Professional Documents
Culture Documents
b1146296 PDF
b1146296 PDF
By
Hiu Tik CHAN
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
HKUST Library
Reproduction is prohibited without the author’s prior written consent
ACKNOWLEDGEMENT
Thank you and praise to my Lord Jesus Christ.
My sincere gratitude is to my supervisor, Professor J.S. Kuang, for his
encouragement, advice and research support throughout my doctoral
study. It would be impossible to complete the project without his help
on any difficulties encountered throughout the duration of this project. I
have enjoyed our discussion, dispute and sharing every successes and
failures.
I would like to express my endless thanks to my thesis advisory
committee, Professor Zongjin Li and Professor Lambros S. Katafygiotis, for
their invaluable advice throughout the course of this research work.
I also wish to thank the instructors of the courses I took during my
graduate studies at HKUST, namely: Professor J.S. Kuang, Professor
Christopher K.Y. Leung, Professor Lambros S. Katafygiotis, Professor C.C.
Chang, Professor Zongjin Li, Professor C.M. Chan, Professor Moe M.S.
Cheung, Professor X.S. Li, and Professor Thomas T.C. Hsu. I learned a
great deal from them all.
I am grateful to all my colleagues for encouragement and willing
assistance in this research, especially Dr. K. Huang, Dr. Y.B. Ho, Ms. Gigi S.P.
Suen, Mr. Vincent H.F. Tsoi, Mr. Terry Y.P. Yuen, Mr. Henry S.H. Luk and Mr
Will Z. Wang.
Finally, I would like to thank my family members for their love,
support, patience and understanding.
iv
Table of Contents
TITLE PAGE i
AUTHORISATION PAGE ii
SIGNATURE PAGE iii
ACKNOWLEDGMENT iv
TABLE OF CONTENTS v
LIST OF FIGURES viii
LIST OF TABLES xiii
ABSTRACT xiv
Chapter 1
Introduction
1.1 Research background 1
1.2 Objectives 4
1.3 Scope of thesis 5
Chapter 2
Literature Review
2.1 Overview of current seismic assessment 6
2.2 Modal analysis 9
2.3 Displacement coefficient method 13
2.4 Drift spectrum 16
2.5 Sheikh’s method 17
2.6 Koru’s method 18
2.7 Lang’s method 19
2.8 Indices based assessment 20
2.9 Capacity spectrum method (CSM) 22
2.10 Nonlinear static analysis 23
v
2.11 Nonlinear response history analysis 24
2.12 FEMA 154 26
2.13 Seismic demand for asymmetric‐plan buildings 27
2.14 Research significance 29
Chapter 3
Quick Estimate of Fundamental Period of Tall Buildings
3.1 Introduction 37
3.2 Continuum flexural‐Shear cantilever 39
3.3 Dynamics properties of the model 40
3.4 A model for quick estimate of fundamental period T1 41
3.5 Prediction of degree of structural interaction 43
3.6 Degree of structural interaction of non‐uniform building 45
3.7 Numerical studies 47
3.8 Concluding remarks 49
Chapter 4
Performance‐Based Seismic Assessment of Wall‐Frame
Building Structures Using Inelastic Seismic Interstorey Drift Ratio
4.1 Introduction 60
4.2 Bifurcation index for elastic and inelastic behaviour 62
4.3 Model of inelastic interstorey drift ratio 65
4.4 Numerical investigation 69
4.5 Concluding remarks 74
vi
Chapter 5
Modified Continuum MDOF Model for Seismic Analysis of
Building Structures
5.1 Introduction 87
5.2 Modified continuum MDOF model 89
5.3 Numerical investigation of the modified continuum MDOF Model 95
5.4 General continuum representation 97
5.5 Concluding remarks 105
Chapter 6
Generalised Inelastic Response Spectra
6.1 Introduction 116
6.2 Methodology 119
6.3 Generalised inelastic response spectra 120
6.4 Procedure of assessment using generalised inelastic response 122
6.5 Numerical investigation 124
6.6 Concluding remarks 127
Chapter 7
Conclusion
7.1 Conclusion 156
7.2 Further research 159
References 160
Appendices
A.1 Continuum flexural‐shear cantilever 164
A.2 Model under uniform lateral loading 171
A.3 Model under inverted triangular lateral loading 176
A.4 Model under point load at the roof 180
vii
List of Figures
Chapter 2
Fig. 2.1 A single‐degree of freedom system 30
Fig. 2.10 Data collection forms for
the three designated seismicity regions 35
Chapter 3
Fig. 3.1 Continuum flexural‐shear model 51
Fig. 3.3 Free body diagram of flexural‐shear cantilever at time t
under free vibration 52
Fig. 3.4 Free body diagram of flexural‐shear cantilever
under lateral loading 53
viii
Fig. 3.7 The floor plans of the building 55
Fig. 3.8 The variation of stiffness and
the equivalent stiffness of the building 56
Fig. 3.9 (a) From PEER with record ID: P0714
(b) From PEER with record ID: P0410 57
Chapter 4
Fig. 4.1 Graph of SR(α) 76
Fig. 4.8 Interstorey drift ratio envelopes of four groups of ground motion 81
ix
Fig. 4.13 Comparison of maximum interstorey drift ratio
by proposed formula and RHA of the first building 85
Fig.4.14 Comparison of maximum interstorey drift ratio
by proposed formula and RHA of the second building 85
Fig.4.15 Comparison of maximum interstorey drift ratio
by proposed formula and RHA of the third building 86
Chapter 5
Fig. 5.1 Strain diagram of a rectangular section 106
Fig. 5.2 (a) Continuum‐MDOF model;
(b) Section of the flexural cantilever 106
Fig. 5.3 Stress strain relationship (a) Fibres in the flexural cantilever
(b) Shear cantilever 107
Fig. 5.5 (a) Excitation from PEER with record ID: P1121;
(b) The roof displacement under the excitation in (a);
(c) The roof displacement under the excitation two times of (a);
(d) The roof displacement under the excitation three times of (a)108
Fig. 5.7 Capacity curve of building α =1.6
(a) Inverted triangular loading (b) Uniform loading 110
Fig. 5.8 Capacity curve of building α =3.0
(a) Inverted triangular loading (b) Uniform loading 111
x
Fig. 5.9 The capacity curve of building α =9.4
(a) Inverted triangular loading
Fig. 5.10 Model with α = 1.6 (a) Roof drift;
(b) Interstorey drift ratio; (c) Curvature 113
Fig. 5.11 Model with α = 3.0 (a) Roof drift;
(b) Interstorey drift ratio; (c) Curvature 114
Fig. 5.12 Model with α = 9.4 (a) Roof drift;
(b) Interstorey drift ratio; (c) Curvature 115
Chapter 6
Fig. 6.1 Generalised roof drift spectra (a) and (b) (Case 1) 129
xi
Fig. 6.15 Generalised interstorey drift ratio spectra (e) and (f) (Case 2) 143
Fig. 6.26 Mean pseudo‐ acceleration spectrum
with 5% viscous damping of the ground motions in Table 6.1 153
Fig. 6.28 Roof drift under response history analysis
and the prediction from the method
of generalised inelastic response spectra 154
Fig. 6.29 Interstorey drift ratio under response history analysis
and the prediction from the method
of generalised inelastic response spectra 155
Fig. 6.30 Curvature under response history analysis
and the prediction from the method
of generalised inelastic response spectra 155
xii
Appendices
Fig. A1 Free body diagram of flexural‐shear cantilever
at time t under free vibration 164
List of Tables
Table 3.3 Comparison of fundamental periods T1 predicted
by proposed formula and finite‐element analysis 49
xiii
INELASTIC DRIFT‐BASED APPROACHES FOR PRELIMINARY
SEISMIC ASSESSMENT OF BUILDINGS
Department of Civil and Environmental Engineering
The Hong Kong University of Science and Technology
Abstract
There is no provision for consideration of seismic resistance in building design codes of
practice in low-to-moderate seismic regions in this world. This could cause a significant
loss of life and economic hardship under a moderate earthquake. This is due to the high
population density and the extensive infrastructure in the urban areas. Thus there is an
urgent need to implement a seismic assessment scheme for building structures composed of a
preliminary assessment and a detailed assessment. In this thesis the methods are proposed
for the preliminary seismic assessment which is a crucial step and a screening procedure for
identifying deficient building structures from a pool of buildings for more detailed
assessment.
Existing preliminary seismic evaluation procedures are mainly for low to medium rise
building structures, in which shear deformation is dominant. Thus an evaluation procedure
for high rise building structures is proposed with consideration of inelastic behaviour and
higher vibration mode effect. A method of analysis for estimating quickly the natural
fundamental period of vibration for tall building structures is first proposed from the continuous
flexural-shear cantilever. Then an inelastic interstorey drift ratio model is derived from the
principle of capacity spectrum analysis and modified modal pushover analysis for the assessed
buildings. Numerical investigations on the representative tall wall-frame structures show that
the seismic responses of the buildings predicted by the proposed inelastic interstorey drift
ratio model agree well with those obtained from response history analysis.
xiv
A systematic and effective assessment procedure for the preliminary assessment is
developed. The proposed procedure first divides the assessed buildings into either elastic or
inelastic one using a bifurcation index. For building structures that deform inelastically may
be further examined by the inelastic interstorey drift ratio model. Performance of building
structures is then defined based on the interstorey drift ratio by comparing those given in the
current codes of practice and the literature. The proposed model is particularly suitable for
assessing the buildings that are analysed in frequency domain.
xv
Chapter 1
Introduction
Earthquakes, unpredictable and unavoidable, have brought catastrophic losses in the economy
and precious human lives. A satisfactory explanation of earthquakes, Elastic Rebound Theory,
was proposed by Henry Fielding Reid to explain the mechanisms of an earthquake after the
event of great 1906 San Francisco earthquake. Unfortunately, knowing of earthquake
mechanisms does not help to predict the occurrence of an earthquake as reliably as the weather
forecast. Prediction of the occurrence of an earthquake is difficult, but potentially dangerous
regions can be identified through the locations of fault lines. Regions of different seismicity
are classified based on the fault’s location and the probability of earthquake occurrence. Three
main regions having low, moderate and high seismicity have been defined.
1
Regions of low to moderate seismicity have a relatively lower probability of earthquakes
than those of high seismicity, but they are not entirely free of earthquakes. Little attention has
been paid on those regions, as the probability of earthquake occurrence is very small. An
earthquake can devastate a whole city if it happens in a region where the structures are not
designed explicitly against earthquakes. This is exactly the situation faced by regions of low
to moderate seismicity. Because of the low probability of earthquake occurrence, buildings
are only designed against the combined effect of static gravity and static lateral load,
particularly wind loads. Designs for static loads are completely different to designs against
earthquakes. The dynamic properties of the buildings which govern the safety of buildings
under earthquake actions are unknown in these designs.
Buildings with seismic design may not be guaranteed to behave as expected as the design
because of uncertainties in earthquakes and invalid assumptions in design. This leads to the
question on “how safe are code-conforming building structures?” being continuously asked (El
Howary and Mehanny, 2011). Therefore buildings without seismic design are undoubtedly
full of uncertainties under earthquake action because of a totally different design philosophy.
In low to moderate seismicity regions there are thousands of non-seismically designed
buildings that are unprotected from seismic loadings and it may be very dangerous to ignore
any vulnerable building in an urban area. An unexpected earthquake will trigger this danger
because damage of buildings and infrastructures are the major source contributing to the losses
in the economy and in life. Citing the 5.6-magnitude Newcastle earthquake in 1989 as an
example, significant loss of life (13 deaths) and economic hardship (A$4 billion loss) was
caused in a region of low to moderate seismicity.
Hong Kong is an example of such a region, in which buildings are not seismically designed and
even worse, it has a densely populated environment. The large population forces both
residential and commercial buildings to occupy higher space and leads to buildings to
2
commonly being 40 storeys or above. Every non-seismically designed tall building may be
regarded as a potential tragedy because the failure of one building in a dense environment may
lead to a worst case, due to a domino effect.
3
1.2 Objectives
The primary objective of this research is to develop a comprehensive and reliable seismic
assessment method that can be implemented into the preliminary seismic assessment procedure,
in order to conduct a rapid screening on a large number of building structures so as to identify
potential hazardous building structures for detailed assessment. To accomplish the primary
objective, the following secondary objectives are needed,
z To develop a practical, yet simple, engineering model for existing building structures
which could best predict their fundamental period and vibration mode shapes.
z To develop methods for predicting the inelastic interstorey drift ratio of the assessed
building structures based on two different analytical approaches depending on the
resources that can be provided in the assessment and the characteristic of the
building structure.
By fulfilling the objectives, the impacts that can be made practically and academically are
z To fulfil the urgent need of society in the management of earthquake risk that has not
been considered in regions of low to moderate seismicity.
4
1.3 Scope of thesis
This thesis is divided into seven chapters. Chapter 1 is the background of this research and it
describes the current situation of seismic assessment in regions of low to moderate seismicity.
The objectives and scope of the thesis are also addressed. In Chapter 2, some typical studies
and provisions in the codes of practice on both preliminary and detailed assessment are
reviewed.
A model for quick estimate of natural fundamental period of tall buildings is presented in
Chapter 3. It is based on the continuum modelling technique derived in detail, and a
numerical investigation is included. Chapter 4 proposes a method extending the previous
continuum modelling technique for building structures to predict the inelastic interstorey drift
ratio under ground motion based on capacity spectrum analysis. A procedure for rapid
assessment when implementing the proposed method is presented.
5
Chapter 2
Literature Review
In countries that are subjected to frequent earthquake threats like China, Japan and U.S.A.,
engineers have to consider earthquake induced forces at the stage of design. This is to ensure
that the structures would behave safely and achieve the desired performance under unexpected
ground motions at any time. Not only is a precaution policy conducted at the design stage, but
a seismic vulnerability assessment scheme is carried out to rehabilitate and strengthen historical
and unqualified buildings. Without doubt, seismic design and assessment are a must in strong
seismicity regions in order to manage earthquake induced problems.
6
which have not been designed against earthquakes, is a worrying problem. More and more
attention has been paid on the evaluation of those building structures, for instance the research
conducted by Lang (2002) to evaluate the seismic risk in Switzerland. To manage seismic risk
has also become a problem in low-to-moderate seismicity regions.
Seismic risk is closely related to the performance of buildings and hence seismic
vulnerability assessment of existing buildings is one of the major tasks in the risk management
policy. With increasing understanding in the structural performance and the increasing
computational power in personal computers, more in-depth research on seismic vulnerability
assessment of buildings is being continuously conducted. Research on seismic vulnerability
assessment in general is composed of two parts, which are the development of the engineering
models of the assessed buildings and the method of analysis.
The development of the engineering model changes from being just a simple enough
SDOF model with mathematical formulae to a more detailed MDOF model in finite-element
based software. In the meantime, the method of analysis has improved from linear static
analysis, which only predicts the maximum response, to nonlinear response history analysis,
which captures the changes of response in a period of time. In spite of the different models and
methods to evaluate the performance of buildings, the main aim of seismic assessment is to
assist practicing engineers to evaluate existing buildings in an organised, reliable, fast and cost
effective way.
7
Preliminary assessment (Level one)
The assessment is based on the observation of buildings and relies on the experience of
engineers. It is usually done by score assignment which involves little or no calculation and is
suitable for quick evaluation. This kind of assessment is also named as a walkthrough
assessment or rapid visual screening procedure because the assessment procedure is mainly
observational. Most of the time, the assessment is developed from models derived from statics.
The purpose of the assessment is to screen out deficient buildings from a larger number of
building structures. Although this kind of assessment is efficient in both time and cost, the
results are subjected to larger uncertainties relative to other in-depth assessments
Detailed assessment requires a full scale model of the assessed building, which includes
modelling of all the structural members and some necessary non-structural members, as well as
the material properties obtained from in-situ inspection. Individual members are represented
by the detailed model (Ile et al., 2000; Kim et al., 2005) thus that the members behave close to
the experimental findings or the real situation. All individual members are modelled carefully
and are combined to form the structural system. The large scale structural system is then
analysed by special methods which involve heavy computational effort, such as nonlinear static
analysis and nonlinear response history analysis using finite-element based software.
8
The assessment has to be done by engineers having a seismic design background and deep
understanding of structural behaviour. Because of the demanding characteristics of this kind
of assessment, it is only applied to very important buildings and buildings which are identified
as deficient buildings, after the preliminary assessment, for future retrofit and strengthening
works. The assessment is cost demanding and time consuming, but it gives results which are
more reliable than the other assessment levels. Since there are many existing methods of
analysis in earthquake engineering, some typical and representative methods of assessment are
selected and discussed in this chapter.
Modal analysis is a linear analysis, which is used to predict the responses of a multi-degree
of freedom (MDOF) model, which behaves linearly under the ground motion. It is a technique
to decompose an MDOF system into a corresponding single-degree of freedom (SDOF) system
for further analysis.
SDOF system
The simplest approach to study the dynamic behaviour of a structure under ground motion is to
idealise the structure into an SDOF system. It has a lumped mass m supported by a zero mass
member with stiffness k and damping coefficient c, as shown in Fig. 2.1. There is only one
single response in a single degree of freedom system, which is the relative lateral displacement
u (t ) in this case. The characteristic equation of the lateral displacement of the system under
ground acceleration u&&g (t ) is written as
9
k c
where the angular frequency is ωn = and the damping ratio is ξ = . The lateral
m 2mωn
displacement solved under a particular angular frequency in Eq. (2-2) is a function of time.
The response of an SDOF system under ground motions is therefore governed by the angular
frequency of the system.
The maximum lateral displacements under different angular frequencies are predicted and
the graph contains the maximum response at the corresponding frequency and is the called
response spectrum. The relative displacement spectra contain curves of the maximum relative
displacement and its corresponding frequency at different damping ratios. Once the
fundamental period of a structure is predicted, the maximum lateral displacement can be
obtained from the relative displacement spectra by assuming the assessed building structure
behaves as an SDOF system.
MDOF system
An SDOF system is simple enough for analysis, but it is too simplified and cannot well capture
the behaviour of a building structure. It is not appropriate to assume a building behaves as an
SDOF system because buildings are more complex so that a single response at the level of the
lumped mass is not enough to model the deformation of building with its height.
A multi degree of system, which is more detailed, is then used to model the building
structure. The simplest MDOF model of a five-storey building is shown in Fig. 2.2. There
are five nodes at the floor level where the floor masses are lumped and the responses are of
concern. The concerned responses are the relative lateral displacements at the floor levels.
When a building is modelled as a multi-degree freedom system similar to Fig. 2.2, the
characteristic equation of lateral displacement of the system ground acceleration is as follows:
M&x& + Cx& + Kx = − Mιu&&g (2-3)
where M is the mass matrix, C is the damping matrix, K is the stiffness matrix of the system, x is
the vector containing the concerned responses and ι is the influence vector describing the
distribution of inertia force.
For the system in Fig 2.2, the response vector x covers the five relative lateral
displacements at the floor levels. The mass matrix, damping matrix and the stiffness matrix
10
are square matrices with size 5μ5, consistent to the numbers of responses. The system has five
degree of freedom with respect to the number of responses in the system. An MDOF system
can predict the behaviour of buildings more closely because more responses are captured in the
system, such as rotation and axial deformation. However the matrix size of the system would
be larger and hence the computation effort for solving the characteristic equation Eq. (2-3)
increases.
For the simplest MDOF model, the concerned responses are usually the relative lateral
floor displacements. The response vector x in the characteristic Eq. (2-3) is solved by
separation of variables. The vector is expressed as the superposition of N modal contributions
which are consistent to the numbers of responses:
N
x(t ) = ∑φn qn (t ) (2-4)
n =1
where φ n is the shape function of the nth mode, N is the number of degree of freedom and
qn (t ) is modal coordinate.
The shape functions are solved by Eigen analysis, and an N-degree of freedom system
would have N shape functions. The shape functions have the orthogonality property which is
φnT KφrT = φnT MφrT = 0 . Because of this property, Eq. (2-3) can be simplified by the following
procedures. The characteristic equation after substitution Eq. (2-4) is given by
N N N
M ∑φn q&&n (t ) + C ∑φn q&n (t ) + K ∑φn qn (t ) = − Mιu&&g (2-5)
n =1 n =1 n =1
Multiplying both sides of the Eq. (2-5) with the shape function of rth mode becomes
φrT Mφr q&&r +φrT Cφr q& r +φrT Kφr q r = −φrT Mιu&&g (2-6)
Dividing both sides of the Eq. (2-6) with φrT Mφr , which is non-zero, becomes
characteristic equation of a single degree of freedom system with angular frequency and
viscous damping ratio calculated as follows
φrT Kφr
ωr = (2-8)
φrT Mφr
11
1 φrT Cφr
ξr = (2-9)
2 ωr φrT Mφr
The response of an MDOF system in Eq. (2-4) hence can be found by the sum of N response
vectors of SDOF systems in Eq. (2-7).
An MDOF system with N-degrees of freedom can be decomposed into N SDOF systems.
Among the N systems, the system with the smallest angular frequency ωr has the largest
participation factor Γr in Eq. (2-4) and hence has the greatest contribution in response.
Therefore most of the preliminary assessment methods estimate the seismic performance based
on the fundamental mode instead of all the modes to reduce computational effort. However
for particular systems that have non-uniform stiffness or unevenly distributed mass and
irregular geometry, etc., the contributions in response from the higher modes will be very
significant. The influence of higher modes has to be considered.
Response spectra developed from the SDOF systems allow a fast prediction of response by
graphical means. It is a powerful tool recording the maximum response of an SDOF system
with all combinations between the fundamental period and viscous damping. It can also be
used to predict the response of an MDOF system by modal analysis in the frequency domain.
For instance, an improved linear elastic procedure has been proposed (Gunay and
Sucuoglu, 2010). In the analysis, members that are expected to behaviour inelastically are
replaced by new members with reduced stiffness. The reduced stiffness of the member is
obtained by the spectral displacement under the equal displacement rule. The original model
is modified to a new linear system with reduced stiffness thus that it can be analysed through
modal analysis. The new linear system presents the inelastic behaviour of the original system.
12
2.3 Displacement coefficient method
Buildings may deform inelastically under ground motion, so a method for prediction of
inelastic displacement under ground motion is important in assessment. A building is a
multi-degree of freedom system, which can be simplified as an SDOF system by modal analysis.
The responses of the building are usually dominated by the first few modes, which have relative
larger participation factors in Eq. (2-4). It is assumed that the building would yield according
to the dominant mode, which contributes the largest part of the response. When a structure
yields, the responses from the higher modes are assumed to be negligible. The problem thus
becomes the prediction of the inelastic response of the dominant SDOF system.
If an SDOF system behaves linearly under ground motion, there are response spectra,
which allow a fast prediction of the response by graphical mean. Because of the convenient
nature and power of the response spectrum, various methods have been proposed to make use
of it for the prediction of the inelastic response. These methods involve multiplication of the
coefficients to the response spectra, to obtain inelastic responses. This is the displacement
coefficient method.
Miranda and Ruiz (2002) compared six different displacement coefficient methods to
predict the inelastic displacement demands of an SDOF system. The six methods are
classified into two distinct approaches. The first approach is to replace the inelastic SDOF
system with an equivalent linear SDOF system with a modified stiffness and damping ratio.
