This action might not be possible to undo. Are you sure you want to continue?
BYG·DTU
R064
2003
ISSN 16012917
ISBN 8778771299
João Luís Domingues Costa
Standard Methods for Seismic
Analyses
DANMARKS
T E K N I S K E
UNIVERSITET
João Luís Domingues Costa
Standard Methods for Seismic
Analyses
Report
BYG·DTU R064
2003
ISSN 16012917
ISBN 8778771269
Contents
ACKNOWLEDGEMENTS...........................................................................................3
1. INTRODUCTION.....................................................................................................5
2. STRUCTURAL DYNAMICS FUNDAMENTALS.....................................................7
2.1 FORMULATION OF EQUATION OF MOTION FOR SDOF SYSTEMS ...........................7
2.2 UNDAMPED FREE VIBRATIONS...........................................................................8
2.3 DAMPED FREE VIBRATIONS................................................................................9
2.4 RESONANT RESPONSE....................................................................................11
2.5 BASE MOTION FOR SDOF SYSTEMS................................................................14
2.6 FORMULATION OF THE EQUATION OF MOTION FOR MDOF SYSTEMS...................15
2.7 FREQUENCY AND VIBRATION MODE SHAPE ANALYSIS .......................................17
2.8 ORTHOGONALITY CONDITIONS.........................................................................19
Orthogonality with respect to the mass matrix .....................................19
Orthogonality with respect to the stiffness matrix.................................19
2.9 MODAL COORDINATES....................................................................................20
2.10 EQUATION OF MOTION IN MODAL COORDINATES...............................................21
2.11 BASE MOTION FOR MDOF SYSTEMS...............................................................23
2.12 VIBRATION ANALYSIS BY THE RAYLEIGH METHOD..............................................26
Basic concepts.....................................................................................26
Approximate analysis of a general system; Selection of the vibration
shape...................................................................................................26
3. SEISMIC ANALYSIS BY RESPONSE SPECTRA................................................31
3.1 RESPONSE SPECTRUM CONCEPT ....................................................................31
3.2 RESPONSE SPECTRUM ANALYSIS APPLIED TO MDOF SYSTEMS........................33
1) SRSS (Square Root of Sum of Squares) ......................................34
2) CQC (Complete Quadratic Combination)......................................35
3.3 DUCTILE BEHAVIOUR CONSIDERATION .............................................................35
4. SEISMIC RESPONSE BY TIMEHISTORY ANALYSIS.......................................39
4.1 RESPONSE OF A SDOF SYSTEM TO GENERAL DYNAMIC LOADING; DUHAMEL’S
INTEGRAL ......................................................................................................39
4.2 LINEAR TIME HISTORY ANALYSIS FOR MDOF SYSTEMS ....................................41
4.3 TIME HISTORY ANALYSIS FOR EARTHQUAKES ...................................................42
Stepbystep integration method with linear variation of the load.........43
Standard Methods for Seismic Analyses
5. EQUIVALENT STATIC METHOD.........................................................................45
6. CASE STUDY.......................................................................................................47
6.1 STRUCTURAL MODEL OF THE BRIDGE...............................................................47
6.2 FREQUENCIES AND VIBRATION MODE SHAPE DETERMINATION FOR THE BRIDGE..49
6.3 RESPONSE SPECTRUM ANALYSIS OF THE BRIDGE.............................................51
6.4 RESULTS OF THE RESPONSE SPECTRUM ANALYSIS...........................................53
Internal forces due to earthquake loading in horizontal direction .........53
Internal forces due to earthquake loading in vertical direction .............54
Displacements .....................................................................................54
Combination of Orthogonal Seismic Effects.........................................54
6.5 TIMEHISTORY RESPONSE ANALYSIS OF THE BRIDGE........................................54
6.6 RESULTS OF THE TIMEHISTORY RESPONSE ANALYSIS......................................55
6.7 EQUIVALENT STATIC ANALYSIS OF THE BRIDGE.................................................58
REFERENCES..........................................................................................................60
2
Standard Methods for Seismic Analyses
Acknowledgements
The work has been carried out at the Department of Structural Engineering and
Materials, Technical University of Denmark (BYG• DTU) under the supervision of
Professor, Dr. techn. M. P. Nielsen.
The author would like to thank his supervisor for giving valuable advice and
inspiration as well as valuable criticism to the present work.
Thanks are also due to the author’s cosupervisor M.Sc. Ph.D. Rita Bento, Instituto
Superior Técnico, Lisbon, Portugal, who has given important and useful comments
and suggestions.
A word of appreciation should also be addressed to Civil Engineer Ph D Junying Liu,
COWI A/S, for providing the example used in the case study and for assistance in
carrying out the study.
The Portuguese institution for scientific research Fundação para a Ciência e
Tecnologia – FCT, sponsors the Ph.D. project under which this report was done. The
author grateful acknowledges this support.
Lyngby, July 2003
João Luís Domingues Costa
3
Standard Methods for Seismic Analyses
4
Standard Methods for Seismic Analyses
1. Introduction
The following report gives a general survey of the most important methods nowadays
at disposal for a structural engineer when performing a seismic design of a given
structure.
The methods to be discussed are the response spectrum method and the linear time
history analysis. The first one is widely used as it applies to the major part of a
seismic analysis necessary for design purpose. The timehistory response method
provides more detailed information regarding the seismic behaviour of a structure
and is therefore used for more specific earthquake analyses. Both methods assume
linear behaviour of the structure, i.e. proportionality between deformations and
forces. For its simplicity, the static equivalent method, usually used in the predesign
