
Miao Z, Ye L, Guan H, Lu X (2011). “Evaluation of modal and traditional pushover analyses
in frame-shear-wall structures.” Advances in Structural Engineering. 14, 815-836.
Miranda E. (1999). “Approximate seismic lateral deformation demands in multistory buildings.”
Journal of Structural Engineering. 125, 417-425.
Miranda E, Reyes CJ (2002). “Approximate lateral drift demands in multistory buildings with
nonuniform stiffness.” Journal of Structural Engineering. 128, 840-849.
Moghaddam H, Hajirasouliha I (2006). “An investigation on the accuracy of pushover analysis
for estimating the seismic deformation of braced steel frames.” Journal of Constructional
Steel Research. 62, 343-351.
NZS (2004) “Structural Design Actions” NZS 1170.5. Standards New Zealand technical
Committee, Wellington, New Zealand.
Ohtori Y, Christenson RE, Spencer Jr BF, Dyke SJ (2004). “Benchmark control problems for
seismically excited nonlinear buildings.” Journal of Engineering Mechanics. 130, 366-385.
Papanikolaou VK, Elnashai AS (2005). “Evaluation of conventional and adaptive pushover
analysis I: Methodology.” Journal of Earthquake Engineering. 9, 923-941.
Poursha M, Khoshnoudian F, Moghadam AS (2014) “The extended consecutive modal
pushover procedure for estimating the seismic demands of two-way unsymmetric-plan tall
buildings under influence of two horizontal components of ground motions.” Soil Dynamics
and Earthquake Engineering. 63, 162-173.
199
Poursha M, Khoshnoudian F, Moghadam AS (2011). “A consecutive modal pushover
procedure for nonlinear static analysis of one-way unsymmetric-plan tall building
structures.” Engineering Structures. 33, 2417-2434.
Poursha M, Khoshnoudian F, Moghadam AS (2009). “A consecutive modal pushover
procedure for estimating the seismic demands of tall buildings.” Engineering Structures. 31,
591-599.
Poursha M, Samarin ET (2015). “The modified and extended upper-bound (UB) pushover
method for the multi-mode pushover analysis of unsymmetric-plan tall buildings.” Soil
Dynamics and Earthquake Engineering. 71, 114-127.
Reyes JC, Chopra AK (2011a). “Evaluation of three-dimensional modal pushover analysis for
unsymmetric-plan buildings subjected to two components of ground motion.” Earthquake
Reyes JC, Chopra AK (2011b). “Three-dimensional modal pushover analysis of buildings
subjected to two components of ground motion, including its evaluation for tall buildings.”
Shakeri K, Shayanfar MA, Kabeyasawa T (2010). “A story shear-based adaptive pushover
procedure for estimating seismic demands of buildings.” Engineering Structures. 32, 174-
183.
Shakeri K, Tarbali K, Mohebbi M (2012a). “An adaptive modal pushover procedure for
asymmetric-plan buildings.” Engineering Structures. 36, 160-172.
Sucuoğlu H, Akkar S (2014). “Basic Earthquake Engineering.” Springer,
Sucuoğlu H, Günay MS (2011). “Generalized force vectors for multi-mode pushover analysis.”
Surmeli M, Yuksel E (2015). “A variant of modal pushover analyses (VMPA) based on a non-
incremental procedure.” Bulletin of Earthquake Engineering. 13, 3353-3379.
Tarbali K, Shakeri K (2014a). “Story shear and torsional moment-based pushover procedure
for asymmetric-plan buildings using an adaptive capacity spectrum method.” Engineering
Structures. 79, 32-44.
Yakut A. (2004). “Preliminary seismic performance assessment procedure for existing RC
buildings.” Engineering Structures. 26, 1447-1461.
Zhang J, Jiang J, Li G (2017) An improved consecutive modal pushover procedure for
estimating seismic demands of multi‐storey framed buildings. The Structural Design of Tall
and Special Buildings 26(4).
200

Spectrum-Based Pushover Analysis For Predicting Seismic Demands of Tall Buildings

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Spectrum-Based Pushover Analysis For Predicting Seismic Demands of Tall Buildings

Uploaded by

Copyright:

Available Formats

Spectrum-based Pushover Analysis for Predicting

Seismic Demands of Tall Buildings

1.1 Background ...................................................................................................................... 2

Chapter 2 Literature Review ...................................................................................................... 7

2.1 Introduction ...................................................................................................................... 8

2.3.1 Nonlinear static analysis method ............................................................................ 11

2.4 Current advanced pushover procedures ......................................................................... 20

Chapter 3 Spectrum-based Pushover Analysis for Estimating Seismic Deformation of Frame

3.1 Introduction .................................................................................................................... 32

3.3.1 Details of ground motions ....................................................................................... 39

3.5.1 Frame structures ...................................................................................................... 56

3.6 Concluding remarks ....................................................................................................... 65

Chapter 4 Modified Spectrum-based Pushover Analysis for Estimating Seismic Deformation

4.1 Introduction .................................................................................................................... 68

4.3.1 Ground motions selection ....................................................................................... 75

Chapter 5 Spectrum-based Pushover Analysis for Estimating Seismic Deformation of Frame

5.1 Introduction .................................................................................................................. 100

5.3.1 Ground motions selection ..................................................................................... 105

Chapter 6 Extended Spectrum-based Pushover Analysis for Estimating Seismic Forces of

6.1 Introduction .................................................................................................................. 162

6.5.1 Comparison of results from ESPA and NLRHA .................................................. 177

6.6 Concluding remarks ..................................................................................................... 190

Chapter 7 Conclusion ............................................................................................................. 191

7.1 Conclusions .................................................................................................................. 192

References .............................................................................................................................. 196

1. Coupling of modes in the nonlinear stages

Wall Frame Wall-frame

4. Generic forces in the buildings

where n is the nth natural vibration mode of the structure, and:

where the modal coordinates are governed by

The solution of Equation (2-6) is given by

where Dn (t ) is governed by the equation of motion for a single-degree-of-freedom (SDOF)

system subjected to ug (t ) :

The floor displacement can then be represented as

2.3 Current provisions in codes of practice

Figure 2.1 Multiple capacity curves required modelling strength degradation.

spectral acceleration; Sd is the spectral displacement;

2. Construct Bilinear Representation of the capacity spectrum;

3. Intersect Capacity Spectrum and Demand Spectrum as illustrated in Figure 2.3.

Figure 2.3 Intersection point of Demand and Capacity Spectrum.

b. FEMA 356 method

the effect of pinching in load-deformation relationships due to degradation in stiffness and

horizontal displacement pattern i :

And the force F * and displacement d * of the equivalent system are

Multi Degree of Freedom (MDOF) system.

Where Se (T * ) is the elastic acceleration response spectrum at the period T * .

a) T *  TC (Short period range)

If Fy / m  Se (T ) , the response is elastic and thus

dt*  d et* (2-29)

If Fy / m  Se (T ) , the response is nonlinear and

Se (T * ) and the structure with limited strength Fy* / m* .

b) T *  TC (medium and long period range)

dt*  d et* (2-31)

The relation between different quantities can be visualized in Figure 2.6.

(b) Medium and long period range

d. Chinese Standard 2008

2.3.2 Nonlinear response time history analysis method

2.4 Current advanced pushover procedures

Envelope-based Pushover Analysis Method

Modal Pushover Analysis Method

Generalised pushover analysis procedure

Consecutive modal pushover analysis methods