The response of the inelastic SDOF system is said to be equal to that of the equivalent linear
SDOF system; this is known as the equivalent linear method. The second approach multiplies
the response of linear SDOF system with an inelastic displacement ratio to obtain the response
of the inelastic SDOF system; this is the inelastic displacement ratio method.
the maximum lateral inelastic displacement demand to the maximum linear displacement
demand on the same SDOF system. Miranda (2000) determined the value of the inelastic
displacement ratio for SDOF systems with different fundamental periods. The maximum
13
lateral inelastic displacement, which is the nominator of the ratio, is determined from the
nonlinear response history analysis of an SDOF model with defined hysteretic behaviour. The
maximum linear displacement, which is the denominator, is determined from the linear
response time history analysis of the SDOF model.
A graph of the inelastic displacement ratio and the displacement ductility, which is the
ratio of the inelastic displacement to the yielding displacement, is plotted. Miranda discovered
that the graph of the inelastic displacement ratio can be described by an implicit function of
building’s period and displacement ductility, given by
−1
⎡ ⎛1 ⎞ ⎤
( )
Cμ = ⎢1 + ⎜⎜ − 1⎟⎟ exp − 12Tμ − 0.8 ⎥ (2-10)
⎣ ⎝μ ⎠ ⎦
where μ is the displacement ductility.
Eq. (2-10) is an implicit equation, in which the response of the SDOF system has to found
by iteration. The iteration is simple enough in that it can either be done by programming or
manually. Given the fundamental period T and the yielding displacement of an SDOF system,
a first guess of the displacement ductility is assumed and is input in Eq. (2-10) to calculate the
inelastic displacement ratio. If the new displacement ductility calculated from the first guess
of inelastic displacement ratio is consistent to the first guess of displacement ductility, the
solution is found. Otherwise the value of ductility is updated by the new inelastic
displacement ratio and becomes the second guess. The iteration process stops when the
inelastic displacement obtained from Eq. (2-10) is consistent to the guess with an acceptable
tolerance.
The methodology is started from the characteristic equation of the deflection of the inelastic
SDOF system, which is given by
14
4πξo 2π
&x& + x& + ( ) 2 f (x ) / ko = − &x&g (2-11)
To To
where ξ o is the inherent damping ratio, f ( x ) is the restoring force of the system and k o is
the initial stiffness of the system. Eq. (2-11) is compared to the characteristic equation of a
linear SDOF system, which is given as follows,
4πξ eq 2π 2
&x&eq + x&eq + ( ) xeq = − &x&g (2-12)
Teq Teq
The equivalent linear method tries to find the equivalent period and damping so that the
maximum x in Eq. (2-11) is equal to the maximum xeq in Eq. (2-12).
In the study of Lin and Miranda (2008), relationships between the inelastic system and
the equivalent linear system are established. A group of inelastic SDOF systems with
various periods is subjected to ground motion under same conditions as in response history
analysis. Then large scale trial and error is done to find out the equivalent system for each
inelastic SDOF system in the group. The combination of the equivalent period Teq and
equivalent damping ratio ξeq which gives the least discrepancy between the maximum
displacement of the equivalent system and the inelastic displacement allows the solution.
The original period and damping ratio in Eq. (2-11) are plotted against the equivalent
period and damping ratio in Eq. (2-12). Two simplified empirical formulas expressed in terms
of strength ratio R, which is the ratio of elastic lateral strength to yield lateral strength, are
derived from regression analysis.
n1
ξ eq = ξ 0 + ( R − 1) (2-13)
T0n2
Teq m1
= 1+ m2
( R1.8 − 1) (2-14)
To T0
where m and n is post yield stiffness ratio.
In the assessment, the fundamental period and the strength ratio of the SDOF model are
first found. Then the corresponding equivalent SDOF system can be defined from the Eqs
(2-13) and (2-14). Finally, the response of the equivalent linear SDOF system is obtained from
the displacement response spectra.
15
2.3.3 Conclusions
The displacement coefficient method is a convenient tool which allows a quick prediction of the
inelastic response of an SDOF system. It empowers the response spectrum of linear SDOF
systems so that it can be used to model nonlinear inelastic behaviour. Despite the benefits of
the method, the fundamentals of the method only give the response of SDOF systems and not
MDOF systems. Prediction based on the nonlinear response of an SDOF system is an indirect
approach to model the nonlinear response of an MDOF system.
Drift spectrum is a seismic assessment tool similar to the response spectrum. It is developed
either from continuum models, which usually have simple mathematical expression in
behaviour. Drift spectrum expresses the maximum interstorey drift ratio as a function of the
fundamental period. Unlike the response spectrum which just gives the maximum response of
the SDOF system, drift spectrum gives the maximum interstorey drift ratio of the model, in
which modal analysis has been carried out to give the resultant response of all the modes.
Interstorey drift ratio is the derivative of the lateral deflection of the continuum model.
When an MDOF model is used, the interstorey drift ratio is defined as the difference in
displacement between successive storeys over the interstorey height. It has been used as one
of the design criteria that have to be satisfied in tall building design. It becomes more
important in nowadays in performance-based design and assessment. The interstorey drift
ratio closely relates to the shear stress at the storey level and it has been widely used as an
indicator of damage. It is more related to damage than lateral displacement, so drift spectrum
is a more direct assessment tool than response spectrum.
The properties of the drift spectrum depend on the type of model for representing the
dominant deformation characteristic. A typical model for drift spectrum is the continuous
shear beam model in Fig. 2.3 (Iwan, 1997). It is a model dominated by shear deformation in
the fundamental mode. It enables a quick assessment of low-rise building structures, which
deform in shear as does a continuum shear beam.
16
A shear beam model would limit the use of the spectrum to low-rise buildings; thus a
continuum flexural-shear cantilever is adopted (Miranda and Akkar, 2006) as shown in Fig. 2.4.
The deformation characteristic of the model is governed by the degree of structural interaction.
By increasing the degree of structural interaction from nil to infinity, the model deflects from a
pure flexural to flexural-shear and from flexural-shear to pure shear. The spectrum developed
from this model is named the generalised drift spectrum. It is an improvement of the original
drift spectrum and it is more powerful.
Drift spectrum provides a direct approach to obtain the interstorey drift ratio by the
fundamental period. It facilitates performance-based assessment, which uses interstorey drift
ratio as the indicator. However the assessment is based on linear analysis and it cannot model
the inelastic behaviour of building structures.
The inelastic displacement method and the equivalent linear method are tools for assessment of
buildings based on the SDOF system. A conversion process from the SDOF system to the
MDOF system is necessary. Moreover, approximate methods based on SDOF most of the
time only suit buildings dominated by the fundamental mode. Tall and slender buildings
would experience a significant influence from higher vibration modes and they are not
appropriate for evaluation by approximate methods that only consider the fundamental mode.
To consider the higher mode influence, which is not included in the SDOF system, and a
give direct estimation of the assessed buildings, Sheikh (2005) proposed an empirical formula
to predict the maximum interstorey drift ratio. It is an extended displacement coefficient
method for MDOF systems. The equation of the maximum interstorey drift ratio is given by
RSD1
θ max = λ2λ1λavg (2-15)
Hb
where RSD1 is the response spectrum displacement of SDOF system of the fundamental mode,
H b is the height of the assessed building, and λ2, λ1, λavg are the drift multipliers.
The drift multipliers are determined from a building database which contains the responses
from a group of full scale computer models in finite-element based software. The drift
17
multiplier λ2 indicates the contribution of the higher vibration modes relative to the
fundamental mode. Sheikh correlated this higher vibration mode drift multiplier λ2 to the ratio
RSD2/RSD1 and proposed the following formula
λ2 = 2.0(RSD2/RSD1) + 0.35 (2-16)
From this formula, the second mode is assumed to be the main influential parameter to the
higher modes. The other drift multipliers are used to convert the maximum lateral
displacement of the SDOF system to the maximum interstorey drift ratio of the corresponding
MDOF system.
The predicted maximum interstorey drift θmax is further decomposed into flexural-rotation
component θflmax and shear-drift component θshmax. Damage which is mainly attributed to
shear-drift component can then be evaluated. This approximate method is conceptually good
because it tries to incorporate the influence of higher modes empirically from a building
database and also identifies the shear-drift component as an indicator for performance
assessment.
Recently Tsang et al. (2009) have made use of the Sheikh’s proposed formula of
interstorey drift ratio to develop a rapid assessment procedure, which can be implemented in the
spreadsheet analysis. The procedure considers the influence of the site characteristics and the
effects of soil amplification. Numerical studies have shown that site conditions significantly
influence the seismic responses of building structures. Earthquakes may excite building
structures into the inelastic range of response, and the inelastic effect on the building is
modelled by an equivalent damping ratio.
A seismic vulnerability assessment method for low-rise building structures was proposed by
Koru (2002). It was developed from a very large amount of post-earthquake data. Four
different methods were investigated with verification from post-earthquake data. Each
method involves only one indicator or criterion for defining the performance. The four criteria
are the following,
1. Roof drift
2. Base-shear strength coefficient
18
3. Shear strength coefficient
4. Priority index
It was found that the roof drift and priority index were more reliable as they successfully
identified most of the buildings damaged in the database. By modifying the assessment
method with suitable material properties and construction practice, the method was adjusted to
apply to the assessed region. Observations from statistical results, in which many structural
parameters like numbers of columns, area of columns, numbers of storeys, height of building,
were taken into consideration for establishing the relationship between the degree of damage
and the indicator.
The growing concern on the management of earthquake risk in Switzerland has initiated a
series of research studies. Part of the research is to establish a method to evaluate the seismic
vulnerability of existing buildings. A seismic vulnerability assessment method was then
proposed by Lang (2002). It was applied to masonry and reinforced concrete buildings whose
lateral resistance is mainly contributed by the walls.
19
capacity curves. The four capacity curves are then superposed to obtain the resultant capacity
curve of the building, like in Fig. 2.5.
Degrees of damage are related to the base shear at different stages of the capacity curve,
for instance the yielding of the first wall will initiate the first damage and then further damage
occurs at the yielding of the second wall. The stages of damage are identified in the capacity
curve of the building. The capacity curve is then converted into a vulnerability curve (Fig. 2.6)
expressed as the stage of damage and the spectral displacement. Once the spectral
displacement is known, the expected degree of damage can be visualised.
The proposed method is comprehensive and the vulnerability curve suggested by the
method facilitates assessment and the retrofit arrangements. The method involves a
reasonable amount of calculations which are mainly the section analyses of the walls. Because
of its analytical approach, which includes section analysis and identifying the force path, the
method is flexible to model buildings with different characteristics. The method however is
only applicable when the structural plan of the building is simple and the coupling between
components is not significant. Otherwise it would be difficult to establish the capacity curve
by an analytical approach.
Damage index is one of the indices based assessment. According to Kappos (1997), a damage
index consists of damage variables which quantitatively describe the degree of damage of local
elements or substructures or even the global structures. It is proposed in seismic engineering
because of its rationality in assessment. It is a number ranging from 0 to 1, which represents a
damage stage varying from no damage to failure or collapse.
Damage indices usually relate to the amount of inelastic deformation and cumulative
damage due to cyclic loading. A representative one is the Park-Ang damage index (Park and
Ang, 1985), taking the form of
D=
θ max
+β
∫ dE (2-18)
θu M yθ u
where θmax and θu are the maximum recorded rotations of a member and the ultimate rotation
20
capacity of the member under monotonic loading respectively. The second term is the ratio of
absorbed hysteretic energy ( ∫ dE ) to the product of moment at yielding and the ultimate
based on experiments. It represents the effect of cyclic loading on the damage of a structural
member.
A damage index commonly considers both the deformation and energy as the factors
governing the degree of damage. It is a good concept that not only includes the deformation,
but also the energy. Therefore, use of damage index to evaluate damage may theoretically be
better than use of interstorey drift ratio. On the other hand, use of damage indices is more
complex. Difficulties arise in determination of the effect of cycling loading β because there
exists only a few choices of β from regression analysis of limited kinds of structural members.
It is inaccurate to apply the results from a small group of data to a large variety of members in a
realistic building (Williums and Sexsmith, 1995).
Moreover, empirical factors used in damage indices always have a large coefficient of
variation because of large uncertainties in the prediction of cumulative damage.
Accumulation of uncertainties may occur because the combination of local indices from
components to intermediate indices of a substructure or entire buildings is usually done through
the summation of weighted local damage indices (Michael et al., 1989). The model of the
probabilistic density function could be a better approximation to the model uncertainties in the
relationship between the damage stage and the index (Stephens and Yao, 1986). Weighted
local damage indices may not be sufficient to reflect the local failure mechanism which causes
progressive partial collapse of a building.
Although the accuracy of the global damage indices and reliability of the methodology is
doubtful, there is great application potential and the concept of indices is good. Further
development of damage indices may reduce the uncertainties in application and make it simpler
to be applied in assessment of buildings. Research on damage indices is continuing and
further descriptions can be found in the existing literature (Kappos, 1997; Ghobarah et al.,
1999).
21
Apart from damage indices, there are assessment indices proposed by Thermou and
Pantazopoulou (2011). It is very similar to the method of assessment proposed by Koru (2002)
and Lang (2002). The method is developed for low-rise reinforced concrete buildings which
sway in ground motion mainly due to the deformation of vertical members. From static
mechanics, the area of the vertical floor members and the reinforcement ratio are used to
develop the composite index which is the governing parameter in the assessment method to
determine the interstorey drift capacity of the building. The vulnerability curves of the
composite index and the interstorey drift capacity under different magnitudes of peak ground
acceleration are compared to the elastic demand obtained from the original response spectrum
method. Based on the results, the performance of the building is determined.
The capacity spectrum method was first introduced in the 1970s and rapidly became a popular
seismic vulnerability assessment and seismic design tool. The development of the capacity
spectrum method is described by Freeman (1998, 2004). The principle of the method is to
compare the lateral resistant capacity of the building to the earthquake demand graphically and
the intersection is the response. For seismic evaluation, the building is said to be safe if the
lateral capacity exceeds the earthquake demand. For seismic design, it is necessary to build a
structure which has enough lateral resistant capacity and at the same time satisfies the
serviceability limit state design.
Fig 2.7 illustrates the principle of the capacity spectrum method. In the method,
earthquake demand is presented by spectra containing demand curves with different damping in
the format of spectral acceleration versus spectral displacement (ADRS). The amount of
damping depends on the hysteretic behaviour of the buildings and it can be related to the
equivalent ductility demand. The lateral capacity of the building is estimated by the nonlinear
static analysis, which is also called pushover analysis, to obtain the capacity curve.
The capacity curve, which is represented in roof displacement versus base shear, is then
converted to the ADRS format for comparison to the response spectrum. The capacity curve
in ADRS format is called capacity spectrum. Two spectra are superposed and the intersection
between them is the expected response, also called as the performance point. The performance
22
point has to satisfy the condition where the equivalent ductility corresponding to the demand
spectrum is equal to the ductility at the capacity spectrum.
The capacity spectrum method graphically presents the relationship between capacity and
demand allowing engineers to visualise the performance of a building. However, the use of
the damped elastic response spectrum as an equivalent inelastic response spectrum is doubtful
because of the unclear relationship between the damping and hysteretic energy (Fajfar, 1999).
It is also limited to buildings subjected to a small influence by higher vibration modes because
higher vibration modes are not considered in the analysis.
Pushover analysis
Pushover analysis is a powerful and popular tool nowadays to assist in the evaluation of the
seismic performance of a structural system. Because of its reliability in capturing the
behaviour of a building, it is widely accepted and promoted by researchers and authorities as a
standard for design and evaluation (ATC 40; FEMA 310; FEMA 356). The main purpose of
the pushover is to estimate the lateral resistant capacity of a building and to present the capacity
graphically for capacity spectrum analysis.
In the analysis of a multi-storey building, which is an MDOF system, inertia force on the
building induced by the earthquake is modelled as invariant distributed lateral force acting
along the building height. The magnitude of the force is continuously increased to simulate
the intensity of the ground motion and the responses are recorded during the process. With the
advanced computer software, the internal force and displacement in the members, substructures
and the entire building are observed. Hence the weaknesses and the deficiencies of the
structural system can be revealed. It provides guidance for strengthening and retrofitting.
By monotonically increasing the magnitude of the invariant force, a capacity curve under
the specific invariant force distribution is obtained. The curve is usually presented in base
shear and roof displacement. The invariant distributed force is usually derived from the
expected deformed shape of the building in the fundamental mode, which is not valid for a
building with a significant higher vibration mode effect or a building with asymmetric plan
23
shape or discontinuity in strength and stiffness. The limitations of pushover analysis have
been described (Krawinkler and Seneviratna, 1998; Huang and Kuang, 2010). There is also
debate on the shape of the distributed force because it reflects the expected failure mechanism
of the building and results in the capacity curve (Mwafy and Elnashai, 2001). Different
distributed forces would give different capacity curves.
The influence of each mode on the building is represented by its invariant force
distribution, which has identical shape to the mode shape. Then N times of pushover analyses
are performed with N different invariant distributed forces respectively. Peak responses
resulting from respective modes under the effective earthquake force are then combined
according to the combination rule to obtain the maximum responses of the building. Pushover
analysis of invariant force distributions of higher mode however are difficult to conduct
because of great numerical instability.
An improved approach is the modified modal pushover analysis (Chopra et al., 2004). In
the analysis, modal contributions of higher modes in the responses are assumed to be linear
elastic. Therefore only nonlinear static analysis is performed on the fundamental mode.
Numerical studies also found that the modified modal pushover is more applicable and gave
results close to the results from other rigorous analysis.
Nonlinear response history analysis differs from nonlinear static analysis (pushover) in the
simulation of earthquake load. Unlike the nonlinear static analysis modelling of earthquake
24
load as a monotonic increasing distributed lateral force, earthquake load in nonlinear dynamic
analysis is completely captured by using the ground motion record, which is a function of time.
This procedure is rigorous and involves less uncertainty because of using ground motion
records and the material hysteretic model, but it is a time consuming procedure due to the step
by step evaluation of the building responses in time domain.
This method is found to be very difficult to apply because the numerical convergence is
difficult to be achieved when a complicated material hysteretic model is used. If a full-scale
model is used in the analysis, the time of analysis is extremely demanding and it does not
guarantee the obtaining of results. Even though the prediction of building responses is more
reliable, the responses are sensitive to the characteristic of the ground motion record.
Responses only reflect the performance under a single earthquake and may not be consistent in
other earthquakes. This drawback can be overcome by averaging a group of ground motions.
The power of the IDA is the linkage to the reliability analysis which accounts for
uncertainties. However it is time consuming to perform such large amount of nonlinear
response history analysis. To overcome the deficiencies in incremental dynamic analysis, the
IDA is modified (Vamvatsikos and Cornell, 2006). The modification uses nonlinear static
analysis to approximate the results of nonlinear dynamic analysis (Sang and Chopra, 2006).
The IDA curves of an SDOF model, nonlinear dynamic analysis, can be estimated from a
quadrilinear capacity curve of an SDOF model obtained by nonlinear static analysis through
empirical equations. The IDA curves can be further used for estimation of the seismic demand
25
and capacity of MDOF systems (Vamvatsikos and Cornell, 2005). This new approach reduces
the computational effort, but it is then subjected to the limitations and drawbacks of pushover
analysis. Reliability is reduced because of more estimation being required.
FEMA (Federal Emergency Management Agency) 154 (2002) refers to “Rapid Visual
Screening of Buildings for Potential Seismic Hazards”. It is a standard rapid visual screening
(RVS) tool for evaluation of building’s seismic performance. It helps engineers who are
experienced in seismic design to identify potential hazardous buildings for further deep
evaluation through a rapid screening means. The RVS requires the engineer to conduct a
“sidewalk survey” of buildings based on visual observation of the buildings from the exterior,
and the interior if possible. Meanwhile, the engineer needs to complete a data collection form
and determine the structural score which is the index used for the identification of the seismic
hazard.
The RVS method is established based on past earthquake records in the United State and
the data on the performance of buildings in earthquakes. Since the seismicity of regions of the
United States is classified into high, moderate and low, based on ground motions having a 2%
probability of exceedance in 50 years, three different data collection forms as shown in Fig.
2.10 are then developed using the data from each of these regions. Thus, this method can be
applied to regions of different levels of seismicity, but it is limited to regions with similar local
construction practice. The extension of this method to other regions requires modifications by
empirical coefficients, capturing the difference in construction practice and site characteristics.
Fifteen general types of buildings are considered in the data collection form. For each
type of building, a basic structural score is assigned based on the damage factor established in
the ATC 13 (an Applied Technology Council report, Earthquake Damage Evaluation Data for
California). The basic structural score is related to the probability of major physical damage,
which is defined by damage exceeding 60 percent of the building value in ATC 13. In the
RVS procedure, the engineer first selects the type of assessed building, which is associated to
the basic structural hazard score, through identification of the structural lateral-load-resisting
system and the major structural materials. To complete the form, the engineer needs to walk
26
around the building to verify and update the building identification information such as its
occupancy, size, shape and the soil type it is constructed on. Also, sketches of the plan and
elevation, as well as a photo of the building, need to be attached to the form.
Then, the engineer has to select appropriate score modifiers shown in Fig. 2.10, which are
related to the observed performance attributes and are derived from statistics using expert’s
opinions. Finally, the score modifiers are added to or subtracted from the basic structural
hazard score to obtain the final structural score. The final score can be interpreted as having 1
event of major damage in 10final score events. The probability of major damage is hence equal to
1/10final score. For example, a final score equal to 3 implies that there is a probability of 1 in 103
(i.e. 1 in 1000) that the building will collapse if such ground motions occur. The higher the
score means the better the expected seismic performance.
Based on present seismic design criteria, a final score of 2 is suggested as a “cut-off” value.
Buildings having a final score higher than the cut-off score have acceptable performance in
earthquakes. Buildings failing to meet the cut-off score are temporarily treated as seismically
hazardous and need further detailed assessment. The cut-off score can also be raised based on
the required level of safety imposed by users or society.
FEMA 154 (2002) is comprehensive and rapid. It requires little resources and is very
useful to rank buildings according to the structural score. Because of these attributes, the
assessment is also subjected to large uncertainties. They are the subjective judgment that is
highly dependent on the experience of the engineer, unavailability of detail interior inspection
and statistical uncertainties in the method. To avoid misleading assessment, the method tends
to underestimate building performance and always stands on the safe side.
It is well agreed that structural asymmetry results in torsion on buildings, which may increase
deflections and stresses in members. The worst condition is the failure of a member due to
additional stress when translation and torsion happen simultaneously under ground motion.
An uncoupled period ratio was defined by Yoon and Smith (1995 a) as the pure translational
period to pure torsional period to determine the influence of torsion and translation. The ratio
27
was derived from wall-frame structures with the corresponding engineering model represented
by a shear-flexure torsion-warping cantilever which does not consider any interaction between
the systems.
The uncoupled period ratio was an indicator for the degree of translational and torsional
coupling. If the ratio was unity, the maximum degree of translational and torsional coupling
had a high possibility of occurring. A small ratio implies that the torsional influence can be
separately considered and the torsional influence is less significant than the translation. In a
further study by Yoon and Smith (1995 b) an improved method using the coupled period ratio
was proposed. The period ratio may become an indicator to assist engineers in determining the
significance of torsional influence in assessment.