phase of regular structures, is also introduced.
The theoretical information given in this report is complemented with analysis of a
bridge similar to one designed for the High Speed Transportation System in Taiwan.
This document is intended for students or civil engineers who want to have a basic
knowledge about earthquake analysis. Before discussing seismic analysis in
particular, the reader is introduced to some of the corresponding basic concepts from
elementary Structural Dynamics.
It should be noted that this report does not intend to be neither a reference book nor
a Structural Dynamics or Earthquake Analysis textbook. For further study a number
of references are given.
5
Standard Methods for Seismic Analyses
6
Standard Methods for Seismic Analyses
2. Structural Dynamics Fundamentals
2. 1 Formul ati on of Equati on of Moti on for SDOF Systems
The essential properties of any linearly elastic structural system subjected to
dynamical loads include its mass, m, its elastic characteristics (stiffness), k, and its
energy loss mechanism (damping), characterized by a number c. In dynamical terms,
a system is called a Single Degree of Freedom (SDOF) system if all these properties
may be modelled by a physical element with only one component of displacement, q.
See figure 1 a).
Figure 1 – a) Simplified sketch of a SDOF system; b) Dynamical equilibrium of a SDOF system
The primary objective in a structural dynamical analysis is to evaluate the time
variation of the displacements and to accomplish this the Equation of Motion must be
formulated and solved.
One of the methods to formulate the Equation of Motion
1
is direct use of Newton’s
second law, which implies that the mass develops an inertia force, f
i
, proportional to
its acceleration and opposing the acceleration. The dynamical equilibrium condition is
given by (2.1)
2
.
( ) ( )
( ) ( ) 0
0
= ⋅ − ⇔
⇔ = −
t q m t F
t f t F
i
& &
(2.1)
Referring to Figure 1 b) the resultant force acting on the mass, F(t), may be defined
as the difference between the external loads p(t) and the sum of the elastic forces, f
k
,
and the damping forces, f
c
. The equilibrium condition may then be written as follows:
( ) t p f f f
k c i
= + + (2.2)
Elastic forces, f
k
, are determined using Hooke’s law:
( ) t q k f
k
⋅ = (2.3)
1
Chapter 15 and chapter 22 of reference 1 on the formulation of the Equation of Motion is
recommended.
2
A dot means differentiation with respect to time
7
Standard Methods for Seismic Analyses
Damping forces, f
c
, of the viscous type are proportional to the velocity, i.e.
3
.
( ) t q c f
c
&
⋅ = (2.4)
Introducing equations (2.3) and (2.4) into equation (2.2) one may write the
equilibrium condition in terms of the coordinate q(t), the system properties, m, k and c
and the external dynamical loads as follows:
( ) ( ) ( ) ( ) t p t q k t q c t q m = ⋅ + ⋅ + ⋅
& & &
(2.5)
This last expression is known as the Equation of Motion of a SDOF system.
2. 2 Undamped Free Vi brati ons
The motion of a SDOF system free from external action or forces is governed by the
initial conditions. If damping is disregarded the equation of motion (2.5) is of the form:
( ) ( ) 0 t q k t q m = ⋅ + ⋅
& &
(2.6)
This is a homogeneous second order linear differential equation with constant
coefficients.
Considering solutions of the form
( ) ( ) t A t q ⋅ ⋅ = ω cos (2.7)
or
( ) ( ) t sin B t q ⋅ ⋅ = ω (2.8)
where A and B are constants, one may easily verify by direct substitution, that these
are solutions to the differential equation (2.6). For instance, the substitution of
equation (2.7) into (2.6) leads to:
( ) ( ) 0 t A k m 
2
= ⋅ ⋅ ⋅ + ⋅ ω ω cos (2.9)
In order to satisfy this condition at any time t, the term in the first parenthesis must be
equal to zero, giving:
m
k
= ω (2.10)
Since the differential equation (2.6) is linear and homogeneous, the superposition of
the two solutions above is also a solution. Therefore one may write the general
solution as:
( ) ( ) ( ) t cos B t sin A t q ⋅ ⋅ + ⋅ ⋅ = ω ω (2.11)
3
Damping forces are always present in any physical system undergoing motion. These forces are part
of a mechanism transforming the mechanical energy of the system to other forms of energy such as
heat. The mechanism is quite complex and still not completely understood. Therefore the damping
influence is usually quantified on the basis of experience.
8
Standard Methods for Seismic Analyses
The constants of integration A and B may be expressed in terms of the initial
conditions, i.e. the displacement, q(0), and the velocity, ( ) 0
&
q , at time t=0. Thus the
solution becomes:
( )
( )
( ) ( ) ( t cos 0 q t sin
0 q
t q
.
⋅ ⋅ + ⋅ ⋅ = ω ω )
ω
&
(2.12)
This last equation is the equation for the motion of an undamped SDOF system
under freevibration conditions. This is a simple harmonic motion, in which the
quantity ω is the circular frequency. Dividing ω by the factor π ⋅ 2 , one obtains the
natural frequency of the system, f, expressed in Hz (cycles per second). As shown by
expression (2.10), this parameter only depends on the system properties k and m.
Expression (2.12) may be used qualitatively to understand how the response is
influenced by the stiffness and inertia properties of the system as well as the initial
conditions:
A very stiff (or very “light”) SDOF system has a large value of k (or low value
for m), and so the response frequency is high and the displacements are
mainly given by ( ) ( ) ( ) t q t q ⋅ ⋅ = ω cos 0 & . Consequently the maximum displacement
will be of the same order as the initial displacement, q(0);
A very flexible (or very “heavy”) SDOF system has a large value for m (or low
value of k). The response frequency is low and the maximum displacement is
mainly governed by ( )
( )
( t sin
0 q
t q ⋅ ⋅ = ω )
ω
&
& . This implies that the maximum
displacements may be larger than the initial displacement, q(0).
2. 3 Damped free vi brati ons
We now discuss a SDOF system vibrating freely but we include the effect of the
damping forces. The equation of motion (2.5) then has the form:
( ) ( ) ( ) 0 t q k t q c t q m = ⋅ + ⋅ + ⋅
& & &
(2.13)
This differential equation is of the same form as before for the undamped case, but
the solution now is:
( )
t s
e C t q
⋅
⋅ = (2.14)
where C is a constant. This is proved substituting (2.14) into (2.13) which leads to
( ) 0 e C k s c s m
t s 2
= ⋅ ⋅ + ⋅ + ⋅
⋅
(2.15)
Requiring the parenthesis to be zero we get:
0 = + ⋅ + ⋅ k s c s m
2
(2.16)
The roots of this quadratic equation are:
9
Standard Methods for Seismic Analyses
2
2
2
1
2 2
ω − 
.