Earthquake induced energy is mainly dissipated through structural damping and hysteretic
behaviour when the building has been excited into an inelastic response range. Hysteretic
dissipation ratio is defined as the ratio of the total hysteretic energy to the total energy
dissipated by the system in an earthquake. It measures the number and duration of yield
excursions in a system and hence reflects the damage potential for the system. A system
dissipating a high proportion of seismic energy through hysteretic response is more vulnerable.
In the study of Chandler and Correnza (1996), a system with large eccentricity dissipates larger
portion energy in viscous damping action as a whole, but it results in a high portion of seismic
energy being dissipated by particular elements, which is undesirable for seismic safety as local
failure arises.
Torsion may be highly coupled to translation and this makes assessment more complicated.
Although there are many studies concerning torsion, practical assessment methods mainly use
28
nonlinear static analysis. They are limited to particular types of structures and have to satisfy
certain assumptions. It is believed that torsion can be revealed from the response history
analysis, but it is difficult to be performed and the results are sensitive to the input model.
Torsion response in the inelastic range would be a challenging task in seismic engineering
especially in assessment. This is the reason why, in the current code of practice, the design of
a highly asymmetric-plan building is prohibited and detailed assessment is recommended for
such buildings.
Comprehensive and quick seismic assessment methods are necessary for minimising and
managing earthquake disaster risks because it is definitely not efficient to evaluate every
building with the same amount of effort and time. It is also not appropriate to retrofit and
strengthen every building against earthquake risk without assessment. The priority order of
strengthening deficient buildings depends on the question “how urgent is the situation in regard
to a particular building?”. Although concern on earthquake risk is rising, the answer is still in
not clear.
Existing comprehensive seismic assessment methods are limited to low-rise buildings and
elastic behaviour. Detailed analysis may be reliable, accurate and applicable for most kinds of
buildings, but they are rather demanding in computational time and professional knowledge in
regard to modelling and interpreting. Therefore, detailed analysis is not the appropriate first
step in the assessment of buildings. A comprehensive seismic assessment should first be
carried out to identify the urgency for detailed analysis.
29
Table 2.1 Assessment level
Level one Level two Level three
30
Shear beam
31
Fig. 2.5 Resultant capacity curve from superposition (Lang, 2002)
32
Fig. 2.7 Capacity spectrum method
33
Fig. 2.9 Fractile curves
(Vamvatsikos and Cornell, 2002)
34
Fig. 2.10 Data collection forms for the three designated seismicity regions
35
Fig. 2.11 Basic score and score modifiers
36
Chapter 3
Quick Estimate of Fundamental Period
of Tall Buildings
3.1 Introduction
In preliminary seismic design and seismic assessment, it is important to evaluate accurately the
natural fundamental frequency of vibration of a building structure, since it is directly related to
the corresponding seismic forces and deformations. Traditionally, the fundamental frequency
required in the analysis is generally estimated from empirical equations which are normally
expressed in terms of the height of a building only. For example, the natural fundamental
period of vibration for moment-resisting frames T1 = 0.1N sec., where N is the number of
storeys, where the effects of the interaction of different structural forms in the building,
rigidities of the structure and intensities of the gravity and applied lateral loads, etc., on the
vibration characteristics are not considered in the calculation. This will clearly result in an
inaccurate estimate of natural vibration frequency or period of a building.
37
Realising the consequences of inaccurate estimate of the natural frequency, improved
formulas have been derived. Rakesh et al. (1997) calibrated the period data of both reinforced
concrete and steel moment-resisting frames to produce two different sets of empirical formulas
depending on the construction materials. Later Goel and Chopra (1998) proposed another set
of empirical period formulae, including the upper and lower bound values for concrete shear
wall buildings, in which the influence of the height of building structure and the area of the
structural walls at the ground floor are considered. To obtain a more accurate estimate of the
vibration frequency, more structural parameters, such as the mass and the degree of structural
interaction, have been implemented to the period formula based on rationale engineering
models.
38
interstorey drifts, and the intensities of the applied loading. Numerical studies pertaining to
determining the fundamental periods of moment-resisting frames, shear walls and a wall-frame
structures show that the results from the proposed formula agree very well to the finite-element
analysis (FEA). The proposed analysis has been shown to provide a simple and quick, yet
accurate, means of determining the fundamental period of vibration for tall buildings behaving
elastically.
39
3.3 Dynamics properties of the model
A free-body diagram of the flexural-shear cantilever at time t under free vibration, with the
mass replaced by its inertia force is shown in Fig. 3.3. Consider the dynamic equilibrium of
both flexural and shear elements in the flexural-shear cantilever, so the two equilibrium
equations in horizontal direction can be developed.
∂u
Vs = −GAχ = −GA (3-5)
∂x
in which M is the bending moment and χ is the shear strain. By combining the equations of
equilibrium of the elements represented in Eqs (3-2) and (3-3) and then substituting the shear
forces given by Eqs (3-4) and (3-5), the governing equation of motion of the continuum
flexural-shear model under free vibration can be derived and simplified as
∂ 2u ∂ 4u ∂ 2u
m 2 + EI 4 − GA 2 = 0 (3-6)
∂t ∂x ∂x
where m = mf +ms is the distributed mass.
where φ(x) is the shape function and q(t) is the modal coordinate, with different independent
variables. Eq. (3-6) can be solved by the technique of separation of variables, and thus
decomposed into two equations with variables x and t respectively,
40
q′′ ( t ) + ω 2 q ( t ) = 0 (3-8)
EI ( 4) GA
φ ( x) − φ ′′ ( x ) − ω 2φ ( x ) = 0 (3-9)
m m
where ω is the natural vibration frequency of the continuum model.
The fourth-order ordinary differential equation (3-9) can be solved by considering four
appropriate boundary conditions, resulting in equation Eq. (3-10) that contains the natural
vibration characteristics of the flexural-shear cantilever, has to be satisfied under the fixed-base
condition,
⎛ α4 ⎞ α2
2 + ⎜ 2 + 2 2 ⎟ cos γ i cosh β i + sin γ i sinh β i = 0 (3-10)
⎝ βi γ i ⎠ βi γ i
where
1
γ i2 = ⎡ −α 2 + α 2 + 4mH 4ωi2 / ( EI ) ⎤ and βi2 = α 2 + γ i2
2⎣ ⎦
The natural periods and shape functions corresponding to the characteristic equation (3-10) are
then given by
m 2π
Ti = H2 (3-11)
EI γ (γ + α
i
2
i
2 2
)
⎛ x⎞ ⎛ x⎞ ⎛ x⎞ ⎛ x⎞
φi ( z ) = Ci1 cos γ i ⎜ ⎟ + Ci 2 sin γ i ⎜ ⎟ + Ci 3 cosh β i ⎜ ⎟ + sinh βi ⎜ ⎟ (3-12)
⎝H⎠ ⎝H⎠ ⎝H⎠ ⎝H⎠
in which Ci is a coefficient composed of α, βi and γi.
It is well known that the flexural rigidity of a building has a significant effect on the natural
fundamental frequencies of vibration. A free body diagram of a flexural-shear cantilever
subjected to an inverted triangular load with a maximum intensity of wo, which gives the
smallest practical lateral load, is shown in Fig. 3.4. The governing equation with respect to
lateral deflection can then be derived and is given in a similar form to that Eq. (3-6),
∂4 y ∂ 2 y wo x
EI − GA = (3-13)
∂x 4 ∂x 2 H
where y is the static lateral deflection. With four appropriate boundary conditions, Eq. (3-13)
can be solved and the lateral deflection under the fixed-base condition is given by
41
wo H 4 ⎡⎛ α sinh α + 2 sinh α ⎞ ⎛ αx ⎞
y ( x) = ⎢⎜ 2 cosh α − α cosh α ⎟ ⎜ cosh H − 1⎟
EIα 4 ⎣⎝ ⎠⎝ ⎠
(3-14)
αx α2 ⎛ x ⎞ ⎛α2 ⎞ x ⎤
3
⎛1 α⎞
+ ⎜ − ⎟ sinh − ⎜ ⎟ +⎜ − 1⎟ ⎥
⎝α 2 ⎠ H 6 ⎝H⎠ ⎝ 2 ⎠ H ⎥⎦
The flexural rigidity EI is directly related to the design roof drift ratio RΔ, which is the ratio
of y(H)/H, and can be obtained by substituting x = H into Eq. (3-14) for derivation,
Substituting Eq. (3-15) into the equation of natural periods of vibration, Eq. (3-11), and
only considering the first mode of vibration (i = 1) lead to the following formula of the natural
fundamental period
mH
T1 = f (α ) i =1
RΔ (3-16)
wo
where
2πα 2
f (α ) i =1
= (3-17)
⎛α sinh α
2
α sinh α + 2 ⎞
γ 12 (γ 12 + α 2 ) ⎜ + −
⎝ 3 α cosh α 2 cosh α ⎟⎠
f(α) is a function of the degree of structural interaction between the flexural and shear
cantilevers and considered as an index of structural interaction. The value of f(α) varies
between 5.9 and 6.8 , as shown in Fig. 3.5, if only the first mode of vibration (i = 1) is considered,
giving a maximum difference of 15%, which indicates the fundamental period T1 can be
changed up to 15% due to the structural interaction.
42
3.5 Prediction of degree of structural interaction
The degree of structural interaction between the flexural and shear cantilevers, α, may be
calculated using Eq. (3-1) by identifying the flexural and shear rigidities of structural systems.
A suggested approach was proposed by Smith and Coull in their book “Tall Building Structures:
Analysis and Design”. The flexural rigidity of a building structure can be regarded as the total
flexural rigidities of the shear walls and the related components, which is
n
EI = ∑ Ei I i (3-18)
i =1
where n is the number of walls, E is the elastic modulus of the material and I is the moment of
inertia of the section.
The shear rigidity of a building is contributed from the moment resisting frames
constructed by columns and beams. The equation of shear rigidity is
12 E
GA = (3-19)
⎛1 1⎞
h⎜ + ⎟
⎝B C⎠
N
⎛I ⎞ M
⎛I ⎞
where h is the storey height, B = ∑ ⎜ b ⎟ , C = ∑ ⎜ c ⎟ ,N is the number of beams, M is the
i =1 ⎝ L ⎠ i i =1 ⎝ h ⎠ i
number of columns, L is the length of the beam. Ib and Ic are the moment of inertia of the beam’s
section and column’s section, respectively. Normally, the exact different lateral systems in a
building are very difficult to be identified due to structural couplings in the systems. Eq. (3-1)
would not be recommended generally for determining the value of α . Deflection, which
results from the action of external forces under the resultant lateral system, could be used to
predict the degree of structural interaction.
43
The deflection of the continuum model under uniform load shown in the appendix A2 is
wH 4 ⎡α sinh α + 1 ⎛ αx ⎞ αx 2⎛ x x 2 ⎞⎤
y(x ) = ⎢ ⎜ cosh − 1⎟ − α sinh + α ⎜
⎜ H 2 H 2 ⎟⎟⎥
− (3-20)
EIα 4 ⎣ cosh α ⎝ H ⎠ H ⎝ ⎠⎦
where w is the magnitude of the uniform load. The roof displacement is
wH 4 ⎡ α 2 α sinh α + 1⎤
y (H ) = ⎢1 + − ⎥ (3-21)
EIα 4 ⎣ 2 cosh α ⎦
A dimensionless expression which is the ratio of deflection to the roof displacement is given by
α sinh α + 1 ⎛ αx ⎞ αx ⎛ x x2 ⎞
⎜ cosh − 1⎟ − α sinh + α 2 ⎜⎜ − ⎟
y (x ) cosh α ⎝ H ⎠ H ⎝ H 2 H 2 ⎟⎠
= (3-22)
y (H ) α 2 α sinh α + 1
1+ −
2 cosh α
Eq. (3-22) is the deflection shape of the model where the roof displacement is adjusted to be
unity. If the model is identical to the building, the difference in deflection shape between the
model and the building should be zero. This is equivalent to the following mathematical
expression
n
y ( xi ) Y ( xi )
Δ=∑ − =0 (3-23)
i =1 y (H ) Y (H )
where Δ is the sum of the absolute differences, Y is the deflection of the building and i is the
floor number.
Since it is a very strict constraint where Δ is equal to zero, a more appropriate approach is
to minimise Eq. (3-23) by changing the degree of structural interaction. Although this method
would give a degree of structural interaction that can best describe the mode of deflection of the
building, it is not possible to obtain displacement of every floor. An alternative is to obtain the
displacement from the corresponding multi-degree-of-freedom (MDOF) model of the building.
Since buildings are usually designed in a finite-element-based program, the particular
displacement can be extracted from the model for analysis.
When the corresponding MDOF model of the building is not available, a more convenient
way is to approximate the value of α from the deformation profile of the building, which is
characterised by RΔ/Rδm, a ratio of the design roof drift ratio to the design maximum interstorey
drift ratio of the building. When a building is subjected to an inverted triangular load wo, the
interstorey drift ratio can be obtained by differentiating Eq. (3-14),
44
wo H 3 ⎡⎛ α sinh α + 2 sinh α ⎞ αx
Rδ ( x ) = 3 ⎢⎜
− ⎟ sinh
EI α ⎣⎝ 2 cosh α α cosh α ⎠ H
(3-24)
α x α ⎛ x ⎞ ⎛ α 1 ⎞⎤
2
⎛1 α⎞
+ ⎜ − ⎟ cosh − ⎜ ⎟ + ⎜ − ⎟⎥
⎝α 2 ⎠ H 2 ⎝ H ⎠ ⎝ 2 α ⎠ ⎥⎦
The design maximum interstorey drift ratio Rδm is obtained by substituting xo into Eq. (3-24),
where xo is the height at which the maximum interstorey drift occurs. The variation of α
against the drift ratio RΔ/Rδm is plotted in Fig. 3.6.
At the preliminary design stage, the roof drift ratio RΔ and the maximum interstorey drift
ratio Rδm may be assumed as H/500 and h/350 respectively, where h is the storey height, and are
consistent with those recommended in design codes of practice. The value of α may then be
quickly determined using Fig. 3.6. With the given value of α, the index of structural interaction
f(α) is found in Fig. 3.5. Hence, the natural fundamental period T1 can be conveniently
calculated using Eq. (3-16).
It is seen from the equation of the natural fundamental period, Eq. (3-16), that when this
equation is used to determine the value of T1 of a building, the effects of the structural
interaction, roof and interstorey drifts, gravity and lateral loads, and height of the structure on
the fundamental period of vibration have been considered.
A constant degree of structural interaction implies uniform rigidity along the height and it is an
approximation to general buildings which are not uniform. The changes of stiffness along the
height most of the time is not significant because the sizes of the structural components are
restricted to provide minimum strength in the design. Significant change in stiffness only
occurs in building structures with transfer storey. A 60-storey building with non-uniform
stiffness and a transfer storey shown in Fig 3.7 is analysed to study the influence of a constant
degree of structural interaction on the response. The building is constructed by three main
floor plans. They are the floors below the transfer storey, the transfer storey and the floors
above the transfer storey. The main differences between the three floor plans are the number of
columns, the arrangement of columns, and dimensions of the columns and beams at the
periphery.
45
Since both fundamental period and the degree of structural interaction are the key
parameters for a continuum model, they have to be similar to those of the equivalent MDOF
model, so that they are equivalent to the MDOF model. A method for constructing a
continuum model from an MDOF model is proposed. The flexural rigidity and shear rigidity,
which construct the degree of structural interaction, are estimated from the deflection
characteristic of the equivalent MDOF model. The deflection of the continuum model under a
point load is given by (Appendix A4),
PH 3 PH 2 PH 3 αx PH 3 αx
y (x ) = tanh α − x− tanh α cosh + sinh (3-25)
EIα 3
EIα 2
EIα 3
H EIα 3
H
where P is the magnitude of the point load at the roof.
Since the continuum model has to be equivalent to the MDOF model, the difference in
deflection between the two models should be zero. This is equivalent to minimising the
differences in deflection between two models by changing two variables: the flexural rigidity
(EI) and the degree of structural interaction (α).
n
Δ = ∑ y ( xi ) − Ym (xi ) (3-26)
i =1
where Ym is the deflection of the MDOF model. Since both models would have similar
fundamental frequency, Eq. (3-10) of the characteristic equation containing the fundamental
frequency should be held as a minimum, which is given by
⎛ α4 ⎞ α2
2 + ⎜⎜ 2 + 2 ⎟ cos γ cosh α 2
+ γ 2
+ sin γ 1 sinh α 2 + γ 12 = 0 (3-27)
⎝ ( )2 ⎟
α + γ1 γ1 ⎠
2 1 1
γ α + γ1
2 2
where
4mH 4ω12
−α2 + α4 +
γ 12 = EI (3-28)
2
When the flexural rigidity (EI) and the degree of structural interaction (α) are obtained, the
shear rigidity can be calculated from Eq. (3-1).
By the proposed method, the rigidities of different floor plans are estimated from uniform
models made with the respective floor plans. The rigidity of the equivalent continuum model,
which has a similar deflection and fundamental frequency, are also constructed. The variation
of rigidity in the MDOF model and the equivalent continuum model are compared in Fig 3.8. It
can be seen that there is a significant change in rigidity at the transfer storey and the change is at
46
least twice that of the other floors while the change of rigidity above the transfer storey is
gradual. The rigidity of the continuum model is close to that of the floors above the transfer
storey.
Both the MDOF model and the equivalent continuum model are then subjected to two
selected excitations in Fig 3.9 from the “PEER strong Motion Database”. The peak ground
acceleration of both ground motion records is approximately 0.3g. The deflection history of
both models is then observed at three different floor levels: 207.7 m (Fig. 3.10), 129.7 m (Fig.
3.11) and the transfer storey 17.5 m (Fig. 3.12).
The comparisons in Figs. 3.10 and 3.11 show that the deflection of the equivalent
continuum model is close to that of the MDOF model in both phase and magnitude. However
at the transfer storey, the continuum model could not capture the response of the MDOF model
as shown in Fig. 3.12. The reason for this difference is the abrupt change of the stiffness at the
transfer storey. When a non-uniform MDOF model is replaced by its corresponding
continuum model, the response of the equivalent continuum model is close to the MDOF model
at the region where the rigidity is similar, but at the region where rigidity changes abruptly, the
responses predicted by the continuum model may not be reliable.
Therefore, it is concluded that an equivalent continuum model could be used to predict the
response of general buildings which do not have significant change in rigidity. For building
structures with significant change in rigidity, the equivalent model could not give a reliable
prediction at the location where rigidity changes abruptly. Thus a constant degree of structural
interaction can be used to establish an equivalent continuum model that represents general
building structures.
The fundamental periods of vibration of three reinforced concrete tall buildings, including a
10-storey moment-resisting frame, a 20-storey shear-wall structure and a 30-storey wall-frame
structure, are calculated using the proposed formula and analysed by the finite element method,
employing ETABS (Computer and Structures Inc., 2005). The three buildings have the
identical storey heights of 3 m, and the floor plans are shown in Fig. 3.13. The dimensions of
47
structural members, as in the practical design, are given in Table 3.1, while the reinforcement
ratios of the members satisfy the strength and minimum reinforcement ratio requirements,
which are 0.25% for both the beam and wall sections, and 0.4% for column sections. The
design concrete cube strength is 30 MPa with density of 2400 kg/m3, and the reinforcement
yield strength is 460 MPa.
The three example buildings are designed to resist a dead load (DL) of 4.5 kN/m2, a live
load (LL) of 8 kN/m2, and lateral loads (WL) of 2.5 kN/m2, 3.0 kN/m2 and 3.5 kN/m2 at the roof
level respectively, which decrease linearly to 0 at the base level. Three load combinations are
considered in the analysis:
(1) 1.4DL + 1.6LL,
(2) 1.2DL +1.2LL + 1.2WL, and
(3) 1.0DL + 1.0LL + 1.0WL.
To evaluate the fundamental period of vibration T1 for the buildings using the proposed
formula given by Eq. (16), the structural parameters, including the degree of structural
interaction α and its corresponding index f (α ) i =1
, design roof drift RΔ and design maximum
interstorey drift ratio Rδm, as well as the distributed mass m, are first calculated, as shown in
Table 3.2. FEM models of the buildings are shown in Fig. 3.14.
Comparison of the results of the natural fundamental periods predicted by the proposed
formula and the finite-element analysis by ETABS is made and presented in Table 3.3. It can
be seen from Table 3.3 that the two sets of the results show very good agreement.
48
3.8 Concluding remarks
In this chapter, a model is developed for the quick estimate of the natural fundamental period of
vibration for tall building structures. The model is based on the continuum modelling, where a
tall-building structure is considered as a continuous, interactive flexural-shear cantilever. The
proposed model is presented in a closed form mathematical expression for use of calculating the
natural fundamental period T1, in which the effects of the interaction of different structural
forms in the building, design roof drift and maximum interstorey drift ratios, intensities of
applied loading and height of the structure on the fundamental period of vibration have been
considered. Numerical studies pertaining to predicting the fundamental periods of the tall
moment-resisting frames, shear walls and wall-frame structures show that the results from the
proposed formula and finite-element analysis agree very well. The proposed formula has been
shown to provide a simple and quick, yet accurate, means of estimating the fundamental period
of vibration for tall building structures, in particular at the preliminary seismic design and
seismic assessment stage.
49
Table 3.1 Dimensions of Structural Members
Column (m) Beam (m) Wall (m)
10-storey 30-storey 10-storey 30-storey 20-storey 30-storey
Floor frame wall-frame frame wall-frame shear walls wall-frame
st th
21 to 30 0.3 × 0.3 0.4 × 0.25 0.4
11th to 20th 0.4 × 0.4 0.5 × 0.3 0.3 0.5
st th a a
1 to 10 0.3 × 0.3 0.5 × 0.5 0.4 × 0.25 0.6 × 0.3 0.4 0.6
a st nd
In the 10-storey frame, columns at 1 floor and central columns at 2 floor are
0.35×0.35 m.