\

⋅
±
⋅
− =
)
`
¹
m
c
m
c
S
S
(2.17)
As in the previous paragraph, the general solution is given by superposition of the
two possible solutions:
( )
t s t s
e C e C t q
⋅ ⋅
⋅ + ⋅ =
2 1
2 1
(2.18)
Depending on the value of c, one gets three types of motion, according to the
quantity under the squareroot sign being positive, negative or zero.
The value making the squareroot quantity zero is called the critical damping value,
ω ⋅ ⋅ = m c
c
2 , and it may be shown that this value represents the largest value of
damping that leads to oscillatory motion in free response. Structural systems under
normal conditions do not have values of damping above this critical value. So, in the
following, only the situation for underdamped systems will be discussed, i.e. systems
with damping below the critical value.
Under these conditions, equation (2.18) may be written in a more convenient form,
introducing the parameters:
ξ , which is the damping ratio to the critical damping value i.e.
ω
ξ
⋅ ⋅
=
m 2
c
d
ω , the damped vibration frequency, i.e.
2
d
1 ξ ω ω − ⋅ =
( ) ( ) ( )   t sin B t A e t q
d d
t
⋅ ⋅ + ⋅ ⋅ =
⋅ ⋅ −
ω ω
ω ξ
(2.19)
Finally, when the initial condition of displacement, q , and velocity, q , are
introduced, the constant of integration A and B can be evaluated and substituted into
equation (2.19), giving:
0 0
&
( ) ( ) (
(
¸
(
¸
⋅ ⋅
⋅ ⋅ +
+ ⋅ ⋅ =
⋅ ⋅ −
t sin
q q
t cos q e t q
d
d
0 0
d 0
t
ω
ω
ω ξ
ω
ω ξ
&
) (2.20)
The term in parenthesis represents simple harmonic motion, as it is of the same form
as equation (2.12). It is of interest to note that the frequency for this harmonic motion
is now given by
d
ω with the expression as above. For common structural systems
(ξ <20%) this value differs very little from the undamped frequency as shown by
equation (2.10), so it may be inferred that, for normal conditions, damping will not
have any significant influence on the frequency of motion.
The effect of damping is more evident when considering the successive peak
responses (see figure 2). It may be shown that the ratio between two successive
peaks, q
n
and q
n+1
, is given approximately by:
ξ π⋅ ⋅ − +
≅
2 1
e
q
q
n
n
(2.21)
10
Standard Methods for Seismic Analyses
Figure 2 – Plot of a freevibration response equation of motion for underdamped SDOF systems
q
n
t
q(t)
q
n+1
We may now formulate the equation of motion for SDOF systems by introducing the
damping ratio, ξ , and the natural vibration frequency, ω :
( ) ( ) ( )
( )
m
t p
t q t q 2 t q
2
= ⋅ + ⋅ ⋅ ⋅ + ω ω ξ
& & &
(2.22)
2. 4 Resonant Response
4
To explain this important phenomenon, taking place when a structure is submitted to
dynamical loading, response to harmonic loading will be considered.
The simplest load of this type is of the form:
( ) ( ) t sin p t p ⋅ ⋅ = ω
0
(2.23)
where p
0
is the maximum value and ω its frequency
The equation of motion (2.5) may now be written as follows:
( ) ( ) ( ) ( ) t sin p t q k t q c t q m
0
⋅ ⋅ = ⋅ + ⋅ + ⋅ ω
& & &
(2.24)
4
The study of SDOF systems cannot be completed without discussing the equations of motion for
harmonic and periodic loading. However these subjects are not directly related to the standard
methods for seismic design to be presented in this document. The reader is referred to, for example,
chapters 4 and 5 of reference 1 or chapter 3 of reference 2.
11
Standard Methods for Seismic Analyses
One has now a nonhomogenous differential equation which solution is of the form:
( ) ( ) ( )  
( ) ( ) (  
( ) ( )
)
2
2
2
2
0
d d
t
2 1
t cos 2 t sin 1
k
p
t cos B t sin A e t q
β ξ β
ω β ξ ω β
ω ω
ω ξ
⋅ ⋅ + −
⋅ ⋅ ⋅ ⋅ − ⋅ ⋅ −
⋅ +
+ ⋅ ⋅ + ⋅ ⋅ ⋅ =
⋅ ⋅ −
(2.25)
Here:
A and B have the same meaning as before i.e. they depend on the initial conditions.
The parameter β is defined as the ratio
ω
ω
β = .
The first term in (2,25) is called the transient response and because of its
dependence on the factor e ,it damps out quickly. Therefore its evaluation is of
little interest for the present discussion.
t ⋅ ⋅ − ω ξ
The second term is called the steadystate response and it may be written in a more
convenient form:
( ) ( ) θ ω ρ − ⋅ ⋅ = t sin t q (2.26)
The term ρ is the amplitude, i.e. the maximum value of the displacement. It may be
shown that this value is given in terms of the static displacement
k
p
0
multiplied by the
factor, D, which is called dynamical magnification factor:
D
k
p
0
⋅ = ρ (2.27)
with D expressed as:
( ) ( )
2
2
2
2 1
1
D
β ξ β ⋅ ⋅ + −
= (2.28)
The value θ is called the phase angle and describes how the response lags behind
the applied load:


.