50
Axially-rigid links
Flexural cantilever
Shear cantilever x
α=0 α=∞
Height
Height
Deflection Deflection
51
∂M
M+ dx
∂x
∂V f
Vf + dx
∂x
ndx
∂ u
2
dx m f dx
∂t 2
Vf
∂Vs
Vs + dx
∂x
ndx χ
∂ 2u
m s dx
∂t 2
Vs
52
∂M
M+ dx
∂x
∂V f
Vf + dx
∂x
wo ( x / H ) ndx
dx Vf
∂Vs
Vs + dx
∂x
ndx
χ
Vs
53
6.90
6.80
6.70
6.60
6.50
f (α ) i =1
6.40
6.30
6.20
6.10
6.00
5.90
5.80
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
α
0.85
0.80
0.75
0.70
RΔ /Rδm 0.65
0.60
0.55
0.50
0.45
0.40
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
54
43.5 m
207.7 m
(a) Floor 1st to2nd
55
Variation of EI along the height Variation of GA along the height
EI Equivalent EI GA Equivalent GA
207.9 207.9
198 198
188.1 188.1
178.2 178.2
168.3 168.3
158.4 158.4
148.5 148.5
138.6 138.6
128.7 128.7
Height (m)
118.8
Height (m)
118.8
108.9 108.9
99 99
89.1 89.1
79.2 79.2
69.3 69.3
59.4 59.4
49.5 49.5
39.6 39.6
29.7 29.7
19.8 19.8
9.9 9.9
0 0
0 50000 100000 0.00 10.00 20.00 30.00
2
EI (GNm ) GA (GN)
Fig. 3.8 The variation of stiffness and the equivalent stiffness of the building
56
0.4 0.4
0.3 0.3
Ground acceleration (g)
0 0
-0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-0.4 -0.4
0 5 10 15 20 25 0 5 10 15 20 25
Time (s) Time (s)
(a) (b)
Fig. 3.9 (a) From PEER with record ID: P0714 (b) From PEER with record ID:P0410
At 207.7m At 207.7m
Displacement (m)
0.1 0.04
0.05 0.02
0 0
-0.05 -0.02
-0.1 -0.04
-0.15 -0.06
0 5 10 15 20 25 0 5 10 15 20 25
Time (s) Time (s)
(a) (b)
Fig. 3.10 Time history response at height of 207.7 m
At 129.7m At 129.7m
Displacement (m)
0.06 0.03
0.04 0.02
0.02 0.01
0 0
-0.02 -0.01
-0.04 -0.02
-0.06 -0.03
-0.08 -0.04
0 5 10 15 20 25 0 5 10 15 20 25
(a) (b)
Fig. 3.11 Time history response at height of 129.7 m
Displacement (m)
0.004 0.006
0.002 0.004
0.002
0 15.5
0
-0.002
-0.002
-0.004 -0.004
-0.006 -0.006
0 5 10 15 20 25 0 5 10 15 20 25
(a) (b)
Fig. 3.12 Time history response at height of 17.5 m
57
Beam
9m
Column
9m
Shear wall
2m 10 m
1m
10 m
Column
Beam
Shear wall
5m 15 m
15 m
58
(a) 10-storey moment-resisting (b) 20-storey shear walls (c) 30-storey wall-frame
frame
59
Chapter 4
Performancebased seismic assessment of
wall frame building structures using
inelastic interstorey drift ratio
4.1 Introduction
60
Preliminary assessment is usually based the methods with engineering models of building
structures, which are suitable for the assessment of a larger number of buildings. To conduct a
large scale assessment in regions or countries, preliminary assessment is a compulsory step
performed by practising engineers to prioritise the assessment order of building structures for
the detailed assessment. The detailed assessment, though more reliable, is cost demanding and
time consuming in nature. Therefore the detailed assessment is only conducted on building
structures that fail the preliminary assessment.
A preliminary assessment for larger numbers of low-rise building structures has been
proposed by Lang (2002) in Switzerland. Assessed buildings can be simplified by obtaining
the strength and ductility through study of architectural plans. Capacity spectrum analysis is
then conducted on the simplified buildings to predict the seismic responses, which can be used
as an indicator to define the performance. Representative buildings are selected for
conducting the analysis and producing the fragility curve, which expresses the possible damage
grade under the expected spectral displacement. Building structures in general are sorted to
conform to one of the representative buildings and have similar fragility curves to the
representative building. It cuts short the detailed analysis for preliminary assessment.
Recently Sheikh (2005) established fragility curves of building structures based on the
maximum interstorey drift ratio, which is determined from a model composed of several drift
multipliers. A group of building structures is first selected for analysis by constructing their
corresponding full-scale models in the finite-element-based program “ETABS” with linear
elastic material properties. The fundamental periods of the models are correlated to the height
of the building and hence are used to determine the corresponding spectral displacements. The
drift multipliers are used to modify the spectral displacement, thus the influence of the
modelling and the higher modes are included. The drift multipliers are determined and
calibrated from the responses of the full-scale finite-element models.
Existing preliminary assessments are usually developed either for low-rise building
structures based on inelastic analysis, or for high-rise building structures based on elastic
analysis. It is well known that elastic analysis is less favourable than inelastic analysis for
high-rise building structures in earthquake engineering because under seismic excitation,
buildings are to behave nonlinearly and probably inelastically. This is the reason why
nonlinear analysis with well defined material properties is preferred for seismic assessment.
61
Nonlinear analysis is, however, cost and time demanding. It is not suitable for preliminary
assessment which has to be quick and simple in nature.
A quick and simple performance-based assessment that is based on the inelastic interstorey
drift model is proposed in this study for preliminary assessment of high-rise building structures.
The assessment first systemically defines building structures into two groups based on their
seismic behaviour, either elastically or inelastically, through a bifurcation index. For building
structures that deformed inelastically, a further assessment step is conducted based on the
model of the inelastic interstorey drift ratio. A model of the inelastic interstorey drift ratio is
derived from the principle of capacity spectrum analysis and the modified modal pushover
analysis (Chopra et al., 2004) on the assessed building structures, which is modelled by the
continuum flexural-shear cantilever. Depending on the interstorey drift ratio, the performance
of building structures is defined by the current codes of practice and existing literature.
In the preliminary assessment, the performance and intensity of damage of the building
structure are usually correlated to the magnitude of deformation. Building structures are
guaranteed safe if they are elastic under ground motion. If inelastic deformation occurs in the
building structures during the excitation, permanent deformation and damage will result. If a
building structure is expected to deform inelastically, appropriate methods of assessment,
which considers inelastic behaviour, should be used.
To distinguish the seismic behaviour of building structures is the first step in preliminary
assessment for selection of the assessment method. An index, from the principle of the
capacity spectrum analysis, is defined as the ratio of the elastic spectra displacement of the 1st
mode Sd1 to the yield displacement Sdy. When it is larger than one, the assessed building
deforms inelastically. Otherwise, the assessed building deforms elastically under the given
ground motions and it is regarded as safe, with no significant damage ensuing.
62
4.2.1 Fundamental period
To estimate the elastic spectra displacement Sd1, the fundamental period of the assessed
building has to be known. A theoretical model for estimating the natural fundamental period
of vibration of building structures has been proposed in Chapter 3. The fundamental period is
derived from the continuous flexural-shear cantilever, which has various modes of deformation,
from flexure to shear, by changing the degree of structural interaction α of the continuous
cantilever. The period is given by
mH
T1 = f (α ) i =1 RΔ (4-1)
wo
where f(α)|i=1 is the index of structural interaction at the 1st mode (Fig. 3.5), m is the distributed
mass, H is the overall height, wo is the magnitude of the design inverted triangular lateral load,
and RΔ is the design roof drift ratio.
63
Celastic
Sd 1 = × T1 (4-3)
4π 2
where Celastic is the proportionality constant of the elastic viscous damping. By substituting Eq.
(4-1) into Eq. (4-3), the spectral displacement of the 1st mode at the corresponding period is
Celastic mH
Sd1 = × f (α ) i =1 RΔ (4-4)
4π 2 wo
From the SEAOC Blue Book (SEAOC, 2009), it is stated that the overstrength coefficient,
a, may vary from about 2.25 to 4.5. The result was obtained from a preliminary investigation
that had been conducted on various lateral systems, namely a bearing wall system, building
frame system and a moment resisting frame system. However, it also suggested that a
conservative value of overstrength coefficient should be either 2.0 or 2.5 depending on the
systems. For the preliminary assessment without performing a nonlinear static analysis, a
suggested value of overstrength coefficient could then be selected.
Building structures in general are designed against static lateral load and hence the design
strength can be approximated from the magnitude of the lateral load. Due to different lateral
patterns, it is not uncommon to consider a pattern that can result either the lower bound or the
upper bound of strength. The lower strength bound is obtained when the design lateral load
pattern is assumed to be an inverted triangular load with a maximum magnitude wo. The
design base shear of the building structure under this load pattern is
wo H
Vbase = (4-5)
2
When a nonlinear static analysis is conducted on the building structure, the elastic region of the
64
capacity curve can be modelled as linear, and yields at (autop, aVbase), including the effect of
overstrength a, where utop is the design roof displacement under the design base shear Vbase.
The capacity spectrum corresponding to the capacity curve gives the yield point expressed in
the format of spectral coordinate (Sdy , Say),
aRΔ H
Sdy = (4-6)
PF1
awo
S ay = 1
(4-7)
2mPF1 ∫ φ1 ( z )dz
0
where PF1 is the participation factor of the 1st mode, z (= x/H) is the dimensionless height, RΔ is
1
the roof drift (= utop/H) and ∫ φ (z )dz
0
1 is the integrand of the 1st mode with respected to z.
If the elastic spectral displacement is larger than the yield displacement, the assessed
building would deform inelastically and would result in damage and permanent deformation.
Therefore a bifurcation index (BI), defined as the ratio of the elastic spectra displacement to the
yield displacement, is established. The ratio of the elastic spectra displacement Sd1 to the yield
displacement Sdy is calculated from dividing Eq. (4-4) by Eq. (4-6) and is given by
S d 1 Celastic m
BI = = j (α ) i =1 (4-8)
S dy a wo RΔ H
PF1 f (α ) i =1
where j (α ) i =1 = ≈ 0.23 for α ∈ [0,20] When the value is larger than one, the
4π 2
building is damaged with unidentified intensity, which has to be defined by further assessment.
To evaluate the performance and the intensity of damage of a building structure, interstorey
drift ratio, which has been recognised as an important instrument for performance-based
seismic assessment, is predicted. The model is based on capacity spectrum analysis and the
principle of modified modal pushover analysis and the continuum technique on the
flexural-shear cantilever model, which may vary its mode of deformation from flexure
dominant to shear dominant. The effects of the structural height, interaction of different
structural forms in the building, top and interstorey drifts, intensities of applied design load, and
the properties of the demand spectra are considered in the inelastic interstorey drift ratio model.
65
4.3.1 Demand spectra on building structures
The demand spectra, which are plots of pseudo-acceleration under natural vibration periods, are
a series of curves representing different equivalent viscous damping of the model. When the
spectra is expressed in the format of the spectral coordinate, Eq. (4-2) becomes
Cζ2
Sa = (4-9)
4π 2 S d
With increasing equivalent damping ζ %, the magnitude of the demand decreases and so
does the proportionality constant Cζ. The relationship between equivalent damping and the
proportionality constant Cζ is modelled as
− d 2Cζ
ζ = d1e (4-10)
where d1 and d2 are the modelling parameters in the regression analysis, dependent on the given
set of ground motions.
S do = 0
(4-11)
2π 2 aw
Sao = S ay (4-12)
where Co is the proportionality constant of the demand spectrum at the performance point. It is
determined when the equivalent damping of the demand spectrum coincides with the equivalent
damping of the model.
The equivalent damping of the model is estimated from the equation proposed by Gulkan
and Sozen (1974). It was developed from the results of shake table experiments of small-scale
reinforced concrete frames. The equation has been compared to other empirical equations by
Miranda (2000) and it is showed that the equation gave estimations close to the others,
⎛ 1 ⎞⎟
ζ = ζ elastic + 0.2⎜1 − (4-13)
⎜ μ ⎟⎠
⎝
66
where ζ and ζelastic are the equivalent damping of the model at a particular μ , which is the
ductility at the performance point, and the elastic damping at the beginning, respectively.
To simplify Eq. (4-16), the 1st mode spectral displacement is extracted as follows
Sd 1
IDRmax ≈ (PF1φ1′(z )Sr )2 + (PF2φ2′ (z )r2 )2 + (PF3φ3′(z )r3 )2 (4-17)
H
where Sr is the inelastic displacement ratio, r2 and r3 are higher mode elastic displacement
ratios. The inelastic displacement ratio Sr is defined as the ratio of the performance point to
the elastic 1st mode spectral displacement,
S do Co2 1 m
Sr = = SR (α ) (4-18)
S d 1 Celastic a RΔ Hwo
67
1
2 PF1 ∫ φ ( z )dz
where SR (α ) = 0
(4-19)
f (α ) i =1
which is plotted in Fig. 4.1. The displacement ratio ri is the ratio of elastic spectral
displacements between the higher modes and the 1st mode, which is given by
ri =
S di
= 12 12
(
γ 2 γ 2 +α2 ) (4-20)
Sd1 (
γi γi +α 2 )
By substituting Eq. (4-4) of the elastic spectral displacement of the 1st mode and zo, where
zo is the dimensionless height at which the maximum interstorey drift ratio occurs, into Eq. (17)
of the interstorey drift ratio, the maximum value is expressed as
mRΔ
IDRmax = Celastic IR(Sr ,α ) (4-21)
wo H
For preliminary assessment, the proportionality constant Co is determined from Eq. (4-14) for
calculating the inelastic displacement ratio Sr in Eq. (4-18). With the inelastic displacement
ratio Sr, the index of inelastic displacement ratio IR(Sr, α) at particular structural interaction α
is found from Fig. 4.2. Hence, the maximum interstorey drift ratio IDRmax can be conveniently
determined from Eq. (4-21).
68
A performance-based preliminary seismic assessment serving the purpose of screening a
larger number of building structures is shown in Fig. 4.3. Building structures under the
expected ground motion intensity are first divided into two groups by the bifurcation index in
which building structures behave either elastically or inelastically. For building structures that
deform inelastically, meaning the occurrence of permanent deformation, a second step in
assessment is conducted based on the model of the inelastic interstorey drift ratio.
Performance of the building structure is then suggested from the deformation limits stated in
codes of practice or the literature.
To evaluate a larger number of building structures, building structures would undergo the
suggested procedure and is divided into two groups by the bifurcation ratio based on their
behaviour. Then a more detailed classification of performance is conducted based on the
prediction of the inelastic interstorey drift ratio by comparing to the deformation limits in codes
of practice and the literature. Three representative buildings are investigated by the suggested
procedure.
Non-seismic design is considered for the three buildings, which are designed to resist a
dead load (D) of 7.0 kN/m2, a live load (L) of 3.0 kN/m2, and lateral wind loads (W) of 2.88 kPa,
2.80 kPa and 2.58 kPa at the roof level respectively, which decrease gradually to 0 at the base
level. Three load combinations are considered in the design: 1.4D + 1.6L, 1.2D +1.2L + 1.2W,
and 1.0D + 1.0L + 1.0W. All the buildings satisfy the roof drift limit of H/500 in the load
combination: 1.0D + 1.0L + 1.0W. The roof drift under this load combination are 0.20 m, 0.14
69
m and 0.07 m. By the drift ratio RΔ/Rδm in Fig. 3.6, the degrees of structural interaction α of the
buildings are approximately equal to 1.6, 3 and 9.4.
The three buildings are then modelled in the nonlinear finite-element-based program
“OpenSees” for response history analysis. Because of the symmetry, the buildings are
converted into two-dimensional models shown in Fig. 4.5, which have similar fundamental
periods and elastic deflections under lateral load in the corresponding buildings, in order to
reduce computational effort. The resultant masses at the floor level of the models are 242 Mg,
274 Mg and 53.32 Mg respectively. The properties of the concrete and reinforcement are
shown in Fig. 4.6 with yield strengths o approximately at 28 MPa and 413 MPa. The
hysteretic behaviour is selected as isotropic hardening.
To study the seismic behaviour of the buildings, the ground motions are selected and given
in Table 4.4. The peak ground acceleration of the ground motion is artificially increased by
amplification scalars from 1 to 4 and it results in four groups of ground motion. The maximum
seismic responses, including the storey drift and the interstorey drift ratio under the four groups
of ground motion, are recorded and the maximum responses at each storey together produce the
seismic response envelope. Fig 4.7 and Fig 4.8 show the displacement envelopes and
interstorey drift ratio envelopes with the static response under the load combination 1.0D +
1.0L + 1.0W, respectively. The ductility demand which is defined as the ratio of the maximum
seismic response to the maximum static response in the loading combination 1.0D + 1.0L +
1.0W are plotted in Fig. 4.9 and Fig. 4.10.
The deflections and interstorey drift ratios in Fig. 4.7 and Fig. 4.8 have demonstrated the
deflection characteristics of the buildings with different degrees of structural interaction. The
three structures behave differently: the 30-storey and 25-storey wall-frame deflect flexurally,
while the 15-storey frame deflects in shear. With increasing peak ground acceleration, the
responses increase. The increments of the ductility demands shown in Fig. 4.9 and Fig. 4.10
are larger than those in the peak ground acceleration. It implies a nonlinear deflection
characteristic of the building structure under ground motion.
70
4.4.2 Performance-based assessment
The applied lateral force on the 30-storey wall frame is 2.88μ15=43.20 kN/m and the
distributed mass is 242/3.5=69.14 Mg/m. The lateral force on the 25-storey wall frame is
2.79μ12.5=34.88 kN/m and a distributed mass is 274/3.5=78.29 Mg/m. The 15-storey frame
is applied a lateral force 2.58μ9=23.22 kN/m and distributed mass of 53.32/3.5=15.23Mg/m.
Nonlinear static analysis is conducted on the models and the overstrength coefficients are
approximately equal to 3, 4 and 4 respectively. The bifurcation index of the 30-storey building
is calculated using Eq. (4-8)
9.63 242 / 3.5
BI1 = × 0.23
3 (2.88 × 15) × 0.0019 × (3.5 × 30)
= 2.09
Similarly the bifurcation index of the 25-storey building is calculated and equal to BI2 =
2.22. The 15-storey frame has bifurcation index of BI3 = 1.70. Since the bifurcation indices
of all buildings are larger than 1, they deform inelastically under the expected ground motion
and need further assessment.
71
Step 3 Inelastic interstorey drift ratio
The further assessment is conducted by predicting the inelastic interstorey drift ratios. The
corresponding indices of structural properties in Eq. (4-15) are firstly calculated. The index of
structural properties of the 30-storey wall frame building is
32 × 0.0019 × 105 × (2.88 × 15)
Bdg = = 1.12
(242 / 3.5)
From Fig. 4.12, when α = 1.6 and Bdg = 1.12, the proportionality constant Co = 6.4. From Fig.
4.2, SR(1.6) = 0.21. Hence, the inelastic ratio is given by Eq. (4-18),
6 .4 2 1 69.14
Sr = × × 0.21
9.63 3 0.0019 × 105 × 43.20
= 0.84
Then the index of inelastic displacement ratio and structural interaction given in Fig. 4.13 is
IR(0.84, 1.6) = 0.78. The maximum interstorey drift ratio determined by Eq. (4-22) is
69.14 × 0.0019
IDRmax = 9.63 × × 0.28
43.20 × 105
= 0.014
Similarly, the index of the structural properties of the 25-storey wall frame building is Bdg =
1.00 and the proportionality constant is Co = 6.5 found from Fig. 4.11. From Fig. 4.12, SR(3) =
0.216 and the inelastic ratio in Eq. (4-18) is Sr = 0.95. Therefore the index of the inelastic
displacement ratio and the structural interaction given in Fig. 4.2 is IR(0.95, 3) = 0.30, which
results the maximum interstorey drift ratio IDRmax = 0.014.
The index of the structural properties of the 15-storey frame building is Bdg = 0.43 and the
proportionality constant is Co = 6.8, found in Fig. 4.11. From Fig. 4.12, SR(9.4) = 0.223 and
the inelastic ratio in Eq. (32) is Sr = 0.82. Therefore the index of the inelastic displacement
ratio and the structural interaction given in Fig. 4.2 is IR(0.82, 9.4) = 0.28, which results the
maximum interstorey drift ratio IDRmax = 0.011.
72
Step 4 Performance
Comparisons between the maximum interstorey drift ratio predicted by the proposed inelastic
drift model and by the response history analysis (RHA) are made and presented in Figs. 4.13
and 4.14. It can be seen from the Fig. 4.13 that the proposed formula, which is presented by
the dotted line, gives slightly conservative predictions for the first building. This may be due
to the use of a smoothed demand spectrum, which is commonly used in preliminary assessment.
In Fig. 4.14, the proposed formula agrees through the results obtained from response history
analysis and also tends to give higher predictions in the second building. Similar observations
are also seen in the results of the 15-storey frame shown in Fig. 4.15. In general, the proposed
formula successfully predicts an expected conservative interstorey drift ratio of the assessed
building under the given set of ground motions.
The maximum interstorey drift ratios of the assessed buildings are compared in Table 4.1.
It can be seen that both structures belong to “Damage Control” from the suggestion of ATC 40.
A more detailed description is given in Table 4.2. Although the assessed buildings are safe,
structural components such as shear walls, are damaged with the concrete crushed and the
reinforcement exposed. If “Damage Control” is the minimum requirement for building
structures to pass the preliminary assessment, the assessed building is regarded as safe, based
on this performance, and the need for further detailed assessment is low.
73
4.5 Concluding remarks
A quick and simple performance-based assessment that is based on the inelastic interstorey drift
model has been proposed in this chapter for the preliminary seismic assessment of high-rise
building structures. The preliminary assessment first divides the assessed building structures
into either elastic or inelastic categories under the expected ground motion using the bifurcation
index, which is derived from the elastic spectrum displacement to yielded spectrum
displacement in capacity spectrum analysis. Building structures that deform inelastically are
further examined by the model of inelastic interstorey drift ratio.
The model of inelastic interstorey drift ratio is derived from the principle of capacity
spectrum analysis and the modified modal pushover analysis on the assessed building structures,
which is modelled by a combined continuous flexural-shear cantilever that can describe a wide
range of modes of deformation and hence capture the deflection characteristics more closely for
multi-storey buildings. The model considers the effects of the structural height, interaction of
different structural forms in the building, top and interstorey drifts, intensities of the applied
design load, and the demand spectra properties. Under the defined demand spectrum, the
inelastic interstorey drift ratio can be quickly calculated through a series of charts.
Performance of building structure is then defined based on the interstorey drift ratio by
comparison to the current codes of practice and the literature.
74
Table 4.1 Deformation Limits suggested in ATC-40
Performance level
Immediate Damage Structural
Interstorey Drift Limit Life Safety
Occupancy Control Stability
Maximum total drift a 0.01 0.01-0.02 0.02 0.33Vi/Pi c
b
Maximum inelastic drift 0.005 0.005-0.015 No limit No limit
a
Maximum total drift is defined as the performance point displacement
b
Maximum inelastic drift is defined as the portion of the maximum total drift beyond the effective yield point.
c
Vi is the total calculated lateral shear force in storey i and Pi is the total gravity load.
Table 4.2 Relationship between Structural Damage and Storey Drift (Li et al., 2006)
Description of structural damage Storey drift
Small cracks on columns in frames 1/1,000–1/1,300
A few number of small cracks on shear walls 1/1,100–1/1,200
Many through-cracks on shear walls 1/300–1/700
Shear walls damaged with concrete crushed and reinforcement exposed 1/80–1/200
75
0.230
0.225
0.220
SR (α )
0.215
0.210
0.205
0.200
0 2 4 6 8 10 12 14 16 18 20
α
Fig. 4.1 Graph of SR(α)
1
Sr = 2.5
0.8 Sr = 2.0
IR(Sr, α)
Sr = 1.5
0.6
Sr = 1.0
0.4
0.2
Sr = 0.2
0
0 2 4 6 8 10 12 14 16 18 20
α
Fig. 4.2 Graphs of IR(Sr, α)
76
A large group Bifurcation index
of buildings Safe
<1
>1
Gathering building
information
Substitution
Calculation
Classification of performance
77
30m
Death load: 7kN/m2
Live load: 3kN/m2 10m 10m
6m
24m
12m
α =1.6
25m
10m 5m
25m
5m
5m
α =3.0
18m
6m 6m
18m
6m
α =9.4
78
2.89kPa
Floor 30 Storey height = 3.5m
Floor 29
Floor 28
2.80kPa
….....
Floor 25
Floor 24
Floor 23
….....
….................
2.58kPa Floor 15
Floor 14
………. Floor 13
….....
…
α =1.6 α =3.0 α =9.4
Fig. 4.5 Models in “OpenSees”
(-0.0035, 27.6)
25 GPa
Concrete Steel
Fig. 4.6 Material properties
79
30 25
1 2 3 4 1 2 3 4
25
20
20
15
Story
Story
15
10
10
5
5
0 0
0 0.5 1 1.5 0 0.5 1 1.5
Displacement (m) Displacement (m)
(a) (b)
30-storey wall frame building 25-storey wall frame building
15
1 2 3 4 Legend
4. Group 4
5
0
0 0.5 1 1.5
Displacement (m)
(c)
15-storey frame
80
30 25
1 2 3 4 1 2 3 4
25
20
20
15
Story
Story
15
10
10
5
5
0 0
0 0.01 0.02 0 0.005 0.01 0.015 0.02
Interstory drift ratio Interstory drift ratio
(a) (b)
30-storey wall frame building 25-storey wall frame building
15
Legend
4. Group 4
1 2 3 4
5
0
0 0.01 0.02 0.03 0.04
Interstory drift ratio
(c)
15-storey frame
Fig. 4.8 Interstorey drift ratio envelopes of four groups of ground motion
81
1 1
0.9 1 2 3 4 0.9 1 2 3 4
0.8 0.8
0.7 0.7
Relative Height
Relative Height
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 1 2 3 4 5 6 0 2 4 6 8
Displacement Ductility Displacement Ductility
(a) (b)
30-storey wall frame building 25-storey wall frame building
0.9 Legend
0.8
1 2 3 4 -x- Static deflection
0.7
1. Group 1
Relative Height
0.6 2. Group 2
0.5 3. Group 3
4. Group 4
0.4
0.3
0.2
0.1
0
0 2 4 6 8 10 12
Displacement Ductility
(c)
15-storey frame
82
1 1
0.9 0.9
0.8 1 2 3 4 0.8
0.7 0.7
Relative Height
Relative Height
0.6 0.6 1 2 3 4
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 1 2 3 4 5 6 0 2 4 6 8
Drift Ductility Drift Ductility
(a) (b)
30-storey wall frame building 25-storey wall frame building
1
Legend
0.9
0.6
3. Group 3
0.5 4. Group 4
1 2 3 4
0.4
0.3
0.2
0.1
0
0 2 4 6 8 10 12
Drift Ductility
(c)
15-storey frame
83
12
10
8
Sa (m/s2)
7 .420
6 Sa =
T
0
0 1 2 3 4 5
Period T (s)
6.4
6.6
6.9
5.9
18
6.1
6.5
6.7
6
6.3
6.2
16
14
12
6.8
7
6.4
α
6.6
6.9
5.9
10
6.1
6.5
6.7
6
6.3
6.2
4
6.8
7
9
2 1 5.
6.4
6.6
6.
6.9
6
6.7
6.5
2
6.
6.
84
RHA Proposed model
0.018
0.016
0.014
Interstory drift ratio
0.012
0.010
0.008
0.006
0.004
0.002
0.000
0 1 2 3 4 5 6 7
Ground motion ID
Fig. 4.13 Comparison of maximum interstorey drift ratio by proposed formula and
RHA of the first example building
0.018
0.016
0.014
Interstory drift ratio
0.012
0.010
0.008
0.006
0.004
0.002
0.000
0 1 2 3 4 5 6 7
Ground motion ID
85
RHA Proposed model
0.018
0.016
0.014
Interstory drift ratio
0.012
0.010
0.008
0.006
0.004
0.002
0.000
0 1 2 3 4 5 6 7
Ground motion ID
86
Chapter 5
Modified ContinuumMDOF Model for
Seismic Analysis of Building Structures
5.1 Introduction
The linear continuum flexural-shear cantilever model has been widely used in seismic
engineering, as discussed in Chapter 2. An example is the model used in the development of
the generalised interstorey drift spectra (Miranda and Akkar, 2006). The drift spectrum has
provided an alternative way to the traditional response spectra, and an additional tool in
assessment, for engineers to predict the seismic response of a building structure in a fast and
simple manner. However, the drift predicted from a linear model is limited to the linear elastic
response range. Although increasing the equivalent viscous damping of the model can be used
to predict the inelastic responses, it is less favourable than response history analysis of a
nonlinear model with hysteretic behaviour. A nonlinear model is preferred for predicting
inelastic response.
87
A nonlinear simplified MDOF model for the estimating of the seismic response of regular
wall-frame buildings was developed by Huang (2009). The model is a modification of the
continuum flexural-shear cantilever model, in which the characteristic of the continuum model
is implemented into a finite-element-method-based computer programme to produce a
continuum-MDOF model. Compared to a detailed model with member-by-member
representation, the degree of freedom of the simplified model is less and hence the
computational demand for the analysis is smaller. Although the use of the continuum-MDOF
model is proposed for individual detailed assessment, it has a great potential in the
development of other types of assessment, especially for preliminary seismic assessment of
large number of buildings.
In preliminary seismic assessment, the time involved is very limited. At the same time,
the effort for analysis has to be kept at minimal as possible; thus the requirement of the
modelling would not be as strict as the modelling for the detailed assessment, for instance
referred to the noniterative equivalent linear method proposed by Lin and Miranda (2008).
Similarly, a continuum-MDOF model, which has several more advantages than the SDOF
model, can also be adopted for the preliminary seismic assessment of existing building
structures after modification. In this research, a modified continuum-MDOF model, which is
an improvement of the continuum-MDOF model for the preliminary seismic assessment, is
developed. Numerical investigations are carried out for three building structures. It is
shown that the modified model gives reasonably good predictions of the interstorey drift
ratios under response history analysis as compared to the corresponding detailed model with
member-by-member representation.
88
5.2 Modified continuum-MDOF model
Wall elements in the assessed building are combined to produce the flexural cantilever
based on the area of the wall elements. The hysteretic behaviour of the flexural cantilever is
then presented from the total steel ratio of the wall elements and the arrangement of the
reinforcement in the walls. The lumped masses at the nodes are obtained from the tributary
area of the floor. An initial value of shear rigidity is assumed for the shear cantilever so that
the initial fundamental periods of the continuum-MDOF mode can be calculated. Through
an iteration process, the most appropriate shear rigidity can be obtained by minimising the
difference between the period of the detailed model and the MDOF model. Lastly, an
appropriate hysteretic behaviour of the shear beam is selected.
It can be seen that the establishment of the continuum-MDOF model requires a detailed
MDOF model to be first developed or given. It greatly depends on the detailed MDOF model
or the structural plan of the assessed building. If neither piece of information corresponding
to the building is given, it would be difficult to establish the continuum-MDOF model.
Therefore it is only applicable in the detailed assessment in which the information has been
provided. For the preliminary assessment, an alternative approach has to be used; this is
because the corresponding detailed model including information of reinforcement details is
usually not available. A new approach is hence proposed for the preliminary seismic
89
assessment. It tries to approximate the behaviour of the assessed building structure based on
accessible information such as the design roof drift and maximum interstorey drift ratios,
intensities of applied loading and the height of the structure, etc.
The magnitude of the couple can be estimated from either the resultant force of the
compressive area or the resultant force of the tensile area, as they are identical. Consider the
resultant compressive force in the section,
Eεbd
F= (5-2)
4
where E is Young’s modulus of the material, ε is the extreme fibre strain, b is the width of the
section and d is the depth of the section. The moment arm of the couple is taken as the
distance between the centroid of the compressive area and that of the tensile area,
90
2d
R= (5-3)
3
Substituting Eq. (5-2) and Eq. (5.3) into Eq. (5.1) and simplifying gives
M = EI × φ (5-4)
where I is the moment of inertia and φ is the curvature, which are given by
bd 3
I= (5-5)
12
2ε
φ= (5-6)
d
To model the nonlinearity of the flexural cantilever, which represents a structural system, a
nonlinear moment and curvature relationship has to be defined in Eq. (5-4). However, the
flexural rigidity in Eq. (5-4) only concerns the properties of the material and the cantilever
section. To represent a structural system which behaves flexurally, a corresponding flexural
rigidity should be defined. Therefore when a flexural cantilever is modelled as a collective of
flexural systems in an MDOF model, a new flexural cantilever section is adopted.
The new section of the flexural cantilever allows easy nonlinear modelling of a flexural
system with known moment of resistance. It consists of two fibres with total area 2A, where A
is the area of the fibre as shown in Fig. 5.2(b). The centroid of two circular fibres is separated
a distance r and the centroid of the section is axially rigid, which does not allow the section to
have a net axial shortening. The expression for the moment of resistance of the new section is
different to the rectangular section defined in Eq. (5-4). The moment of resistance of the
section is given by
m = f ×r (5-7)
where f is the magnitude of force in either the compressive fibre or the tensile fibre. The
magnitude of force depends on the following
f = EεA (5-8)
where ε is the fibre strain. Substituting Eq. (5-8) into Eq. (5-7), the moment of resistance
becomes
m = EI s × φ (5-9)
where Is is the moment of inertia of the simplified section and φ is the curvature, which are
given by
91
Ar 2
Is = (5-10)
2
2ε
and φ = (5-11)
r
By using the new section of the flexural cantilever with a known moment of resistance, the
nonlinear flexural hysteretic behaviour of corresponding system can be modelled by the axial
stress-strain relationship of the fibre defined in Eq. (5-8).
The yield moment can be approximated from the magnitude and distribution of the
external force and the degree of structural interaction. For instance, when the lateral load is a
uniformly distributed load with magnitude w, the yield moment at the flexural cantilever and
the shear force along the height of the shear cantilever, which are derived and presented in
Appendix A2, are given by,
92
wH 2 ⎡ α sinh α + 1 ⎤
My = − 1⎥ (5-13)
α 2 ⎢⎣ cosh α ⎦
wH ⎡α sinh α + 1 αx αx αx ⎤
VS ( x ) = ⎢ sinh − α cosh + α − ⎥ (5-14)
α ⎣ cosh α H H H⎦
The maximum shear occurs at a particular model height which can be found conveniently in
spreadsheet analysis.
where RΔ is the roof drift ratio under uniform load w. The shear rigidity is then calculated from
the degree of structural interaction, which has been discussed in Chapter 3.
2
⎛α ⎞
GA = ⎜ ⎟ EI (5-16)
⎝H ⎠
The internal forces under an inverted triangular distributed load and point load are shown in the
Appendix A3.
93
cantilever, the maximum shear in the shear cantilever, and the stiffness of the model are
calculated as,
M y = 124129 kNm
Vmax (37 m ) = 2095 kNm
EI = 699222880 kNm 2
GA = 821943 kN
The corresponding continuum-MDOF model is then built with the above information as
shown in Fig. 5.2. Each cantilever consists of 26 nodes separated at a distance 3.5 m, which is
the interstorey height of the building. The flexural cantilever section consists of two fibres
with areas of A = 10 and separation of r = 20. The yield stress and Young’s modulus of the
fibres are calculated by Eq. (5-7), Eq. (5-10) and Eq. (5-15), given by
fy My
σy = = = 620.6kN
A Ar
EI 2 EI 2 × 699222880
E= = = = 349611.4kN/m
I Ar 2 10 × 202
The hysteretic behaviour of the fibre used in the flexural cantilever is bilinear with 2% isotropic
hardening as shown in Fig. 5.3(a). Similarly, the shear cantilever is determined as shown in
Fig. 5.3(b). Lump masses are then assigned on nodes of the continuum-MDOF model to
produce the desired fundamental period of the model, which is 2.0 s in this example.
The capacity curve of the continuum-MDOF model under a uniform point load at the
nodes is shown in Fig. 5.4. The continuum model deforms 0.175 m at the roof with base shear
70 × 87 .5 = 5985 kN ; however, the continuum MDOF model achieves the same roof
displacement with a smaller base shear 5741 kN due to the difference between the discrete and
distributed loads. A discrete loading in the continuum-MDOF model would create a larger
over-turning moment due to larger moment arm. Hence a smaller base shear results at the
target roof displacement. The difference can be minimised by increasing the number of nodes
of each cantilever, so that the discrete load is close to the uniform distributed loading, but this
will increase the computational effort.
94
Fig. 5.5(b, c, d) that increasing the magnitude of the ground acceleration would force the model
to behave more inelastically and is reflected in the permanent roof displacement after the
excitation. The inelastic behaviour of the model depends on the selected hysteretic behaviour
of the cantilever. Appropriate hysteretic behaviour can be used in the model.
Three modified continuum-MDOF models with bilinear isotropic hysteretic behaviour are built
to model the three representative buildings which have been discussed in Chapter 4. Their
degrees of structural interaction are 1.6, 3 and 9.4, respectively. The detailed models of the
representative buildings are then compared to their corresponding modified continuum-MDOF
models in nonlinear analysis for both the static and dynamic approaches.
In the static approach, both the detailed models and the modified continuum-MDOF
models have similar capacity curves under invariant lateral forces of the inverted triangularly
distributed loading and the uniformly distributed loading, as shown in Fig. 5.7 to Fig. 5.9. The
detailed models possess a slightly softening characteristic while the modified models have a
slightly strength hardening characteristic. This is due to the progressive yielding of the
members in the detailed model system and the softening characteristic in the materials.
Although there is a slight difference in the responses of the nonlinear static analysis, the
detailed models and the modified models are very much alike.
In the dynamic approach, both the detailed models and the modified models are analysed
by nonlinear response history approach. The maximum responses including the roof
displacement, the interstorey drift ratio and the curvature of the modified models, are plotted
against the responses of the detailed models as shown in Fig. 5.10 to Fig. 5.12. If the responses
of both models are close each another, the data will be located around the solid line which has a
slope equal to 1. If the predictions from the modified continuum-MDOF model are smaller
than those from the detailed model, the data would fall into the region above the solid line.
Otherwise, it would fall into the region below the solid line. For example, the data above the
solid line indicates an underestimated response by using the modified continuum-MDOF model.
The dotted lines, which are above the solid line, have gradient increases from 1.2 to 2
representing underestimation to 1.2 to 2 times.
95
In general, the curvature response of the modified model cannot be predicted well, which
is consistent with the findings of Huang (2009). However, it is also included in the numerical
study for the reason of understanding the behaviour of the modified continuum-MDOF model
in the nonlinear inelastic response range. Only the comparisons in the roof displacement and
the interstorey ratio are made in each modified model and the detailed model.
For the building with degree of structural interaction equal to α = 1.6, it is observed that in
the graphs of the roof displacement and the interstorey drift ratio, most of data are located
around the solid line. It indicates that the maximum responses from the modified
continuum-MDOF model under response history analysis are close to those from the detailed
model in terms of the roof displacement and the interstorey drift ratio.
For the building with degree of structural interaction equal to α = 3.0, the straight line
passing through most of the data in the graph of the roof displacement has gradient equal to 1.4.
Similarly the gradient of the straight line in the graph of the interstorey drift ratio is 1.2. It
indicates that the prediction of the modified model tends to underestimate the seismic response
of the roof displacement and the interstorey drift ratio.
Consider the data of the model with degree of structural interaction of α = 9.4. The
centroid lines of the data in the graphs of the roof displacement and the interstorey drift ratio are
below the solid line. The modified continuum-MDOF model overestimates the response of
the detailed model. The modified continuum-MDOF model is said to be conservative.
Common observations and conclusions can be drawn from the above comparisons.
Under the assumption of an isotropic hardening behaviour, a typical approximation without a
detailed investigation, the interstorey drift ratio of the detailed model does not differ from that
of the continuum-MDOF model significantly. The interstorey drift ratio is well predicted by
the modified continuum-MDOF model. In a comprehensive preliminary assessment, different
hysteretic behaviours can be selected in the analysis depending on the preference of the
engineer. Since the actual behaviour of an assessed building is always unknown before a
detailed investigation, which is both time and resource consuming, a simple but typical
behaviour like bilinear isotropic hardening behaviour can be used to give an approximate result
with a reasonable amount of effort.
96
5.4 General continuum representation
A full-scale detailed model can be constructed specifically for any assessed building in a
detailed assessment, but it is not usual in the preliminary seismic assessment because of limited
resources. Therefore a simplified model has to be used. The proposed modified
continuum-MDOF model is a good simplified model for the preliminary assessment of a large
number of building structures. A continuum-MDOF model can represent building structures
with various lateral deformation characteristics, from flexure to shear. Although the effort and
requirement for the establishment of the model and the analysis have been reduced for
individual building structures, it is still a considerable task to establish models and conduct
analyses for every building structure. It would be beneficial if a single model and its
analysis result could be applied to other building structures which have similarity to the single
model; for instance, having similar deformation characteristics and dynamic properties.
For a linear elastic SDOF model, its behaviour is governed by the fundamental frequency,
which is related to the ratio of stiffness to mass. For any two SDOF models, if their ratios of
stiffness to mass are identical though their values of stiffness and mass are not equal, they
would have identical fundamental frequency and an identical response under same ground
motion. An SDOF model is objective as its behaviour is only governed by an objective
parameter, which is the frequency. A linear elastic continuum model is subjective in contrast
to an SDOF model. This is because the responses of a continuum model depend on both the
fundamental frequencies and the height of the model. However, a dimensionless continuum
model, which has taken the influence of height into the mode shape, is objective. Any two
dimensionless continuum models with identical deformation characteristics and dynamic
properties would have identical responses.
A nonlinear inelastic model is characterised by its hysteretic behaviour and its behaviour is
governed by more parameters compared to linear elastic models. The response of a nonlinear
inelastic SDOF model depends on the fundamental frequency, the initial stiffness and the
force-displacement relationship. A nonlinear inelastic SDOF model is unique when it is
analysed in the time domain. Capacity spectrum analysis, which is a static analysis equivalent
to response history analysis, is less restrictive and a nonlinear SDOF model could be related to
a similar nonlinear SDOF model.
97
An SDOF model in capacity spectrum analysis is characterised by the capacity curve, the
demand curve and the equivalent viscous damping ratio. Since it is a graphical means of
analysis, the performance point of the model depends on the geometry of the intersection
between two curves and the viscous damping ratio. For any two SDOF models with
proportional geometry of the intersection, in capacity spectrum analysis, the two performance
points would be proportional under the same viscous damping. The two SDOF models are
similar because of geometric proportion in graphical analysis.
A continuum model, which can be decomposed into several SDOF models, can be related
to other models through the same approach as in the SDOF model. By comparison of two
different continuum flexural-shear models with known similarity, the relationship between
their responses is defined. A general continuum representation can then be established based
on the relationship, which describes how two similar models, the prime model and the sub
models, are related.
By considering two continuum models and two sets of parameters shown in Fig. 5.13, the
mathematical relationships between two similar models can be found under the above
conditions. The subscript notation “p” of the parameter refers to the parameters of the prime
model, while the subscript notation “s” refers to the sub model. The first four conditions are
expressed mathematically from Eqs (5-17) to (5-20), which can be found in the appendix.
98
GAp GAs
Hp = Hs (5-17)
EI p EI s
m p H p4 2π ms H s4 2π
=
EI p γ 2 (γ 2 + α 2 ) EI s γ 2 (γ 2 + α 2 )
m p H p4 ms H s4
= (5-18)
EI p EI s
3. Identical static roof drift ratio when their lateral yield strengths are reached under lateral
load, either in uniform or inverted triangular distributions,
yp ys
=
Hp Hs
w p H 3p ws H s3
= (5-19)
EI p EI s
ws H s wp H p y yp
÷ = s ÷
ms H s ∫ φdz m p H p ∫ φdz PF PF
ws m p y
⋅ = s
w p ms yp
ws m p H s
⋅ = (5-20)
w p ms H p
The above four conditions involve five pairs of variables. They are the height of the
model (H), the distributed mass (m), the magnitude of the lateral force (w), the flexural rigidity
(EI) and the shear rigidity (GA). From the fifth condition, the height of the sub model is
expressed in terms of the height of the prime model as
H s = kH p (5-21)
99
where k is an arbitrary proportional constant and is referred as the height ratio. By using Eqs
(5-17), (5-18), (5-19), (5-20) and (5-21), the parameters of the sub model are solved in as
follows,
ws = bkwp (5-22)
ms = bm p (5-23)
EI s = bk 4 EI p (5-24)
The mass-per-unit-floor area of different buildings is more or less the same, especially for
building structures with similar functions. For similar building structures, the difference in
mass is mainly due to the dimensions of the floor area. Therefore the difference in the
distributed mass along the height should be proportional to the floor area. From the fifth
condition, when the height is increased by a factor of k, the floor area should be increased by k2
as does the distributed mass. By assuming b = k2, Eqs (5-22), (5-23), (5-24) and (5-25)
become
ws = k 3 w p (5-26)
ms = k 2 m p (5-27)
EI s = k 6 EI p (5-28)
Two continuum flexural-shear cantilever models are said to be similar if their parameters
satisfy the above equations. By satisfying the above conditions, the relationship of the
responses between two similar continuum models are investigated.
wp H 3p ⎡α sinh α + 1 ⎤
IDR p ( z ) = 3 ⎢
sinh αz − α cosh αz + α (1 − z )⎥ (5-31)
EI pα ⎣ cosh α ⎦
100
wp H p2 ⎡α sinh α + 1 ⎤
curp ( z ) = 2 ⎢
cosh αz − α sinh αz − 1⎥ (5-32)
EI pα ⎣ cosh α ⎦
x
where z = is the dimensionless height of the model and z ∈ [0,1] . Based on the properties
H
defined in the Eqs (5-21) and (5-26) to (5-29), the static responses of two similar continuum
models under uniform lateral loading can be found and are related by,
ys ( z ) = k ⋅ y p ( z ) (5-33)
1
curs ( z ) = ⋅ curp (z ) (5-35)
k
The above relationships also hold for static responses under inverted triangular loading.
1
IDRp ( z, t ) = ∑φi′(z )qi (t ) (5-37)
Hs
1
curp ( z, t ) = ∑φ ′′(z )i qi (t ) (5-38)
H s2
where
q&&i + 2ξωi q&i + ωi2 qi = − PFi u&&g (5-39)
Since both the prime and the sub model have identical fundamental frequencies and degrees of
structural interaction, they would have same modal coordinate in Eq. (5-39) under the same
damping. By substituting Eq. (5-21) into Eqs (5-37), (5-38) and (5-39), the linear dynamic
responses of the prime and that of the sub models are as follows,
ys ( z , t ) = y p ( z , t ) (5-40)
1
IDRs ( z, t ) = ⋅ IDRp (z, t ) (5-41)
k
1
curs ( z, t ) = ⋅ curp ( z, t ) (5-42)
k2
101
5.4.3. Performance point of two similar models under capacity spectrum analysis
Response history analysis is regarded as one of the most reliable methods for seismic
assessment, but it depends on the modelling and the numerical method which cannot be
expressed easily. An equivalent analysis, capacity spectrum analysis, is used to approximately
predict the inelastic behaviour of the prime and the sub models.
From the conditions of similar models, the capacity spectrum of the prime and the sub
model are proportional to each another under the same secant stiffness. The proportional
constant is k, which is the proportional constant of the dimension or the height ratio of the
building structures stated in Eq. (5-20) and Eq. (5-21). If the performance point of the prime
model is (Sdo , Sa) under specific excitation u&& g , the performance point of the sub model would
be (kSdo , kSa) under same strength ratio or excitation ku&&g . It is explained in the followings.