\

−
⋅ ⋅
=
2
1 
1
2
tan
β
β ξ
θ (2.29)
Several plots of the dynamical magnification factor with respect to β are shown in
figure 3 for values of damping, ξ , usually found in common structures.
As it may be seen the peak values of D are reached when β is very close to 1 (in
fact, when
2
2 1 ξ β ⋅ − = ). This means that when the load frequency approaches the
natural vibration frequency of the SDOF system, the response will increase more and
more. This phenomenon is called resonance.
12
Standard Methods for Seismic Analyses
Figure 3 – The dynamical magnification factor D as a function of β
0
4
8
12
0 1 2 3
ξ=2%
ξ=5%
ξ=10%
ξ=20%
D
β
Substituting the value of β for which D is maximum, one has the following
expression for the maximum response, q
max
:
k
p
1 2
1
q
0
2
max
⋅
− ⋅ ⋅
=
ξ ξ
(2.30)
The effect of damping on the resonant response is seen clearly: The lower is the
damping value, ξ , the bigger the response. Theoretically for undamped conditions
the value is infinite.
The physical explanation for resonance is of course that both load frequency and
natural vibration frequency of the system are so close that most part of the time the
response and the load signals are in the same phase. This means that when the
system is moving in a certain direction the load is in the same direction. This will lead
to a consecutive amplification of the response in each cycle until the limit given by
expression (2.30) is reached. For undamped conditions the response will grow
indefinitely.
It should be also noticed that for values of β near 0, i.e. when the natural vibration
frequency of the system is much higher than the load frequency, D approaches unity.
This means that the response will be closer to the static response. In fact, for highly
stiff systems the quantity ( ) t q k ⋅ is expected to play an important role in the final
response.
13
Standard Methods for Seismic Analyses
2. 5 Base Moti on for SDOF Systems
Figure 4 shows a sketch of a SDOF system when submitted to base motion.
Figure 4 – SDOF system submitted to base motion
When a SDOF system is submitted to base motion, one may write the absolute
displacement, q, in terms of the sum of the relative displacement, q*, and the support
displacement, q
s
(figure 4).
s
q q q + = * (2.31)
The formulation of the equation of motion leads to the same form as (2.2). However it
should be noted that no load is acting on the system. The only action able to induce
deformation on the system is the support displacement, . Therefore, as in (2.2),
one may write the dynamical equilibrium condition:
s
q
0 = + +
k c i
f f f (2.32)
Here:
Inertia forces, f
k
, are in terms of absolute coordinates, . q
& &
Elastic forces, f
k
, and damping forces, f
c
, are in terms of relative coordinates,
and q , respectively. * q *
&
( ) ( ) ( ) 0 * * = ⋅ + ⋅ + ⋅ t q k t q c t q m
& & &
(2.33)
By means of (2.31) it’s possible to write the previous equation in terms of relative
coordinates. This is more convenient for the purpose of achieving the effects on the
system due to base motion:
( ) ( ) ( ) ( ) t q m t q k t q c t q m
s
& & & & &
⋅ − = ⋅ + ⋅ + ⋅ * * * (2.34)
Equation (2.34) is of the same form as (2.5). Therefore the response analysis of a
SDOF system submitted to ground motion, in terms of relative coordinates, may be
treated assuming a load applied on the system equal to ( ) (t q m t p
s
& &
⋅ ) − = .
14
Standard Methods for Seismic Analyses
Equation (2.34) may also be formulated in the same way as (2.22):
( ) ( ) ( ) t q t q t * q 2 t q
s
2
& & & & &
− = ⋅ + ⋅ ⋅ ⋅ + ) ( * * ω ω ξ (2.35)
Again we have a nonhomogeneous differential equation and so it is necessary to
find a particular solution, which depends on the form of . In chapter 4, the
solution for base acceleration of general form will be discussed.
( ) t q
s
& &
2. 6 Formul ati on of the Equati on of Moti on for MDOF systems
From the discussion in the previous paragraphs, a degree of freedom is defined as
an independent coordinate, necessary to specify the configuration or position of a
system at any time, q(t).
a) b)
Figure 5 – Examples of MDOF systems
A structural system composed by more than one degree of freedom is called a Multi
Degree of Freedom system (MDOF). Figure 5 shows two examples of MDOF
systems.
The establishment of the equations of motion for several degrees of freedom
proceeds analogously as for the SDOF systems, which leads to a dynamical
equilibrium condition of the same form as (2.2) for each degree of freedom. The
result is a system of N differential equations, in which N is the number of degrees of
freedom.
( )
( )
( ) t p f f f
.. .......... .......... ..........
t p f f f
t p f f f
N N k, N c, N i,
2 k,2 c,2 i,2
1 k,1 c,1 i,1
= + +
= + +
= + +
(2.36)
Each of the resisting forces, f
i,i
, f
c,i
or f
k,i
developed for a certain degree of freedom, i,
is due to the motion of one degree of freedom. For example the elastic force
produced for the degree of freedom 1, f
k,1
, is the sum of the different elastic forces
acting at point 1, each one due to the displacement of each of any of the other
degrees of freedom.
15
Standard Methods for Seismic Analyses
Most conveniently the resisting forces may be expressed by means of a set of
influence coefficients. Considering again the example above one has:
( ) t q k f
i
N
1 i
1i k,1
⋅ =
∑
=
(2.37)
in which k
1i
is called the stiffness influence coefficient. It may be defined as the force
at degree of freedom 1 due to a unit displacement corresponding to degree of
freedom i.
In figure 6 is illustrated the analysis of the stiffness coefficients in a twostorey frame
with masses M
1
and M
2
, bending stiffness of the columns EI and lengths of the
columns L.
( ) ( ) ( ) ( )
3 3
22
3
12 21
3
11
12
2
12
2
12
2
12
2
b
b
a
a
a
a
a
a
L
EI
L
EI
k
L
EI
K k
L
EI
k
⋅
⋅ +
⋅
⋅ =
⋅
⋅ − = =
⋅
⋅ =
Figure 6 – Analysis of frame stiffness coefficients
Analogously one may define the damping forces produced for degree of freedom 1:
q1=1
q2=1
M1
(EI)a, La
(EI)b, Lb
M2
k11
k22
k21
k12
( ) t q c f
i
N
1 i
1i c,1
&
⋅ =
∑
=
(2.38)
in which c
1i
are called the damping influence coefficients. They may be defined as the
force at degree of freedom 1 due to unit velocity of the degree of freedom i.
16
Standard Methods for Seismic Analyses
Finally the inertia forces produced for degree of freedom 1:
( ) t q m f
i
N
1 i
1i i,1
& &
⋅ =
∑
=
(2.39)
in which m
1i
are called the mass influence coefficients and may be defined as the
force at degree of freedom 1 due to unit acceleration corresponding to degree of
freedom i.
It is important to notice that the principle of superposition may be applied only if linear
behaviour is assumed.
The set of equations in (2.36) may be written in matrix form:
  ( ) { }   ( ) { }   ( ) { } ( ) { } t p t q K t q C t q M = ⋅ + ⋅ + ⋅
& & &
(2.40)
This equation is equivalent to (2.5) for a given MDOF system as it expresses the N
equations of motion defining its response
5
. In the following paragraphs until 2.11 the
procedures leading to the solution of this system will be discussed.
2. 7 Frequency and Vi brati on Mode Shape Anal ysi s
The problem of determining the vibration frequencies in MDOF systems is solved as
for SDOF systems, i.e. assuming undamped conditions and no loads applied. For
this situation equation (2.40) is written as follows:
  ( ) { }   ( ) { } 0 t q K t q M = ⋅ + ⋅
& &
(2.41)
By analogy with the behaviour of SDOF systems, it is assumed that the freevibration
motion response is simple harmonic, i.e. of the form (2.12):
( ) { } { } ( ) θ ω + ⋅ ⋅ = t sin q t q (2.42)
Here
{ } q represents the vibration shape of system, (constant in time)
ω is the vibration frequency and
θ the phase angle.
Introducing the equation of motion into (2.41) and observing that
( ) { } { } ( ) θ ω ω + ⋅ ⋅ − = t q t q sin
2
& &
one has (after omitting the sine term):
  { } 0 q M K
2
= ⋅ ⋅ −ω (2.43)
5
For further study of the formulation of the equations of motion for MDOF systems, the reader is
referred to chapter 11 in reference 1 regarding the evaluation of the matrices [M], [C] and [K] .
17
Standard Methods for Seismic Analyses
The only nontrivial solution of this equation is the one making the determinant of the
matrix   M K
2
⋅ −ω equal to 0, (2.44). Thus the problem of determining the frequencies
in a MDOF system results in an eigenvalue problem of the nonstandard form. The
eigenvalues are the squares of the frequencies and the eigenvectors are the
vibration modes associated with the frequencies.
0 M K
2
= ⋅ −ω (2.44)
Equation (2.44) is called the frequency equation for MDOF systems. Expanding the
determinant gives a polynomial expression of the Nth degree. Therefore one has a
set of N solutions ( )
2
N
2
2
2
1
,..., , ω ω ω , each one representing a possible vibration mode.
Each shape vector, { , is determined substituting the corresponding frequency, } q
i
ω ,
into equation (2.43).
The lowest frequency (also called the natural frequency) corresponds to the first
vibration mode, the next higher corresponds to the second vibration mode, etc.
It is of interest to notice that, as for SDOF systems, the frequencies and the
corresponding vibration mode shape depend only on the mass, , and the
stiffness, , of the system.
M
  K
It should also be noticed that the system resulting from substituting a given
frequency,
i
ω , into equation (2.43) is homogeneous, with linear dependent equations
and therefore indeterminate. This means that it’s impossible to determine the
amplitudes of each degree of freedom in the corresponding vibration shape by simply
resorting to equation (2.43). Only ratios between these amplitudes may be
established.
It is obvious that there are infinitely many ways of computing the relations between
the values of each vibration mode shape. It is usual to do it so to obtain an easy
interpretation and comparison of the several vibration modes.
One of these ways is to normalize the vectors so that the largest value corresponds
to unity. Another way is to assign the same value for a given degree of freedom in
each vibration mode vector.
Either way it is convenient to express the vibration mode shapes in the normalized
form, i.e., in dimensionless terms by dividing all the components by one reference
component. The resulting vector is called the nth mode shape φ
n
. The matrix,   Φ ,
assembling each of the mode shapes in a column is called the mode shape matrix
and may be written as follows:
 