When the performance point of the prime model is (Sdo, Sao), it has to satisfy two criteria,
1. The performance point is on the demand spectra, which is the maximum response of
the corresponding SDOF model under the ground motion at a particular angular
frequency and equivalent damping. The x-coordinate of the performance point is
S do = max[qo (t )] (5-43)
where
q&&o (t ) + 2ξ oωo q&o (t ) + ωo2 qo (t ) = −u&&g (t ) (5-44)
Eq. (5-44) is the equation of motion of the corresponding SDOF model; ξ o is the
2. The performance point is on the demand spectra and is also on the capacity spectrum,
hence the equivalent damping provided by the corresponding ductility on the
capacity curve is consistent to the demand curve in Eq. (5-45). That is
102
⎛ S do ⎞⎟
ξ eq = f ⎜ μ = = ξo (5-46)
⎜ S y ⎟⎠
⎝
where ξeq = f (μ ) is the function of damping ratio at particular ductility μ .
Similarly, assuming the performance point of the sub model to be (Dk , Ak ) and would
I. The x-coordinate is
Dk = max[qk (t )] (5-47)
where
q&&k + 2ξ k ωo q& k + ωo2 qk = − ku&&g (5-48)
The y-coordinate is
Ak = ω k2 Dk (5-49)
By comparing Eq. (5-44) and Eq. (5-48) and assuming ξ o = ξ k , the modal coordinate of
Ak = k ⋅ S ao (5-53)
Checking the assumption by substituting Eq. (5-51) into Eq. (5-50), the damping ratio at
the performance point is equal to that of the prime model, so the assumption is correct.
103
5.4.4 Modified modal pushover analysis
Nonlinear inelastic behaviour of building structures is analysed by the modified modal
pushover analysis by Chopra (2004), which is a well recognised method of analysis. It argues
that the inelastic range of responses is mainly contributed by the yielding of the fundamental
mode. The influence of the higher modes could be considered as linear elastic and is
combined to the yielding of the fundamental mode, similar to that in modal analysis. Based
on this principle, the inelastic responses are predicted from the modified modal pushover
analysis.
When the prime model and the sub model have the same strength ratio, the higher mode
response of the sub model is k times of the prime model. The same conclusion can be found
in the response from the capacity spectrum analysis of the fundamental mode of the prime and
the sub model according to section 5.4.3. Therefore the lateral displacement of the sub
model is k times that of the prime model under modal combination. It is given by,
ys ( z , μ ) = k ⋅ y p ( z , μ ) (5-54)
The interstorey drift ratio is the derivative of the lateral displacement which is
∂ys ( z, μ )
IDRs ( z, μ ) = = IDRp (z, μ ) (5-55)
∂x
The curvature is the second derivative of the lateral displacement which is
1
curs ( z, μ ) = ⋅ curp ( z, μ ) (5-56)
k
When the two continuum models are similar and they have the same strength ratio, the response
of the prime model and the sub model is related by Eqs. (5-54), (5-55) and (5-56). It indicates
that two buildings with identical deformation characteristics, fundamental periods, the same
proportional capacity and dimensions, their seismic responses for the same strength ratio are
proportional to each another according to the proportionality constant in their dimensions.
5.4.5 Conclusion
A modified continuum-MDOF model has been proposed for preliminary seismic assessment
of a larger number of building structures. Although the effort and requirement for the
establishment of the model and the analysis have been reduced for individual building
structures, it is still an involved task to establish the models and to conduct analysis for each
individual building structure. It would be beneficial if a single model and analysis method
could be applied to other building structures which have similarity with the single model.
104
A general continuum representation is thus established, in which models with similarity
as in Section 5.4.1, are related in responses by the height ratio for the same strength ratio.
When there is a continuum model, its responses can be applied to other models in regard to
linear static behaviour, linear dynamic behaviour and modified modal pushover analysis.
From the general continuum representation, a modified continuum-MDOF model and its
analysis result could be applied to other similar continuum-MDOF models by the height ratio
for the same strength ratio.
Based on the investigation of the properties of the modified continuum-MDOF model and the
general continuum representation, the following conclusions can be drawn.
105
b
Shear cantilever
87.5
A = 10
3.5
R = 20
Fibre element
(a) (b)
106
620.64 2095.93
E=349611.44 GA=821943.63
εy γy
(a) (b)
Fig. 5.3 Stress strain relationship (a) Fibres in the flexural cantilever; (b) Shear cantilever
8000
7000
6000
(0.175, 5741)
Base shear (kN)
5000
4000
3000
2000
1000
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8
Roof displacement (m)
Fig. 5.4 Capacity curve of continuum-MDOF model
107
0.2 0.2
Ground acceleration (g)
0 0
-0.1 -0.1
-0.2 -0.2
0 30 60 90 120 150 180 0 30 60 90 120 150 180
Time (s) Time (s)
(a) (b)
0.4 1
Roof displacement (m)
0.75
0.2
0.5
0 0.25
0
-0.2
-0.25
-0.4 -0.5
0 30 60 90 120 150 180 0 30 60 90 120 150 180
Time (s) Time (s)
(c) (d)
Fig. 5.5 (a) Excitation from PEER with record ID: P1121; (b) The roof displacement under
the excitation in (a); (c) The roof displacement under the excitation two times of (a); (d) The
roof displacement under the excitation three times of (a)
108
Prime model Sub model
ms
ws
mp
wp
EIs GAs Hs
EIp GAp Hp
Prime model
Sa
Sd Sds
Spectra displacement
109
10000
8000
Base shear (kN)
6000
4000
2000
continuum
detail
0
0 0.5 1 1.5 2
Roof displacement (m)
10000
8000
Base shear (kN)
6000
4000
2000
continuum
detail
0
0 0.5 1 1.5 2
Roof displacement (m)
Fig. 5.7 Capacity curve of building α =1.6. (a) Inverted triangular loading;
(b) Uniform loading
110
12000
10000
Base shear (kN)
8000
6000
4000
2000 continuum
detail
0
0 0.5 1 1.5 2
Roof displacement (m)
12000
10000
Base shear (kN)
8000
6000
4000
2000 continuum
detail
0
0 0.5 1 1.5 2
Roof displacement (m)
Fig. 5.8 Capacity curve of building α =3.0. (a) Inverted triangular loading;
(b) Uniform loading
111
10000
8000
Base shear (kN)
6000
4000
2000
continuum
detail
0
0 0.2 0.4 0.6 0.8 1
Roof displacement (m)
10000
8000
Base shear (kN)
6000
4000
2000
continuum
detail
0
0 0.2 0.4 0.6 0.8 1
Roof displacement (m)
Fig. 5.9 The capacity curve of building α =9.4. (a) Inverted triangular loading;
(b) Uniform loading
112
Roof drift (%) Interstory drift ratio (%)
2 2
1.5 1.5
Original model
Original model
1 1
0.5 0.5
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Continuum model Continuum model
(a) (b)
-1
x 10
-3
Curvature (m )
3
2.5
2
Original model
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
Continuum model x 10
-3
(c)
Fig. 5.10 Model with α = 1.6. (a) Roof drift; (b) Interstorey drift ratio; (c) Curvature
113
Roof drift (%) Interstory drift ratio (%)
2 2
1.5 1.5
Original model
Original model
1 1
0.5 0.5
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Continuum model Continuum model
(a) (b)
-1
x 10
-3
Curvature (m )
3
2.5
2
Original model
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
Continuum model x 10
-3
(c)
Fig. 5.11 Model with α = 3.0. (a) Roof drift; (b) Interstorey drift ratio; (c) Curvature
114
Roof drift (%) Interstory drift ratio (%)
2 2
1.8
1.6
1.5
1.4
Original model
Original model
1.2
1 1
0.8
0.6
0.5
0.4
0.2
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Continuum model Continuum model
(a) (b)
-1
x 10
-3
Curvature (m )
3
2.5
2
Original model
1.5
0.5
0
0 0.5 1 1.5 2 2.5 3
Continuum model x 10
-3
(c)
Fig. 5.12 Model with α = 9.4. (a) Roof drift; (b) Interstorey drift ratio; (c) Curvature
115
Chapter 6
Generalised Inelastic Response Spectra
6.1 Introduction
In preliminary seismic assessment, response spectra developed from an SDOF model has
provided an efficient and accurate means for predicting the linear elastic response of building
structures. Based on the fundamental period which is usually estimated from empirical
equations, the corresponding response is read from the spectrum at the chosen damping ratio
and the predicted fundamental period. The use of drift spectra is an alternative mean similar to
the response spectra, in which the interstorey drift ratio is directly read from the spectra at the
fundamental period. The drift spectrum (Iwan, 1997) and the generalised interstorey drift
spectrum (Miranda and Akkar, 2006) are the examples of such spectra. They were developed
from the continuum modelling technique and linear response history analysis.
116
Although response spectra and drift spectra have been used widely in seismic assessment,
their limitations in only being applicable for linear elastic responses are also well known. To
predict the nonlinear inelastic response probably occurring in an earthquake, two different
approaches developed from the SDOF models, namely the equivalent linear method and the
nonlinear inelastic method are used. The two approaches produce the equivalent linear
response spectrum and the inelastic response spectrum.
The equivalent linear response spectrum is developed from the response of an equivalent
linear SDOF model of its corresponding nonlinear inelastic system. A nonlinear system is
represented by an equivalent linear system which has a different period and damping ratio (Lin
and Miranda, 2008). The period and the damping ratio of the equivalent linear system are
defined through a large scale comparison between different models under response history
analysis. An alternative method to define the equivalent system is to replace the nonlinear
system with a linear system of a greater damping ratio while the system’s period is not changed.
A displacement coefficient is used to convert the linear response to the inelastic response
(Miranda, 2000). In both approaches, the equivalent linear system will not behave identically
to the nonlinear system in the time domain, but is intended to capture the maximum response of
the nonlinear system.
Inelastic response spectra are developed from SDOF models which possess hysteretic
behaviour. Inelastic response spectra are of two different types depending on the independent
variables used to predict the response. The response of the first type is expressed by the
ductility ratio while the second type is expressed by the strength ratio. For preliminary seismic
assessment, the strength ratio is more convenient than the ductility ratio as the ductility ratio is
an implicit parameter for non-seismically designed buildings. Iteration is needed to determine
the ductility ratio under the selected ground motion. Therefore, for preliminary assessment, in
which the assessment effort has to be minimised, inelastic response spectra that are expressed
by the strength ratio are preferred.
The strength ratio (R) from the nonlinear static analysis of a model, which is also referred
to as the strength reduction factor, is defined as,
PF1MSa
R= (6-1)
fy
where PF1 is the participation factor of the fundamental mode; M is the mass of the system; Sa is
117
the linear pseudo-acceleration in spectral ordinates; fy is the lateral yield strength of the system
under lateral load. The numerator represents the maximum base shear induced by the ground
motion of the fundamental mode. The denominator is the base shear at yield. The ratio hence
represents the induced base shear by the earthquake to the yield strength of the system. When
the ratio is larger than unit, yielding of the system occurs. Given any system, the value of the
pseudo acceleration Sa can be obtained from the linear elastic response spectra in the design
codes by the fundamental period. The lateral yield strength of the buildings can either be
obtained from the pushover analysis or estimation from the design lateral load of the building.
High-rise building structures are complicated in behaviour and cannot be described only
by an SDOF model, which only monitors one single response. They are represented by MDOF
models for capturing different responses simultaneously. When an SDOF model is used to
assess a building structure, conversion factors or procedures have to be considered to convert
the responses of an SDOF system to the responses of an MDOF system. In the process, it is
assumed that the mode shapes exist in the nonlinear inelastic response range or they are
identical to the mode shapes in the linear elastic region. The result obtained from the SDOF
system is used to amplify the linear elastic mode so that an inelastic result of an MDOF system
can be predicted. The inelastic response however differs from the elastic mode not only by
magnitude, but also by shape. The conversion process from the nonlinear SDOF model to a
nonlinear MDOF model is unfavourable.
A tall building structure experiences higher mode effects which also cannot be represented
by an SDOF model. Several SDOF models could be combined based on modal analysis to
capture the additional force and deformation due to the higher mode effect, but this can only be
valid in the linear elastic response range. Once the building structure is excited into the
inelastic response range, methods based on the SDOF model are not accurate but can be
regarded as approximate. In situations with significantly higher mode effects and an inelastic
response range, an MDOF model is usually used to represent the building structure and is
analysed by response history analysis.
In conclusion, response spectra developed from SDOF models are not suitable for
high-rise building structures which may experience an inelastic range of responses and
significantly higher mode effects. The modified continuum-MDOF model discussed in
Chapter 5 is therefore proposed for preliminary seismic assessment. Its simplicity in
118
modelling and its generality for application to building structures make it a favoured model for
the development of a response spectrum that suits high-rise building structures having
significantly high mode effects. From the general continuum representation, the responses
and the analysis results of a modified continuum-MDOF model can be applied to other similar
continuum-MDOF models. This will help to establish a generalised inelastic response
spectrum, which is a spectrum-based method, for high-rise building structures with various
deformation characteristics for use in preliminary seismic assessment.
6.2 Methodology
From the general continuum representation, the inelastic responses obtained from a continuum
model can be used to approximate the inelastic response of other similar models using the
height ratio. If a prime model (modified continuum-MDOF model) is studied, a sub model
(assessed building) which is similar to the prime model can be approximated. The
relationships for the maximum inelastic responses between the prime model and the sub model
given by Eqs (5-53), (5-54) and (5-55) are
⎧
⎪ ys = k ⋅ y p
⎪⎪
⎨ IDRs = IDR p (6-2)
⎪
⎪ curs = 1 ⋅ curp
⎩⎪ k
where the subscript “s” refers to the sub model while “p” refers to the prime model. The value
k is the height ratio and is the ratio of the height of the sub model to that of the prime model.
By substituting the expression of the height ratio k into Eq. (6-2), the inelastic responses of
the sub model are expressed as follows,
⎧ yp
⎪ ys = × Hs
⎪ H p
⎪
⎨ IDRs = IDR p (6-3)
⎪
⎪ curs = curp × H p × 1
⎪⎩ Hs
In Eq. (6-3), any information concerning the prime model is extracted and is in dimensionless
form. The dimensionless responses given in Eq. (6-4) are referred to as the generalised
inelastic responses,
119
⎧ yp
⎪ Gy =
⎪ Hp
⎪
⎨G IDR = IDR p (6-4)
⎪
⎪Gcur = curp × H p
⎪⎩
By multiplying the inelastic generalised responses to the height of the sub model, the responses
of the sub model can be predicted according to Eq. (6-5).
Eq. (6-5) is derived from the general continuum representation, which applies the
responses to similar models in nonlinear inelastic behaviour by the modified modal pushover
analysis. The modified modal pushover analysis and nonlinear response history analysis
both can reveal the nonlinear inelastic behaviour of a building structure. They are regarded
as the equivalent method of each another and their results are consistent. Therefore, the
observations in the general continuum representation made by the modified modal pushover
analysis are also applied to the nonlinear response history analysis. Nonlinear response
history analysis is then conducted on the prime model to obtain the generalised inelastic
responses.
Generalised inelastic response spectra are developed from the modified continuum-MDOF
model. The modified continuum-MDOF gives reliable predictions of the roof displacement
and the interstorey drift ratio. On the other hand, it tends to underestimate the curvature; hence
the generalised inelastic curvature spectra are not recommended for the purpose of assessment,
but it included for a general background understanding.
120
6.3.1 Generalised inelastic response
The generalised inelastic responses of a continuum flexural-shear model depend on the degree
of structural interaction, the fundamental period and the strength ratio of the particular
hysteretic behaviour under the expected ground motion. The relationships between the
responses and the various parameters are presented in the form of response spectra that are
developed from the results of the response history analysis of 168 MDOF models under
selected ground motions.
The yield strength of the model for defining the strength ratio is obtained from pushover
analysis under an inverted triangular distributed force (Case 1) and a uniform distributed force
(Case 2). Thus two different kinds of strength ratios for different applied loads are defined.
They represent two different equivalent seismic forces experienced by the model and they have
been recognised as the upper and lower bound values of the equivalent seismic force. Each
group of excitations in Table 6.1 contains 11 ground motions which are obtained from FEMA
440 (Appendix F) and are scaled to achieve the defined strength ratio.
121
6.3.3 Generalised interstorey drift ratio spectra
The spectra of the generalised interstorey drift ratios are shown in Fig. 6.9 to Fig. 6.16. Similar
to the roof displacement spectra, the response increases with the strength ratio. The minimum
response occurs at a fundamental period of about 2.2s under the selected ground motions. The
spectra with a degree of structural interaction of 8 give the maximum response.
122
A suggested method for obtaining information on the building for establishing the
corresponding continuum-MDOF model is presented. The height of the assessed building is
approximately equal to
No. of storeys μ interstorey height
The degree of structural interaction can be estimated by the drift ratio as shown in Fig. 3.6 in
chapter 3 and the fundamental period of the building can be predicted by Eq. (3-16) also
suggested in chapter 3, and is used to determine the linear elastic spectral acceleration for
calculating the strength ratio. Moreover there are various alternatives for obtaining the
fundamental period, for instance empirical formulas suggested in codes of practice or
estimations from engineering experience.
The strength ratio presented by Eq. (6-1) requires information about the participation factor,
the mass of the building, the pseudo acceleration at the fundamental period and the yield
strength of the building. The participation factor of the fundamental mode of the flexural-shear
model can be read from Fig. 6.25 by the degree of structural interaction. The mass of the
building is approximately equal to
The pseudo acceleration is simply obtained from the corresponding pseudo-acceleration spectra
by the fundamental period at the initial elastic viscous damping ratio.
The yield strength of the building can be predicted from the design lateral strength. It is
approximated by
The value of the overstrength coefficient is an implicit design parameter which is unknown.
However, from the suggestions in the SEAOC Blue Book (SEAOC, 2009), a conservative value
of the overstrength coefficient can be assigned as either 2.0 or 2.5 depending on the system,
which has bee discussed in Section 4.2.3. Moreover, a more precise value can be defined from
pushover analysis.
123
6.4.2 Estimating responses of assessed building
To predict the response of the assessed building using the generalised inelastic response, the
strength ratio of the assessed building under the expected ground motion has to be defined.
When the strength ratio is known, the generalised inelastic responses can be found from Fig. 6.1
to Fig. 6.24. The inelastic responses of the assessed building can then be approximated
according to Eq. (6-5)
A representative building is selected from Chapter 4 for illustration in the proposed generalised
inelastic response spectra. The selected building is the 25-storey wall-frame building. The
height of the wall-frame building is 87.5 m with an interstorey height of 3.5 m. The design
roof drift is 0.14m under a lateral pressure of 2.80 kPa, which decreases parabolically to 0 at the
base. The degree of structural interaction (α) of the building is approximately equal to 3, using
the method of drift ratio RΔ/Rδm in Fig. 3.6. The mass at the floor level of the models is 137 Mg
with fundamental period of 2 seconds.
124
The height of the building = 87.5 m
The degree of structural interaction = 3
The fundamental period = 2 s
The floor mass of the building = 137 Mg
The pseudo acceleration obtained from Fig. 6.26 = 2.4 ms-2
From pushover analysis with the inverted triangular load pattern in Fig. 6.27, the yield
strength of the building = 7500 kN
From the generalised inelastic response spectra, the dimensionless responses are
From Eq. (6-5), the inelastic responses of the assessed building are
⎧
⎪ ymax = G y × H
⎪
⎨ IDRmax = GIDR
⎪ 1
⎪ curmax = Gcur ×
⎩ H
Hence,
⎧
⎪ ymax = 0.005 × 87.5
⎪
⎨ IDRmax = 0.007
⎪ 1
⎪ curmax = 0.10 ×
⎩ 87.5
then
125
⎧ ymax = 0.44 m
⎪
⎨ IDRmax = 0.007
⎪ −1
⎩ curmax = 0.0011 m
Comparisons between the results obtained from the response history analysis and the
predictions obtained from the generalised response spectra method are shown in Figs. 6.28 to
6.30. It can be seen that the predictions from the generalised response spectra, the solid line,
are more conservative than the results obtained from the response history analysis. The reason
is due to the simplification of structural system from a detailed model to the modified
continuum-MDOF model. An equivalent cantilever is used for each storey of the modified
continuum-MDOF model. The modified continuum-MDOF model with isotropic hysteretic
hardening behaviour would dissipate more energy with smaller displacement compared to the
detailed model. For both the roof displacement and the interstorey drift ratio, the proposed
generalised inelastic response spectra method gives a reasonably good prediction, as compared
to the response history analysis which requires significantly larger effort.
126
6.6 Concluding Remarks
A preliminary assessment method using the generalised inelastic response is proposed. The
generalised inelastic responses are derived from the modified continuum-MDOF model
through the general continuum representation, where the responses of a prime model can be
applied to other similar models using the height ratio. The generalised responses are expressed
in the form of spectra, which describe the relationship between various parameters, including
the fundamental period, the degree of structural interaction and the strength ratio of the building
under the expected ground motions. The generalised inelastic response spectra are developed
from the responses of 168 modified continuum-MDOF models, of heights 100 m, having 8
different degrees of structural interaction and 23 different fundamental periods using nonlinear
response history analysis.
The responses of the assessed building structure are determined from the suggested
general inelastic response spectra and the height of the assessed building, and are given by
⎧
⎪ y = Gy × H
⎪
⎨ IDR = GIDR
⎪ 1
⎪ cur = Gcur ×
⎩ H
The numerical investigation of a 25-storey wall-frame building shows that the proposed method
using the generalised inelastic response spectra has successfully predicted the seismic response
of the building and it agrees satisfactorily with the response history analysis of the
corresponding 2-dimensional finite-element based model. The proposed method has provided
a simple and quick, yet accurate, means of estimating the inelastic seismic response of a
structure, particularly suited for preliminary seismic assessment.
127
Table 6.1 Ground Motion records
Extracted from FEMA 440, Appendix F, Table F-6
Magnitude PGA PGV
# Identifier Earthquake Date Station Location Source
(Ms) (g) (cm/sec)
El Centro imp. Co. Cent
1 ICCC000 Superstitn 11-24-87 6.6 0.358 46.4 CDMG
(01335)
Canyon Country-W Lost
2 LOS000 Northridge 1-17-94 6.7 Cany 0.41 43 USC
(90057)
3 G02090 Loma Prieta 10-18-89 7.1 Gilroy Array #2 (47380) 0.322 39.1 CDMG
Chi-Chi,
4 TCU122N 9-20-99 7.6 (TCU122) 0.261 34 CWB
Taiwan
5 G0390 Loma Prieta 10-18-89 7.1 Gilroy Array #3 (47380) 0.367 44.7 CDMG
Canoga Park-Topanga Can
6 CNP196 Northridge 1-17-94 6.7 0.42 60.8 USC
(90053)
7 CHY101W Chi Chi, Taiwan 9-20-99 7.6 (CHY101) 0.353 70.6 CWB
El Centro imp. Co. Cent
8 ICC090 Superstitn 11-24-87 6.6 0.258 40.9 CDMG
(01335)
Canoga Park-Topanga Can
9 CNP106 Northridge 1-17-94 6.7 0.356 32.1 USC
(90053)
10 E02140 Imperial Valley 10-15-79 6.9 El Centro Array #2 (5115) 0.315 31.5 USGS
11 E11230 Imperial Valley 10-15-79 6.9 El Centro Array #11 (5058) 0.38 42.1 USGS
CDMG: California Division of Mines and Geology
CWD: Central Weather Bureau, Taiwan
USC: University of Southern California
USGS: U.S. Geological Survey
128
α =1
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.1 Generalised roof drift spectra (a) and (b) (Case 1)
129
α =3
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.2 Generalised roof drift spectra (c) and (d) (Case 1)
130
α =5
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
0.04
0.035
0.03
0.025
y
0.02
G
0.015 R=4
R=3
0.01
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.3 Generalised roof drift spectra (e) and (f) (Case 1)
131
α =7
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.4 Generalised roof drift spectra (g) and (h) (Case 1)
132
α =1
0.04
0.035
0.03
0.025
y
0.02
G
0.015 R=4
R=3
0.01
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
0.04
0.035
0.03
0.025
y
0.02
G
0.015 R=4
R=3
0.01
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.5 Generalised roof drift spectra (a) and (b) (Case 2)
133
α =3
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.6 Generalised roof drift spectra (c) and (d) (Case 2)
134
α =5
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
0.04
0.035
0.03
0.025
y
0.02
G
0.015 R=4
R=3
0.01
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.7 Generalised roof drift spectra (e) and (f) (Case 2)
135
α =7
0.04
0.035
0.03
0.025
y
0.02
G
0.015
R=4
0.01 R=3
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
0.04
0.035
0.03
0.025
y
0.02
G
0.015 R=4
R=3
0.01
R=2
0.005 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.8 Generalised roof drift spectra (g) and (h) (Case 2)
136
α =1
0.04
0.035
0.03
0.025
IDR
0.02 R=4
G
R=3
0.015
R=2
0.01
R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
0.04
0.035
0.03
0.025
IDR
0.02 R=4
G
R=3
0.015
R=2
0.01
R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.9 Generalised interstorey drift ratio spectra (a) and (b) (Case 1)
137
α =3
0.04
0.035
0.03
0.025
IDR
0.02 R=4
G
0.015 R=3
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
0.04
0.035
0.03
0.025 R=4
IDR
0.02
R=3
G
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.10 Generalised interstorey drift ratio spectra (c) and (d) (Case 1)
138
α =5
0.04
0.035
0.03
0.025 R=4
IDR
0.02
R=3
G
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
0.04
0.035
0.03 R=4
0.025 R=3
IDR
0.02
G
0.015 R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.11 Generalised interstorey drift ratio spectra (e) and (f) (Case 1)
139
α =7
0.04
0.035
0.03
0.025 R=4
IDR
0.02
R=3
G
0.015
R=2
0.01
R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
0.04
0.035
R=4
0.03
0.025 R=3
IDR
0.02
G
0.015 R=2
R=1.5
0.01
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.12 Generalised interstorey drift ratio spectra (g) and (h) (Case 1)
140
α =1
0.04
0.035
0.03
0.025
R=4
IDR
0.02
G
R=3
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
0.04
0.035
0.03
0.025
IDR
0.02 R=4
G
R=3
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.13 Generalised interstorey drift ratio spectra (a) and (b) (Case 2)
141
α =3
0.04
0.035
0.03
0.025
R=4
IDR
0.02
G
R=3
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
0.04
0.035
0.03
0.025 R=4
IDR
0.02 R=3
G
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.14 Generalised interstorey drift ratio spectra (c) and (d) (Case 2)
142
α =5
0.04
0.035
0.03
0.025 R=4
IDR
0.02 R=3
G
0.015
R=2
0.01 R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
0.04
0.035
R=4
0.03
0.025 R=3
IDR
0.02
G
R=2
0.015
R=1.5
0.01
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.15 Generalised interstorey drift ratio spectra (e) and (f) (Case 2)
143
α =7
0.04
0.035
0.03
0.025 R=4
IDR
0.02 R=3
G
0.015
R=2
0.01
R=1.5
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
0.04
0.035 R=4
0.03
R=3
0.025
IDR
0.02
G
R=2
0.015
R=1.5
0.01
0.005
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.16 Generalised interstorey drift ratio spectra (g) and (h) (Case 2)
144
α =1
1
0.8
0.6
cur
G
0.4
R=4
0.2 R=3
R=2
R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
1
0.8
0.6
cur
G
0.4 R=4
R=3
0.2 R=2
R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.17 Generalised curvature spectra (a) and (b) (Case 1)
145
α =3
1
0.8
0.6
cur
G
R=4
0.4
R=3
R=2
0.2
R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
1
0.8
0.6
R=4
cur
G
R=3
0.4
R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.18 Generalised curvature spectra (c) and (d) (Case 1)
146
α =5
1
0.8
0.6 R=4
cur
R=3
G
0.4
R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
1
0.8 R=4
0.6 R=3
cur
G
R=2
0.4
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.19 Generalised curvature spectra (e) and (f) (Case 1)
147
α =7
1
0.8
R=4
0.6
R=3
cur
G
0.4 R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
1
R=4
0.8
R=3
0.6
cur
R=2
G
0.4 R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.20 Generalised curvature spectra (g) and (h) (Case 1)
148
α =1
1
0.8
0.6
cur
G
0.4
R=4
0.2 R=3
R=2
R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(a)
α =2
1
0.8
0.6
cur
G
0.4 R=4
R=3
0.2 R=2
R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(b)
Fig. 6.21 Generalised curvature spectra (a) and (b) (Case 2)
149
α =3
1
0.8
0.6
cur
R=4
G
0.4 R=3
R=2
0.2 R=1.5
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(c)
α =4
1
0.8
0.6 R=4
cur
G
R=3
0.4
R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(d)
Fig. 6.22 Generalised curvature spectra (c) and (d) (Case 2)
150
α =5
1
0.8
0.6 R=4
cur
R=3
G
0.4
R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(e)
α =6
1
0.8 R=4
0.6 R=3
cur
G
R=2
0.4
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(f)
Fig. 6.23 Generalised curvature spectra (e) and (f) (Case 2)
151
α =7
1
0.8
R=4
0.6 R=3
cur
G
0.4 R=2
R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(g)
α =8
1.2
R=4
1
0.8 R=3
cur
0.6 R=2
G
0.4 R=1.5
0.2
0
1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
(h)
Fig. 6.24 Generalised curvature spectra (g) and (h) (Case 2)
152
2
1.9
Participation factor of 1st mode
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0 2 4 6 8 10 12 14 16 18 20
Degree of structural interaction
Fig. 6.25 Participation factor of 1st mode
10
9
Spectral acceleration (ms )
-2
8
7
6
5
4
3
2
1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Fundamental period (s)
Fig. 6.26 Mean pseudo- acceleration spectrum with 5% viscous damping of ground motions
in Table 6.1
153
10000
8000
Base shear (kN)
6000
4000
2000
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Roof drift ratio
1
0.9
0.8
0.7
Roof drift (m)
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Ground motion record
Fig. 6.28 Roof drift under response history analysis and prediction from method of
generalised inelastic response spectra
154
0.016
0.014
0.012
0.01
IDR
0.008
0.006
0.004
0.002
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Ground motion record
Fig. 6.29 Interstorey drift ratio under response history analysis and prediction from the
method of generalised inelastic response spectra
0.0016
0.0014
0.0012
Curvature (m )
0.001
-1
0.0008
0.0006
0.0004
0.0002
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Ground motion record
Fig. 6.30 Curvature under response history analysis and prediction from method of
generalised inelastic response spectra
155
Chapter 7
Conclusion
7.1 Conclusion
The principal objectives of the research into the development of quick and simple, yet
accurate, performance-based methods at the stage of the preliminary seismic vulnerability
assessment, where a large number of buildings has to be evaluated in a short period of time,
have been attained. On the basis of the theoretical studies combining with numerical
investigations, the following conclusions can be drawn.
(1) A simple and effective, yet accurate model is presented for quick estimation of the natural
fundamental period of vibration for tall building structures. It is derived based on the
continuum technique, where a tall-building structure is considered as a continuous,
interactive flexural-shear cantilever. The proposed model is presented in a closed-form
mathematical expression for calculating the natural fundamental period T1, in which the
effects of the interaction of different structural forms in the building, design roof drift ratio
and maximum interstorey drift ratio, intensities of applied loading and height of the
structure on the fundamental period of vibration have been considered.
156
(2) An inelastic interstorey drift ratio model is proposed, with the use of graphical
representation for the preliminary seismic assessment of high-rise buildings that may
deform inelastically. The inelastic interstorey drift ratio model is derived from the
principle of capacity spectrum analysis and modified modal pushover analysis for the
assessed buildings. Numerical investigations on representative tall wall-frame structures,
which are designed against gravity and lateral loads, show that the seismic responses of the
buildings predicted by the proposed inelastic interstorey drift model agree well with those
obtained from response history analysis for a given set of ground motions. The proposed
model is particularly suitable for assessing the building structures that are analysed in
frequency domain.
(3) A systematic and effective procedure for preliminary seismic assessment is developed,
which serves as a tool for screening and evaluating a large number of existing buildings in
a short period of time and for identifying potential hazardous structures. The proposed
preliminary seismic assessment procedure first divides the assessed building structures
into either elastic or inelastic ones under the expected ground motion using a bifurcation
index. Buildings that behave elastically are safe, whereas buildings that may deform
inelastically are further examined by the inelastic interstorey drift model which can be
quickly implemented through a series of aided charts. Performance of the building
structure is then defined based on the interstorey drift ratio by comparing those given in
the current codes of practice and the literature.
(4) A modified continuum-MDOF model is derived for analysing seismic responses of tall
building particularly for slender building that has significantly higher vibration modes
and inelastic response ranges, which is more suitable for rapid preliminary seismic
assessment as compared to the original continuum-MDOF model because both the
required information and the effort for analysis are lesser. The numerical investigation
shows that the proposed model gives reasonably good predictions of the interstorey drift
ratio in response history analysis, as compared to the corresponding detailed model.
157
(5) A general continuum representation is derived, in which the seismic responses of two
similar continuum models are related by the height ratio for the same strength ratio in
linear static behaviour, linear dynamic behaviour and modified modal pushover analysis.
When a modified continuum-MDOF model is analysed corresponding to a particular
fundamental period and degree of structural interaction, the seismic responses of other
building structures with similar models can be approximated through the height ratio for
the same strength ratio.
(6) A spectrum-based method using the generalised inelastic response spectra is proposed and
recommended for preliminary seismic assessment of tall buildings that are relatively
slender with significant higher mode effects and are normally analysed in time domain.
The responses of a building structure can be predicted using the generalised inelastic
responses, which are derived from the modified continuum-MDOF model through the
general continuum representation. The generalised inelastic response spectra are
developed by the responses of 168 modified continuum-MDOF models, having 8 different
degrees of structural interaction and 21 different fundamental periods under nonlinear
response history analysis. The numerical investigation shows that the seismic responses
from the proposed method agree well with those from the response history analysis of the
corresponding 2-dimensional finite-element-based model. The proposed method
provided a simple and quick, yet accurate, mean of estimating the inelastic seismic
response of building structures, thus particularly being suitable for the preliminary seismic
assessment.
158
7.2 Further research
A systematic and effective procedure using the methods of inelastic interstorey drift ratio model
and generalised inelastic response spectrum has been developed for preliminary seismic
assessment. It serves as a seismic assessment tool for screening and evaluating a large number
of existing buildings in a short period of time. Therefore the principal objective of the research
has been attained.
Following the analytical development, further calibrations on the demand spectrum and
damping function may improve the accuracy of the method. Parametric study on the building
properties may simplify the analysis and increase the efficiency of the method.
The effects of torsion due to asymmetric floor plan should be further studied by
considering models with eccentricity difference in mass and stiffness centres. Adverse seismic
effect due to vertical stiffness and strength variations may be modelled. Drift multipliers that
have included the above effects could be derived either statically or analytically. They could
be implemented in the proposed inelastic interstorey drift model and generalised inelastic
response spectrum, thus that the proposed method of analysis may be applied to more variety of
situations.
159
References
ATC 40. (1996). Seismic Evaluation and Retrofit of Concrete Buildings, Applied Technology
Council, Redwood City, California.
Chopra, A. K., Goel, R. K. (2002). “A modal pushover analysis procedure for estimating
seismic demands for buildings,” Earthquake Engineering and Structural Dynamics, 31(3),
561-582.
Chopra, A. K., Goel, R.. K. (2004). “A modal pushover analysis procedure to estimate seismic
demands for unsymmetric-plan buildings,” Earthquake Engineering and Structural Dynamics,
33(8), 903-927.
Fajfar, P. (1999). “Capacity spectrum method based in inelastic demand spectra,” Earthquake
Engineering and Structural Dynamics, 28(9), 979-993.
FEMA 154. (2002). Rapid Visual Screening of Buildings for Potential Seismic Hazards: A
Handbook (second edition), Federal Emergency Management Agency, Washington, D.C.
FEMA 273. (1997). NEHRP Guidelines for the Seismic Rehabilitation of Buildings, Federal
Emergency Management Agency, Washington, D.C.
FEMA 310. (1998). Handbook for the Seismic Evaluation of Buildings, Federal Emergency
Management Agency, Washington, D.C.
FEMA 356. (2000). Prestandard and Commentary for the Seismic Rehabilitation of Buildings,
Federal Emergency Management Agency, Washington, D.C.
FEMA 440. (2005). Improvement of Nonlinear Static Seismic Analysis Procedures, Federal
Emergency Management Agency, Washington, D.C.
160
Freeman, S. A. (1998). “The capacity spectrum method as a tool for seismic design,”
Proceedings of the 11th European Conference on Earthquake Engineering, Paris, France.
Freeman, S. A., (2004). “Review of the development of the capacity spectrum method,” ISET
Journal of Engineering Technology, paper no. 438, 41(1), 1-13.
Goel, R. K., Chopra, A. K. (1998). “Period formulas for concrete shear wall buildings,” Journal
of Structural Engineering, 124(4), 426-433.
Gunay, M. S., Sucuoglu, H. (2010). “An improvement to linear-elastic procedures for seismic
performance assessment,” Earthquake Engineering and Structural Dynamics, 39, 907-931.
Huang, K. (2009). Continuum MDOF Model for Seismic Analysis of Wall-frame Structures,
Ph.D. dissertation, Department of Civil and Environmental Engineering, The Hong Kong
University of Science and Technology.
Huang, K., Kuang, J. S. (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(5), 573-588.
Ile, N., Reynouard, J. M. (2000). “Nonlinear analysis of reinforced concrete shear wall under
earthquake loading,” Journal of Earthquake Engineering, 4(2), 183-213.
Iwan, W. D. (1997). “Drift spectrum measure of demand for earthquake ground motions,”
Journal of Structural Engineering, 124(4), 397-404.
Kappos, A. J. (1997). “Seismic damage indices for RC buildings: evaluation of concepts and
procedures,” Progress in Structural Engineering and Materials, 1(1), 78–87.
Kim, T. W., Foutch, Douglas A., LaFave, James M. (2005). “A practical model for seismic
analysis of reinforced concrete shear wall buildings,” Journal of Earthquake Engineering, 9(3),
393-417.
Krawinkler, H., Seneviratna, GDPK. (1998). “Pros and cons of a pushover analysis of seismic
performance evaluation,” Engineering Structures, 20(4-6), 452-464.
161
Lang, K. (2002). Seismic Vulnerability of Existing Building, IBK Report No. 273, Swiss
Federal, Institute of Technology (ETH), Zurich, Switzerland.
Li, C. S., Lam, S. S., Zhang, M. Z. (2006). “Shaking table test of a 1:20 scale high-rise
building with a transfer plate system,” Journal of Structural Engineering, 132(11),
1732-1744.
Lin, Y. Y., Miranda, E. (2008). “Noniterative equivalent linear method for evaluation of
existing structures,” Journal of Structural Engineering, 134(11), 1685-1695.
Mazzoni S., McKenna F., Scott M. H., Fenves G. L. (2006). OpenSees Manual,
http://opensees.berkeley.edu/OpenSees/manuals/usermanual/ (last visited in August 2010).
Michael, J. S., Andrei, M. R., Young, J. P. (1989). “Seismic damageability assessment of R/C
buildings in Eastern U.S,” Journal of Structural Engineering, 115(9), 2184-2203.
Miranda, E. (2000). “Inelastic displacement ratios for structures on firm sites,” Journal of
Structural Engineering, ASCE, 126(10), 1150-1159.
Miranda, E., Reyes, C. J. (2002). “Approximate lateral drift demands in multistorey buildings
with nonuniform stiffness,” Journal of Structural Engineering, ASCE, 128(7), 840-849.
Mwafy, A. M., Elnashai, A. S. (2001). “Static Pushover versus dynamic collapse analysis of RC
Buildings,” Engineering structures, 23(5), 407-424.
Pacific Earthquake Engineering Research Centre. (2000). PEER Strong Motion Database,
http://peer.berkeley.edu/smcat/ (last visited in August 2010).
Park, Y. J., Ang, AHS. (1985). “Mechanistic seismic damage model for reinforced concrete,”
Journal of Structural Engineering, 111(4), 722-739.
Rakesh, Goel, R. K., Chopra, A. K. (1997). “Period formulas for moment-resisting frame
buildings,” Journal of Structural Engineering, 123(11), 1454-1461.
Rutenberg, A. (1975). “Approximate natural frequencies for coupled shear walls,” Earthquake
Engineering & Structural Dynamics, 4(1), 95-100.
162
Sang, W. H., Chopra, A. K. (2006). “Approximate incremental dynamic analysis using the
modal pushover analysis procedure,” Earthquake Engineering and Structural Dynamics,
35(15), 1853-1873.
SEAOC (2009). SEAOC Blue Book: Seismic Design Recommendations, Structural Engineers
Association of California.
Sheikh, M. N. (2005). Seismic Assessment of Buildings in Hong Kong with Special Emphasis
on Displacement Based Approaches, Ph.D. dissertation, Department of Civil Engineering, The
University of Hong Kong, Hong Kong.
Smith, S. B., Coull, A. (1991). Tall Building Structures: Analysis and Design, Wiley, New York.
Thermou, G. E., Pantazopoulou, S. J. (2011). “Assessment indices for the seismic vulnerability
of existing R.C. buildings,” Earthquake Engineering and Structural Dynamics, 40(3), 293-313.
Tsang, H. H., Ray Su, K. K., Nelson Lam, T. K., Lo, S. H. (2009). “Rapid assessment of
seismic demand in existing building structures,” The Structural Design of Tall and Special
Buildings, 18, 427-439.
Vamvatsikos, D., Cornell, C. A. (2005). “Direct estimation of the seismic demand and capacity
of multidegree-of-freedom systems through incremental dynamic analysis of single-degree-of
freedom approximation,” Journal of Structural Engineering, 131(4), 589-599.
Vamvatsikos, D., Cornell, C. A. (2006). “Direct estimation of the seismic demand and capacity
of oscillators with multi-linear static pushovers through incremental dynamic analysis,”
Earthquake Engineering and Structural Dynamics, 35(9), 1097-1117.
Wang, Y. P., Reinhorn, A. M., Soong, T. T. (1992). “Development of design spectra for actively
controlled wall frame buildings,” Journal of Engineering Mechanics, ASCE, 118(6),
1201-1220.
Williams, M. S., Sexsmith, R. G., (1995). “Seismic damage indices for concrete structures: a
state of the art review,” Earthquake Spectra, 11(2), 319-349.
Yoon, Y. S., Smith, S. B. (1995)(a). “Estimating period ratio for predicting torsional coupling,”
Engineering Structure, 17(1), 52-56.
163
Appendices
A.1 Continuum flexural-shear cantilever
q
EI GA EI GA
•• •• ••
ug ug ug
∂M
M + dx
∂x
∂V f ∂Vs
Vf + dx Vs + dx
∂x ∂x
χ
dx ∂u
2
ndx ∂ 2u
m f dx 2 ms dx 2
∂t ∂t
Vf Vs
Fig. A1 Free body diagram of flexural-shear cantilever at time t under free vibration
164
A free-body diagram of the flexural-shear cantilever at time t under free vibration with the
mass replaced by its inertia force is shown in Fig. A1. Consider the dynamic equilibrium of
both flexural and shear elements in the flexural-shear cantilever, and two equilibrium
equations in horizontal direction can be developed.
∂u
Vs = −GAχ = −GA (4)
∂x
in which M is the bending moment and χ is the shear strain. By combining the equations of
equilibrium of the elements presented by Eq. (1) and Eq. (2) and then substituting the shear
forces given by Eqs (3) and (4), the governing equation of motion of the continuum
flexural-shear model under free vibration can be derived and simplified as
∂ 2u ∂ 4u ∂ 2u
m + EI − GA =0 (6)
∂t 2 ∂x 4 ∂x 2
where m = mf + ms is the distributed mass.
where φ(x) is the shape function with respected to coordinate x along the height and q(t) is the
modal coordinate, with different independent variables. Eq. (6) can be solved by the technique
of separation of variables and becomes
165
EI (4 ) GA ''
φ (x ) − φ (x )
− q (t ) m
''
m
= (8)
q(t ) φ (x )
Functions at the right and left hand side have augment independent of each another. Since
equation (8) holds for any x and t, equation (8) is forced to equal to a constant, say ω 2 . Thus
Eq. (8) is decomposed into two equations with variables x and t respectively,
q′′ ( t ) + ω 2 q ( t ) = 0 (9)
EI ( 4) GA
φ ( x) − φ ′′ ( x ) − ω 2φ ( x ) = 0 (10)
m m
where ω is the natural vibration frequency of the continuum model. To solve for the shape
function, let φ ( x ) = Ae ax and substitute it into equation (10). A quadratic equation of a2 is
formed as follows and the values of a2 are solved.