(
(
(
(
(
¸
(
¸
= Φ
NN N N
N
N
φ φ φ
φ φ φ
φ φ φ
...
.. .......... ..........
...
...
2 1
2 22 21
1 12 11
(2.45)
18
Standard Methods for Seismic Analyses
2. 8 Orthogonal i ty Condi ti ons
The free vibration mode shape vectors, φ
n
, have certain special properties called
orthogonality conditions, which are very useful in structural dynamical analysis.
Orthogonal i ty wi th respect to the mass matri x
The dynamical equilibrium equation in the form (2.43) may be written for the modes n
and m as follows.
   
n n n
M K φ ω φ ⋅ ⋅ = ⋅
2
(2.46)
   
m m m
M K φ ω φ ⋅ ⋅ = ⋅
2
(2.47)
Multiplying equation (2.46) by one has:
T
m
φ
   
n
T
m n n
T
m
M K φ φ ω φ φ ⋅ ⋅ ⋅ = ⋅ ⋅
2
(2.48)
Transposing equation (2.47) and noticing that   M and   K are symmetrical, i.e.:
and , one has:    
T
M M =    
T
K K =
    M K
T
m m
T
m
⋅ ⋅ = ⋅ φ ω φ
2
(2.49)
If equation (2.49) is multiplied on the righthand side of each member by
n
φ , the
following expression is achieved:
   
n
T
m m n
T
m
M K φ φ ω φ φ ⋅ ⋅ ⋅ = ⋅ ⋅
2
(2.50)
Subtracting equation (2.50) from equation (2.48) results in:
( )   0
2 2
= ⋅ ⋅ ⋅ −
n
T
m m n
M φ φ ω ω (2.51)
It is evident from the discussion in paragraph 2.7 that if m ≠ n, the corresponding
frequencies will be different, making the following equation (2.52) valid:
  0 = ⋅ ⋅
n
T
m
M φ φ (2.52)
This condition shows that the vibration mode shapes are orthogonal with respect to
the mass matrix.
Orthogonal i ty wi th respect to the sti ffness matri x
Dividing equations (2.48) and (2.50) by and , respectively, one has:
2
n
ω
2
m
ω
   
n
T
m n
T
m
n
M K φ φ φ φ
ω
⋅ ⋅ = ⋅ ⋅ ⋅
2
1
(2.53)
   
n
T
m n
T
m
m
M K φ φ φ φ
ω
⋅ ⋅ = ⋅ ⋅ ⋅
2
1
(2.54)
19
Standard Methods for Seismic Analyses
Subtracting equation (2.54) from (2.53) gives the following condition:
  0
1 1
2 2
= ⋅ ⋅ ⋅


.

\

−
n
T
m
m n
K φ φ
ω ω
(2.55)
Thus for different vibration mode shapes the following orthogonality condition with
respect to the stiffness matrix is valid:
  0 = ⋅ ⋅
n
T
m
K φ φ (2.56)
The results (2.52) and (2.56) lead to:
       
G
T
M M = Φ ⋅ ⋅ Φ (2.57)
       
G
T
K K = Φ ⋅ ⋅ Φ (2.58)
in which the matrices  and 
G
M  
G
K are of diagonal form.
2. 9 Modal Coordi nates
For dynamical analysis of linear systems with any kind of property (damped or
undamped; with or without loading) it is assumed that the displacements are
represented in terms of the free vibration mode shapes,
n
φ . These shapes constitute
N independent displacement patterns, the amplitudes of which may serve as
generalized coordinates to express any form of displacement. This is the same to say
that any displacement vector, { } q , may be written by superimposing suitable
amplitudes, Y, of the N modes of vibration.
{ }
{ }
n
N
n
n
N N
Y q
Y Y Y q
⋅ = ⇔
⇔ ⋅ + + ⋅ + ⋅ =
∑
=1
2 2 1 1
...
φ
φ φ φ
(2.59)
It is evident that the modeshape matrix serves to transform from the generalised
coordinates, Y, to the geometric coordinates, q. These generalized modeamplitude
coordinates are called modal coordinates.
{ }   { } Y q ⋅ Φ = (2.60)
The problem lies now in determining the modal coordinates vector, { , so that it may
be used in equation (2.60) in order to determinate the response of the system in
geometrical coordinates. The procedure of determining the displacement vector,
} Y
{ } q
using (2.60) is called mode superposition method.
It should be noted that the modeshape matrix,   Φ , is composed by N independent
modal vectors and therefore it is nonsingular and may be inverted. This means that
20
Standard Methods for Seismic Analyses
it may always be solved directly for the modal coordinates amplitude, Y, associated
with any given displacement vector, { } q .
{ } Y
T
n
φ
  { } q ⋅ Φ =
−1
(2.61)
2. 10 Equati on of Moti on i n Modal Coordi nates
The equation of motion (2.40) represents a set of N simultaneous differential
equations coupled by the offdiagonal terms in the mass and stiffness matrices. It will
now be shown that, with an appropriate normalizing procedure for the vectors
n
φ and
regarding the orthogonality conditions observed previously, it is possible to transform
the equation of motion into a set of N independent modal coordinate equations.
Solving each of these equations and applying the mode superposition method leads
to the establishment of the dynamical response of the system.
The normalising procedure is called normalization with respect to the mass matrix,
, and may consist in writing the vibration mode shape vector, M
n
φ , so that the
following condition will be valid:
  1 = ⋅ ⋅
n
M φ (2.62)
In order to determine,
n
φ , the reference component by which the nth vibration mode
shape, { } , should be divided is:
n
q
{ }   { } (2.63)
n
T
n
q M q ⋅ ⋅
Finally the normalized vibration mode shape vector
n
φ :
{ }
{ }   { }
n
T
n
n
n
q M q
q
⋅ ⋅
= φ (2.64)
As a consequence of this normalization, using (2.62), one has:
        I M
T
= Φ ⋅ ⋅ Φ (2.65)
with as the N x N identity matrix.   I
Another important result deriving from this type of normalization may also be shown:
1) Multiplying both members of equation (2.46) by , one obtains:
T
n
φ
   
n
T
n n n
T
n
M K φ φ ω φ φ ⋅ ⋅ ⋅ = ⋅ ⋅
2
(2.66)
2) Using the result expressed in (2.65) and remembering (2.58),
 
2
,
2
n n G n n
T
n
K K ω ω φ φ = ⇒ = ⋅ ⋅ (2.67)
21
Standard Methods for Seismic Analyses
Therefore, the diagonal element at line n of the stiffness matrix,  , equals the
square of the nth vibration mode frequency.

G
K
Regarding damping, it will be assumed that, as for the mass and stiffness matrices,
the damping matrix is written in a way that the orthogonality conditions are satisfied:
       
G
T
C C = Φ ⋅ ⋅ Φ (2.68)
It may be shown that, if the mode shape matrix,   Φ , is normalized according to
(2.64), then the matrix is a diagonal matrix with each diagonal element as:  
G
C
nn G
c
,
 