GA 2 mω 2
a4 − a − =0 (11)
EI EI
where a12 and a22 are the roots of Eq. (11)
2
⎛ GA ⎞ 4mω
2
GA
− ⎜ ⎟ +
EI ⎝ EI ⎠ EI
a12 = or (12)
2
2
⎛ GA ⎞ 4mω
2
GA
+ ⎜ ⎟ +
EI ⎝ EI ⎠ EI
a22 = (13)
2
From Eqs (12) and (13), four solutions of a can be obtained a = ±iγ o and a = ± β o , where
2
GA ⎛ GA ⎞ 4mω 2
− + ⎜ ⎟ +
EI ⎝ EI ⎠ EI
γo = (14)
2
2
GA ⎛ GA ⎞ 4mH 4ω 2
+ ⎜ ⎟ +
EI ⎝ EI ⎠ EI
βo = (15)
2
Therefore the shape function is expressed as
φ (x ) = A1e β x + A2e − β x A3eiγ x + A4e −iγ
o o o ox
(16)
From Euler’s Formula, Eq. (16) is converted into trigonometric functions as follows
φ (x ) = C1 cos γ o x + C2 sin γ o x + C3 cosh β ox + C4 sinh β o x (17)
166
The derivatives of the shape function with respected to x are
φ ′(x ) = −C1γ o sin γ o x + C2γ o cos γ o x + C3 β o sinh β o x + C4 β o cosh β o x (18)
φ '' ( x ) = −C1γ o2 cos γ o x − C2γ o2 sin γ o x + C3βo2 cosh βo x + C4 βo2 sinh βo x (19)
The shape function in Eq. (17) has four unknown constants Ci , which have to be solved from
u' (0, t ) = 0
φ ' (0)q(t ) = 0
φ' (0) = 0
− C1γ o sin0 + C2γ o cos0 + C3βo sinh0 + C4βo cosh0 = 0
γ oC2 + βoC4 = 0
βo
Therefore C2 = − C4 (22)
γo
3. Zero moment at the roof
M (H ) = 0
∂ 2u ( H , t )
EI =0
∂x 2
φ '' (H )q(t ) = 0
φ ' ' (H ) = 0
− C1γ o2 cos γ o H − C2γ o2 sin γ o H + C3 β o2 cosh β o H + C4 β o2 sinh β o H = 0
Substituting Eq. (21) and (22) into the above equation will obtain the follows
C3γ o2 cos γ o H + C4 β oγ o2 sin γ o H + C3 β o2 cosh β o H + C4 β o2 sinh β o H = 0
( ) (
C3 γ o2 cos γ o H + β o2 cosh β o H + C4 β oγ o sin γ o H + β o2 sinh β o H = 0 )
167
C4 (β oγ o sin γ o H + β o2 sinh β o H )
Therefore C3 = − (23)
γ o2 cos γ o H + β o2 cosh β o H
From Eqs (21), (22) and (23), it can be observed that the constants C1, C2 and C3 depend on C4
which can be arbitrary values. The magnitude of the shape function hence can be selected as
any arbitrary value C4. The shape function is
φ (x ) = C1 cos γ o x + C2 sin γ o x + C3 cosh β ox + C4 sinh β o x
where
C4 (β oγ o sin γ o H + β o2 sinh β o H )
C1 =
γ o2 cos γ o H + β o2 cosh β o H
βo
C2 = − C4
γo
C4 (β oγ o sin γ o H + β o2 sinh β o H )
C3 = −
γ o2 cos γ o H + β o2 cosh β o H
4. Zero resultant shear force at the top,
∂ 3u (H , t ) ∂u (H , t )
EI − GA =0
∂x 3
∂x
EIφ (3 ) (H ) − GAφ ' (H ) = 0
GA '
φ (3 ) ( H ) − φ (H ) = 0
EI
C1γ o3 sin γ o H − C2γ o3 cos γ o H
+ C3 β o3 sinh β o H + C4 β o3 cosh β o H
GA ⎛ − C1γ o sin γ o H + C2γ o cos γ o H ⎞
− ⎜⎜ ⎟=0
EI ⎝ + C3 β o sinh β o H + C4 β o cosh β o H ⎟⎠
⎛ GA ⎞
C1 ⎜ γ o3 sin γ o H + γ o sin γ o H ⎟
⎝ EI ⎠
⎛ GA ⎞
− C2 ⎜ γ o3 cos γ o H + γ o cos γ o H ⎟
⎝ EI ⎠
⎛ GA ⎞
+ C3 ⎜ β o3 sinh β o H − β o sinh β o H ⎟
⎝ EI ⎠
⎛ GA ⎞
+ C4 ⎜ β o3 cosh β o H − β o cosh β o H ⎟ = 0
⎝ EI ⎠
Substituting Eqs (21), (22) and (23) into the above equations lead to,
168
(
C4 β oγ o sin γ o H + β o2 sinh β o H ⎛ 3 )
⎜ γ o sin γ o H +
GA ⎞
γ o sin γ o H ⎟
γ o cos γ o H + β o cosh β o H ⎝
2 2
EI ⎠
β ⎛ GA ⎞
+ o C4 ⎜ γ o3 cos γ o H + γ o cos γ o H ⎟
γo ⎝ EI ⎠
−
(
C4 β oγ o sin γ o H + β o2 sinh β o H ⎛ 3 )
⎜ β o sinh β o H −
GA ⎞
β o sinh β o H ⎟
γ o cos γ o H + β o cosh β o H ⎝
2 2
EI ⎠
⎛ GA ⎞
+ C4 ⎜ β o3 cosh β o H − β o cosh β o H ⎟ = 0
⎝ EI ⎠
GA 2
By assigning α 2 = H , β = β o H , γ = γ o H , C4 = 1 , the above equation is simplified,
EI
βγ sin γ + β 2 sinh β 3
γ 2 cos γ + β 2 cosh β
(γ sin γ + α 2γ sin γ )
(
+ βγ 2 cos γ + α 2 β cos γ )
βγ sin γ + β 2 sinh β 3
−
γ 2 cos γ + β 2 cosh β
(β sinh β − α 2 β sinh β )
(
+ β 3 cosh β − α 2 β cosh β = 0 )
βγ sin γ + β 2 sinh β 3
γ cos γ + β cosh β
2 2
(γ sin γ + α 2γ sin γ )
(
+ βγ 2 cos γ + α 2 β cos γ )
βγ sin γ + β 2 sinh β 3
−
γ 2 cos γ + β 2 cosh β
(β sinh β − α 2 β sinh β )
(
+ β 3 cosh β − α 2 β cosh β = 0 )
β 3 cos γ + βγ 2 cosh β
βγ sin γ + β 2 sinh β 2
+
γ 2 cos γ + β 2 cosh β
(β γ sin γ − βγ 2 sinh β ) = 0
β 2γ 2 (cos 2 γ + sin 2 γ ) + (β 4 + γ 4 )cos γ cosh β
( ) ( )
+ β 2γ 2 cosh 2 β − sinh 2 β + βγ β 2 − γ 2 sin γ sinh β = 0
(( )2
)
2 β 2γ 2 + β 2 − γ 2 + 2 β 2γ 2 cos γ cosh β + βγα 2 sin γ sinh β = 0
⎛ α4 ⎞ α2
Therefore 2 + ⎜⎜ 2 + 2 2 ⎟⎟ cos γ cosh β + sin γ sinh β = 0 (24)
⎝ β γ ⎠ βγ
⎛ α4 ⎞ α2
2 + ⎜⎜ 2 + 2 ⎟ cos γ cosh α 2
+ γ 2
+ sin γ sinh α 2 + γ 2 = 0 (25)
⎝ ( α +γ γ ⎠
2 2 ⎟
) γ α +γ
2 2
where
169
GA
α= H
EI
4mH 4ω 2
−α + α +
2 4
γ = EI
2
γ is a function of the natural frequency ω , so Eq. (25) is actually the governing equation of
the angular frequencies of the model.
170
A.2 Model under uniform lateral loading
n
w
EI GA EI GA
∂M
M+ dx
∂x
∂V f ∂Vs
Vf + dx Vs + dx
∂x ∂x
χ
w dx ndx
Vf
Vs
171
A free-body diagram of the flexural-shear cantilever under uniform lateral load is shown in
Fig. A2. Consider the equilibrium of both flexural and shear elements in the flexural-shear
cantilever, and two equilibrium equations in horizontal direction can be developed.
For the flexural element,
∂V f
Vf + dx − V f + ndx − wdx = 0 (26)
∂x
and for the shear element,
∂Vs
Vs + dx − Vs − ndx = 0 (27)
∂x
By combining the equations of equilibrium of the elements presented by Eqs (26) and (27) and
then substituting the shear forces given by Eqs (3) and (4), the governing equation of deflection
of the continuum flexural-shear model under uniform lateral load can be derived and simplified
as
∂4 y ∂2 y
EI 4 − GA 2 = w (28)
∂x ∂x
GA 2
By assigning α 2 = H , Eq. (28) is expressed as
EI
2
(4 ) ⎛α ⎞ w
y − ⎜ ⎟ y '' = (29)
⎝H⎠ EI
Eq. (29) is an ordinary differential equation and the corresponding homogeneous solution can
be solved by letting yh = Aesx in Eq. (29)
2
⎛α ⎞
As e − ⎜ ⎟ As 2e sx = 0
4 sx
⎝H⎠
2
⎛α ⎞
s −⎜ ⎟ = 0
2
⎝H ⎠
α
s=±
H
Therefore the homogenous solution is
αx αx
− αx αx
yh = A1e H + A2e H
= C3 cosh + C4 sinh (30)
H H
The particular solution of Eq. (29) can be solved by considering,
wH 2
y =−''
p
EIα 2
wH 2 x
y 'p = − + C2
EIα 2
172
Therefore the particular solution is
wH 2 x 2
yp = − + C2 x + C1 (31)
2 EIα 2
Then the general solution to Eq. (29) is obtained from summation of Eqs (30) and (31)
αx αx wH 2 x 2
y = C1 + C2 x + C3 cosh + C4 sinh − (32)
H H 2 EIα 2
The derivatives of the deflection with respected to x are
⎛α ⎞ αx ⎛α ⎞ αx wH 2 x
y′( x ) = C2 + C3 ⎜ ⎟ sinh + C4 ⎜ ⎟ cosh − (33)
⎝H⎠ H ⎝H⎠ H EIα 2
2 2
⎛α ⎞ αx ⎛α ⎞ αx wH 2
y′′(x ) = C3 ⎜ ⎟ sinh + C4 ⎜ ⎟ cosh − (34)
⎝H⎠ H ⎝H⎠ H EIα 2
3 3
⎛α ⎞ αx ⎛α ⎞ αx
y′′′( x ) = C3 ⎜ ⎟ cosh + C4 ⎜ ⎟ sinh (35)
⎝H⎠ H ⎝H⎠ H
The unknown constants Ci in Eq. (32) are solved from the boundary conditions of the model.
⎛α ⎞
Therefore C2 = −C4 ⎜ ⎟ (37)
⎝H⎠
3. Zero moment at the roof,
M (H ) = 0
EIy′′(H ) = 0
2 2
⎛α ⎞ ⎛α ⎞ wH 2
C3 ⎜ ⎟ cosh α + C4 ⎜ ⎟ sinh α − =0
⎝H⎠ ⎝H⎠ EIα 2
wH 4 C sinh α
Therefore, C3 = − 4 (38)
EIα cosh α
4
cosh α
173
4. Zero resultant shear force at the roof,
V f (H ) + Vs (H ) = 0
EIy′′′(H ) − GAy′(H ) = 0
2
⎛α ⎞
y′′′(H ) − ⎜ ⎟ y′(H ) = 0
⎝H⎠
3 3
⎛α ⎞ ⎛α ⎞
C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α
⎝H⎠ ⎝H⎠
2
⎛α ⎞ ⎛ ⎛α ⎞ ⎛α ⎞ wH 3 ⎞
−⎜ ⎟ ⎜⎜ C2 + C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α − ⎟=0
⎝H⎠ ⎝ ⎝H⎠ ⎝H⎠ 2 EIα 2 ⎟⎠
Substituting Eqs (36), (38) and (39) into the above equation gives,
3
⎛ α ⎞ wH
C4 ⎜ ⎟ + =0
⎝H⎠ EI
wH 4
Therefore C4 = − (40)
EIα 3
The coefficients Ci in Eq. (32) are given by
wH 4 wH 4 sinh α
C1 = − −
EIα 4 cosh α EIα 3 cosh α
wH 3
C2 =
EIα 2
wH 4 wH 4 sinh α
C3 = +
EIα 4 cosh α EIα 3 cosh α
wH 4
C4 = −
EIα 3
Substituting the above coefficients into equation (32), leads to
⎡ wH 4 wH 4 sinh α wH 3 x ⎤
⎢ − − + ⎥
⎢ EIα 4
cosh α EIα 3
cosh α EIα 2 ⎥
⎢ ⎛ wH 4 wH 4 sinh α ⎞ αx ⎥
y (x ) = ⎢+ ⎜⎜ + ⎟⎟ cosh ⎥
⎢ ⎝ EIα cosh α EIα cosh α ⎠
4 3
H⎥
⎢ wH 4 αx wH 2 x 2 ⎥
⎢− sinh − ⎥
⎣ EIα 3 H 2 EIα 2 ⎦
⎡ 1 α sinh α α 2 x ⎤
4 ⎢
− − + ⎥
wH ⎢ cosh α cosh α H ⎥
=
EIα 4 ⎢ ⎛ 1 α sinh α ⎞ αx αx α 2 x 2 ⎥
⎢+ ⎜ cosh α + cosh α ⎟ cosh H − α sinh H − 2 H 2 ⎥
⎣ ⎝ ⎠ ⎦
174
Therefore the equation of deflection of the model under uniform loading is
wH 4 ⎡α sinh α + 1 ⎛ αx ⎞ αx 2⎛ x x 2 ⎞⎤
y(x ) = ⎢ ⎜ cosh − 1⎟ − α sinh + α ⎜ H 2 H 2 ⎟⎟⎥
⎜ − (41)
EIα 4 ⎣ cosh α ⎝ H ⎠ H ⎝ ⎠⎦
The roof displacement is then given by from Eq. (41)
wH 4 ⎡ α 2 α sinh α + 1⎤
y (H ) = ⎢1 + − ⎥ (42)
EIα 4 ⎣ 2 cosh α ⎦
175
A.3 Model under inverted triangular lateral loading
q
w
EI GA EI GA
Similar to the derivation of Appendix A.2 Eq. (29), the governing equation of deflection is
2
(4 ) ⎛α ⎞ x w
y − ⎜ ⎟ y '' = (49)
⎝H⎠ H EI
The particular solution of Eq. (49) can be solved by considering the follows,
wHx
y′p′ = −
EIα 2
wHx 2
′
yp = − + C2
2 EIα 2
Therefore the particular solution is
wHx 3
yp = − + C 2 x + C1 (50)
6 EIα 2
Then the general solution to Eq. (49) is obtained from summation of Eqs (30) and (50)
wHx 3
y = C1 + C 2 x + C 3 cosh αx + C 4 sinh αx − (51)
6 EIα 2
The derivatives of the deflection with respected to x are
⎛α ⎞ ⎛α ⎞ wHx 2
y ′( x ) = C2 + C3 ⎜ ⎟ sinh αx + C4 ⎜ ⎟ cosh αx − (52)
⎝H⎠ ⎝H ⎠ 2 EIα 2
176
2 2
⎛α ⎞ αx ⎛α ⎞ αx wHx
y ( x ) = C3 ⎜ ⎟ cosh + C4 ⎜ ⎟ sinh
''
− (53)
⎝H⎠ H ⎝H⎠ H EIα 2
3 3
⎛α ⎞ αx ⎛α ⎞ αx wH
y′′′( x ) = C3 ⎜ ⎟ sinh + C4 ⎜ ⎟ cosh − (54)
⎝H⎠ H ⎝H ⎠ H EIα 2
The unknown constants Ci in Eq. (32) are solved from the boundary conditions of the model.
1. Zero displacement at the fixed support,
y (0) = 0
C1 + C3 = 0
⎛α ⎞
Therefore C2 = −C4 ⎜ ⎟ (56)
⎝H⎠
3. Zero moment at the roof
M (H ) = 0
EIy′′(H ) = 0
2 2
⎛α ⎞ ⎛α ⎞ wH 2
C3 ⎜ ⎟ cosh α + C4 ⎜ ⎟ sinh α − =0
⎝H⎠ ⎝H⎠ EIα 2
wH 4 C sinh α
Therefore C3 = − 4 (57)
EIα cosh α
4
cosh α
4. Zero resultant shear force at the roof
V f (H ) + Vs (H ) = 0
EIy′′′(H ) − GAy′(H ) = 0
2
⎛α ⎞
y′′′(H ) − ⎜ ⎟ y′(H ) = 0
⎝H⎠
3 3
⎛α ⎞ ⎛α ⎞ wH
C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α −
⎝H⎠ ⎝H⎠ EIα 2
2
⎛α ⎞ ⎛ ⎛α ⎞ ⎛α ⎞ wH 3 ⎞
−⎜ ⎟ ⎜⎜ C2 + C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α − ⎟=0
⎝H⎠ ⎝ ⎝H⎠ ⎝H⎠ 2 EIα 2 ⎟⎠
177
Substituting Eqs (55), (56) and (57) into the above equation will obtain the follows,
3
⎛ α ⎞ wH wH
C4 ⎜ ⎟ + − =0
⎝ H ⎠ 2 EI EIα
2
wH 4 wH 4
Therefore C4 = − + (58)
2 EIα 3 EIα 5
The coefficients Ci in Eq. (51) are given by
wH 4 wH 4 sinh α wH 4 sinh α
C1 = − − +
EIα 4 cosh α 2 EIα 3 cosh α EIα 5 cosh α
wH 3 wH 3
C2 = −
2 EIα 2 EIα 4
wH 4 wH 4 sinh α wH 4 sinh α
C3 = + −
EIα 4 cosh α 2 EIα 3 cosh α EIα 5 cosh α
wH 4 wH 4
C4 = − +
2 EIα 3 EIα 5
Substituting the above coefficients into equation (51), leads to
⎡⎛ α sinh α + 2 sinh α ⎞⎛ αx ⎞ ⎤
⎢ ⎜ − ⎟⎜ cosh ⎟ ⎥
wH 4 ⎢⎝ 2 cosh α α cosh α ⎠⎝ H⎠ ⎥
y(x ) = (59)
EIα 4 ⎢ ⎛ 1 α ⎞ ⎞ x⎥
3
αx α ⎛ x ⎞ ⎛ α
2 2
⎢+ ⎜ − ⎟ sinh − ⎜ ⎟ + ⎜⎜ − 1⎟⎟ ⎥
⎢⎣ ⎝ α 2 ⎠ H 6 ⎝H⎠ ⎝ 2 ⎠ H ⎥⎦
The roof displacement is then given by from Eq. (59),
wH 4 ⎡ sinh α α 2 α sinh α + 2 ⎤
y (H ) = ⎢ + − ⎥ (60)
EIα 4 ⎣ α cosh α 3 2 cosh α ⎦
178
The moment in flexural cantilever is
⎡⎛ α sinh α + 2 sinh α ⎞ αx ⎤
2⎢⎜ 2 cosh α − α cosh α ⎟ cosh H ⎥
wH ⎝ ⎠
M ( x ) = EIy′′( x ) = 2 ⎢ ⎥ (64)
α ⎢ ⎛1 α⎞ αx x ⎥
⎢+ ⎜ − ⎟ sinh − ⎥
⎣ ⎝α 2 ⎠ H H ⎦
The shear in flexural cantilever is
⎡⎛ α sinh α + 2 sinh α ⎞ αx ⎤
⎢⎜ 2 cosh α − α cosh α ⎟ sinh H ⎥
wH ⎢⎝ ⎠
V f (x ) = EIy′′′(x ) = ⎥ (65)
α ⎢ ⎛1 α⎞ αx 1 ⎥
⎢+ ⎜ − ⎟ cosh − ⎥
⎣ ⎝α 2 ⎠ H α ⎦
The shear in shear cantilever is
⎡⎛ α sinh α + 2 sinh α ⎞ αx ⎤
⎢⎜ 2 cosh α − α cosh α ⎟ sinh H ⎥
wH ⎢⎝ ⎠
Vs ( x ) = GAy′( x ) = ⎥ (66)
α ⎢ ⎛1 α⎞ αx α ⎛ x ⎞ ⎛ α 1 ⎞⎥
2
⎢+ ⎜ − ⎟ cosh − ⎜ ⎟ + ⎜ − ⎟⎥
⎢⎣ ⎝ α 2 ⎠ H 2 ⎝ H ⎠ ⎝ 2 α ⎠⎥⎦
179
A.4 Model under point load at the roof
P q
EI GA EI GA
Similar to the derivation of Appendix A.2 Eq. (29), the governing equation of deflection is
2
⎛α ⎞
y (4 ) − ⎜ ⎟ y '' = 0 (67)
⎝H⎠
The particular solution of Eq. (49) can be solved by considering,
y ′p′ = 0
y ′p = C 2
Therefore the particular solution is
y p = C2 x + C1 (68)
Then the general solution to Eq. (49) is obtained from summation of Eqs (30) and (50),
αx αx
y = C1 + C2 x + C3 cosh + C4 sinh (69)
H H
The derivatives of the deflection with respected to x are
α αx α αx
y′(x ) = C2 + C3 sinh + C4 cosh (70)
H H H H
2 2
⎛α ⎞ αx ⎛α ⎞ αx
y′′(x ) = C3 ⎜ ⎟ cosh + C4 ⎜ ⎟ sinh (71)
⎝H ⎠ H ⎝H⎠ H
3 3
⎛α ⎞ αx ⎛α ⎞ αx
y′′′( x ) = C3 ⎜ ⎟ sinh + C4 ⎜ ⎟ cosh (72)
⎝H⎠ H ⎝H ⎠ H
180
The unknown constants Ci in Eq. (32) are solved from the boundary conditions of the model.
1. Zero displacement at the fixed support,
y (0) = 0
C1 + C3 = 0
⎛α ⎞
Therefore C2 = −C4 ⎜ ⎟ (74)
⎝H⎠
5. Zero moment at the roof,
M (H ) = 0
EIy′′(H ) = 0
2 2
⎛α ⎞ ⎛α ⎞
C3 ⎜ ⎟ cosh α + C4 ⎜ ⎟ sinh α = 0
⎝H⎠ ⎝H⎠
Therefore C3 = −C4 tan α (75)
6. Zero resultant shear force at the roof,
V f (H ) + Vs (H ) = 0
EIy′′′(H ) − GAy′(H ) = 0
2
⎛α ⎞
y′′′(H ) − ⎜ ⎟ y′(H ) = 0
⎝H⎠
3 3
⎛α ⎞ ⎛α ⎞
C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α
⎝H⎠ ⎝H⎠
2
⎛α ⎞ ⎛ ⎛α ⎞ ⎛α ⎞ ⎞ P
−⎜ ⎟ ⎜⎜ C2 + C3 ⎜ ⎟ sinh α + C4 ⎜ ⎟ cosh α ⎟⎟ =
⎝H⎠ ⎝ ⎝H⎠ ⎝H⎠ ⎠ EI
Substituting Eqs (73), (74) and (75) into the above equation gives,
3
⎛α ⎞ P
C4 ⎜ ⎟ =
⎝H ⎠ EI
PH 3
Therefore C4 = (76)
EIα 3
181
The coefficients Ci in Eq. (69) are given by
PH 3 tan α
C1 =
EIα 3
PH 2
C2 = −
EIα 2
PH 3 tan α
C3 = −
EIα 3
PH 3
C4 =
EIα 3
Substituting the above coefficients into equation (51), get
PH 3 PH 2 PH 3 αx PH 3 αx
y (x ) = tanh α − x− tanh α cosh + sinh (77)
EIα 3
EIα 2
EIα 3
H EIα 3
H
The roof displacement is then given by from Eq. (77)
PH 3 PH 3
y (H ) = tanh α − (78)
EIα 3 EIα 2
The roof drift ratio from Eq. (77) is
y (H ) PH 2 PH 2
RΔ = = tanh α − (79)
H EIα 3 EIα 2
The interstorey drift ratio from Eq. (77) is
PH 2 PH 2 αx PH 2 αx
Rδ ( x ) = − − tanh α sinh + cosh (80)
EIα 2
EIα 2
H EIα 2
H
The curvature from Eq. (77) is
PH αx PH αx
cur ( x ) = − tanh α cosh + sinh (81)
EIα H EIα H
The moment in flexural cantilever is
PH ⎛ αx αx ⎞
M (x ) = EIy′′(x ) = ⎜ − tanh α cosh + sinh ⎟ (82)
α ⎝ H H⎠
The shear in flexural cantilever is
⎛ αx αx ⎞
V f (x ) = EIy′′′(x ) = p⎜ − tanh α sinh + cosh ⎟ (83)
⎝ H H⎠
The shear in shear cantilever is
⎛ αx αx ⎞
Vs ( x ) = GAy′(x ) = p⎜ − 1 − tanh α sinh + cosh ⎟ (84)
⎝ H H⎠
182