n n n
T
n nn G
C c ξ ω φ φ ⋅ ⋅ = ⋅ ⋅ = 2
,
(2.69)
where
n
ξ represents the nth mode damping ratio. This parameter may be interpreted
as an energy loss mechanism associated with the corresponding vibration mode
6
.
In the following the steps that allow writing equation (2.40) in terms of modal
coordinates and therefore as a set of independent equations are described.
1) Equation of motion in terms of the geometrical coordinates.
  ( ) { }   ( ) { }   ( ) { } ( ) { } t p t q K t q C t q M = ⋅ + ⋅ + ⋅
& & &
(2.40)
2) Multiplication of both members by  
T
Φ and introduction of the neutral element
      I = Φ ⋅ Φ
−1
in the first member.
        ( ) { }         ( ) { }
        ( ) { }   ( ) { } t p t q K
t q C t q M
T T
T T
⋅ Φ = ⋅ Φ ⋅ Φ ⋅ ⋅ Φ +
+ ⋅ Φ ⋅ Φ ⋅ ⋅ Φ + ⋅ Φ ⋅ Φ ⋅ ⋅ Φ
−
− −
1
1 1
& & &
(2.70)
3) Simplification considering the results (2.65), (2.58) and (2.68).
  ( ) { }     ( ) { }     ( ) { }   ( ) { t p t q K t q C t q
T
G G
⋅ Φ = ⋅ Φ ⋅ + ⋅ Φ ⋅ + ⋅ Φ
− − − 1 1 1
& & &
} (2.71)
It is evident now that one may write the previous equation for the modal coordinate,
Y
n
, considering the transformation expressed in (2.61) and simplifying by means of
(2.67) and (2.69), in the following form.
( ) { } t p Y Y Y
T
n n n n n n n
⋅ = ⋅ + ⋅ ⋅ ⋅ + φ ω ξ ω
2
2
& & &
(2.72)
Two comments should be made about this equation:
i. The mode shape matrix,   Φ , does not change with time which implies:
{ }   ( ) { } t q Y
&
&
⋅ Φ =
−1
(2.73)
6
The conditions regarding damping orthogonality are discussed in detail in chapter 133 of reference 1
and section 12.3 of reference 2.
22
Standard Methods for Seismic Analyses
{ }   ( ) { } t q Y
& &
& &
⋅ Φ =
−1
(2.74)
ii. Equation (2.72) is written in terms of modal coordinates, in which the
normalizing procedure has been done with respect to the mass matrix.
Therefore the following equation may be inferred from (2.72) using (2.57) and
(2.65):
( ) { }
n G
T
n
n
2
n n n n n
M
t p
Y Y Y
,
2
⋅
= ⋅ + ⋅ ⋅ ⋅ +
φ
ω ξ ω
& & &
(2.75)
The similarity between the previous expression and equation (2.22), describing the
equation of motion for SDOF systems, is evident. This similarity is the basic principle
for carrying out a dynamical analysis using the mode superposition method assuming
that the system behaves linearly. In fact it is assumed that the motion response for
the mode n (modal coordinate Y
n
) is the same as the motion response computed for
a SDOF system with the properties m, ω and ξ having the same values as the
corresponding ones written in modal coordinates ,
n G
M
, n
ω and
n
ξ . As already
discussed in the chapters referring to SDOF systems it is possible to solve equations
(2.72) or (2.75) for each of the N modes and therefore achieve the modal coordinates
vector . As mentioned before, once the vector { } Y { } Y is determined, application of
the transformation (2.60) leads to the global response of the system in terms of
single degree of freedom equations in geometric coordinates.
However, for common structural systems subjected to extreme dynamical loading, as
in a strong earthquake, it may be rather unrealistic to assume linear behaviour. For
instance, in reinforced concrete structures submitted to dynamic loading, the stiffness
distribution successively changes, not only due to the fact that certain elements are
near yielding but also due to cracking. These are effects very difficult to take into
account with the mode superposition method, since this method assumes that the
structural properties remain constant in time. Therefore no information beyond the
elastic limit is provided such as the inelastic energy dissipation. It is known that the
formation of plastic hinges in a structure designed in a redundant way leads to the
dissipation of energy transmitted by dynamic loading. This has a similar effect as
damping and has a significant contribution to the structural response after yielding.
2. 11 Base Moti on for MDOF Systems
The establishment of the equations of motion for several degrees of freedom MDOF
systems follows the reasoning described above. Again, relative coordinates, q*,
presented in (2.31), are used due to the convenience regarding the effects of base
motion on the system.
23
Standard Methods for Seismic Analyses
Because no dynamic load is applied on any degree of freedom, the set of equations
of motion, in the form of (2.36), will be written as follows:
0 f f f
.... .......... ..........
0 f f f
0 f f f
N k, N c, N i,
k,2 c,2 i,2
k,1 c,1 i,1
= + +
= + +
= + +
(2.76)
As in SDOF systems, only the inertia forces, , are in terms of absolute coordinates.
Reducing (2.76) to relative coordinates and expressing the equation in matrix form
leads to:
i i
f
,
  ( ) { }   ( ) { }   ( ) { }   ( ) { } t q M t * q K t * q C t * q M
s
& & & & &
⋅ − = ⋅ + ⋅ + ⋅
7
(2.77)
The vector { is the support acceleration vector and depends on the particular
support conditions. However, it is reasonable to neglect this fact due to simplification
regarding the common structural dimensions.
( ) t q
s
& &
}
} }
It should also be noted that the support acceleration vector has three components
, and { , corresponding to direction X, Y, and Z. It will be assumed
here that the first two directions are in the surface plane and Z corresponds to
vertical direction.
( ) { } t q
sX
& &
( ) { t q
sY
& &
( ) t q
sZ
& &
Referring to the explanation, given in paragraph 2.6 about the influence coefficients,
m
i,i
, c
i,i
and k
i,i
, composing the matrices   M ,   C and   K , it is evident that if a degree
of freedom, i, is under direction J, only the motion of the degrees of freedom under J
direction will affect the motion of the actual degree of freedom i. Therefore it is
practical to introduce into equation of motion (2.77) a set of vectors , { }
X
1 { }
Y
1 and
. These are written so that nth line corresponds to the nth degree of freedom and
the corresponding value will be unity, if the degree of freedom is in the same
direction as that of the vector, otherwise it is zero.
{ }
Z
1
Introducing the above vectors the equation of motion for MDOF systems submitted to
base motion will have the following form:
  ( ) { }   ( ) { }   ( ) { }
  { } ( ) { } ( ) { } ( ) ( ) t q t q t q M
t q K t q C t q M
sZ Z sY Y sX X
& & & & & &
& & &
⋅ + ⋅ + ⋅ ⋅ − =
= ⋅ + ⋅ + ⋅
1 1 1
(2.78)
It is obvious that this equation is of the same form as (2.40). The procedures
described in the previous paragraph, regarding the equation of motion in modal
coordinates, may then be applied. Considering again the nth modal coordinate, one
has:
7
In the present document whenever support motion is discussed for MDOF systems, the relative
coordinates are used. Therefore the symbol * will be omitted in the following expressions.
24
Standard Methods for Seismic Analyses
  { } ( )   { } ( )   { } ( ) t q M t q M t q M
Y Y Y
sZ Z
T
n sY Y
T
n sX X
T
n
n
2
n n n n n
& & & & & &
& & &
⋅ ⋅ ⋅ − ⋅ ⋅ ⋅ − ⋅ ⋅ ⋅ − =
= ⋅ + ⋅ ⋅ ⋅ +
1 1 1
2
φ φ φ
ω ξ ω
(2.79)
The term , affecting each acceleration value q , is denominated the
modal participation factor of the nth mode for direction J, P
  { }
j
T
n
M 1 ⋅ ⋅ φ
sJ
& &
nJ
. As it may be inferred it
only depends on the vibration mode shape, the mass distribution and the direction of
each degree of freedom. By superposition analysis, regarding the linear behaviour of
the system, is possible to solve the equation separately for each direction, which will
lead, for mode n and direction J, to the following differential equation.
( ) t q P Y Y 2 Y
sJ nJ n
2
n n n n n
& &
& & &
⋅ − = ⋅ + ⋅ ⋅ ⋅ + ω ξ ω (2.80)
It was mentioned before that the support motion from an earthquake is of the form of
an excitation. Therefore the minus sign in (2.80) is of minor interest. Generally the
sign of the response does not have any important significance in an earthquake
analysis. From now on it will be omitted due to simplification.
Equation (2.80) is of the same form as (2.72), which means that it may be solved
analogously as for a SDOF system. Moreover, as the modal participation factor is a
dimensionless parameter and the behaviour of the system is linear, it is possible to
solve the equation of motion in the form (2.80) without using P
nj
(first line in (2.81)).
This parameter may be used again to compute the actual modal coordinate by simply
multiplying it by the solution determined as mentioned above (second line in (2.81)).
( )
n nJ n n
n sJ n
2
n n n n n
Y P Y Y Coordinate Modal Final
Y t q Y Y 2 Y
′ ⋅ = ⇒
′ ⇒ = ′ ⋅ + ′ ⋅ ⋅ ⋅ + ′
,
& &
& & &
ω ξ ω
(2.81)
As before, the equation of motion under direction J for the nth degree of freedom
may be computed applying the transformation (2.60).
( ) ( ) t Y P t q
n iJ
N
i
i n J n
′ ⋅ ⋅ =
∑
=1
, ,
φ (2.82)
It appears from this expression, that the modal participation factor serves also as a
measure of each mode contribution for the response in geometric coordinates. For
instance, consider the response of a degree of freedom under X direction, in a given
MDOF system. It is expected that modes with displacements mainly under X direction
will contribute more to this response, than other modes having their displacements
mainly in other directions.
25
Standard Methods for Seismic Analyses
2. 12 Vi brati on Anal ysi s by the Rayl ei gh Method
The Rayleigh method is widely used as it provides a simple method of evaluating the
natural frequency both for SDOF and MDOF systems.
Basi c concepts
The basic concept in this method is the principle of conservation of energy. This
implies that the energy of a SDOF system, as shown in figure 1, must remain
constant if no damping forces, f
c
, act to absorb the energy when the system is freely
vibrating. The total energy in this case consists of the sum of the kinetic energy of the
mass, T, and the potential energy of the spring, V.
The motion of this system may be assumed harmonic i.e.:
( ) ( ) t sin Z t q ⋅ ⋅ = ω
0
(2.83)
where Z
0
is the amplitude and ω the frequency.
Under these conditions is evident that:
when the systems is in its neutral position, q(t)=0, the force of the spring is 0
and the velocity is maximum, ω ⋅
0
Z . The entire energy of the system is then
given by the kinetic energy of the mass:
(
2
0 max
2
1
ω ⋅ ⋅ ⋅ = Z m T ) (2.84)
when the system is at maximum displacement the velocity of the mass equals
0 which means that the entire energy of the system is the potential energy of
the spring:
2
0 max
2
1
Z k V ⋅ ⋅ = (2.85)
According to the principle of conservation of energy, for the present conditions, the
previous expressions must be equal. Thus the same result is established as in (2.10):
( )
m
k
Z k Z m
= ⇔
⇔ ⋅ ⋅ = ⋅ ⋅ ⋅
ω
ω
0
2
0
2
1
2
1
(2.86)
Approxi mate anal ysi s of a general system; Sel ecti on of the
vi brati on shape
The main advantage of this method is that it provides a simple procedure to
determine a good approximation of the natural frequency of MDOF systems.
Consider a simply supported beam as shown in figure 7.
26
Standard Methods for Seismic Analyses
Figure 7 – Simply supported beam with a
selected deformed shape possible
This beam may be considered as a MDOF system as it has an infinite number of
degrees of freedom. To apply the Rayleigh method one has to assume a deformed
shape for the fundamental mode of vibration so that it may be possible to compute
the maximum potential and kinetic energy.
This may be achieved writing the deformed shape in terms of a shape function, ( ) x ψ ,
representing the ratio of the displacement at any point x to a reference displacement,
Z(t), varying harmonically in time (see figure 7):
( ) ( ) ( ) t Z x t x q ⋅ =ψ , (2.87)
with ( ) ( ) t Z t Z ⋅ ⋅ = ω sin
0
.
The previous assumption of the shape function, ( ) x ψ , effectively reduces the beam to
a SDOF system as the knowledge of a single function allows the evaluation of the
displacement of the entire system.
The flexural strain energy, V, of a prismatic beam, as shown in figure 7, is given by
the following expression, EI being the bending stiffness,
( )
( )
dx
dx
t x q d
x EI V
L
2
0
2
2
,
2
1
∫
(
¸
(
¸
⋅ ⋅ = (2.88)
Introducing equation (2.87) into this expression and letting the reference
displacement, Z(t), take its maximum value one finds the following expression for the
maximum strain energy, V
max
:
( )
( )
dx
dx
x d
x EI Z V
L
2
0
2
2
2
0 max
2
1
∫
(
¸
(
¸
⋅ ⋅ ⋅ =
ψ
(2.89)
The kinetic energy of the beam vibrating as assumed in (2.87) is:
( )
( )
dx
dt
t x dq
x m T
L
2
0
,
2
1
∫ (
¸
(
¸
⋅ ⋅ = (2.90)
where m(x) is the mass per unit length.
Proceeding as above to find the maximum strain energy, one may write the maximum
kinetic energy as follows:
27
Standard Methods for Seismic Analyses
( ) ( ) dx x x m Z T
L
2
2
0
2
0 max
2
1
∫
⋅ ⋅ ⋅ = ψ ω (2.91)
The application of the principle of conservation of energy leads to the following
natural vibration frequency:
( )
( )
( ) ( ) dx x x m
dx
dx
x d
x EI
L
L
2
0
2
0
2
2
∫
∫
⋅
(
¸
(
¸
⋅
=
ψ
ψ
ω (2.92)
The accuracy of the vibration frequency obtained by the Rayleigh method depends
entirely on the shape function assumed, ( ) x ψ . Any shape function satisfying the
geometrical boundary conditions may be selected as it represents a possible
vibration shape. However, any shape other than the natural vibration shape requires
the action of additional external constraints that contribute to stiffen the system and
therefore to increase the corresponding frequency. Consequently from the infinity of
vibration shapes possible in a general system, the true vibration shape yields the
lowest frequency.
A good approximation to the natural frequency / vibration shape may be obtained
considering the static performance of the system.
One common assumption is to identity the inertia forces with the weight of the
masses in the system. The frequency is then evaluated assuming that the vibration
shape, ( ) x ψ , is the deflected shape resulting from the application the weight in the
direction where the principal vibratory motion is expected to take place. Therefore
considering the system in figure 5a) one would assume the weight load being vertical
as this is the direction where the vibration motions are expected to take place. In a
multistorey building the vibration shape is mainly due to horizontal displacements of
each storey and so the inertia forces should be put in the horizontal direction.
In the following the application of this procedure in determining the natural frequency
of a MDOF system with N degrees of freedom is explained.
According to (2.87) the displacements for the degree of freedom n is given by the
expression:
( ) ( ) t sin Z t q
n n
⋅ ⋅ = ω (2.93)
Here Z
n
is the amplitude, which depends on the position of the mass and may be
taken as the displacement at the degree of freedom when the system is acted upon
by the weight load.
The potential energy is given by the sum of the work of each weightload, W
n
. The
maximum potential energy is given by:
28
Standard Methods for Seismic Analyses
∑
=
⋅ ⋅ =
N
i
n n
Z W V
1
max
2
1
(2.94)
The maximum velocity of mass number n, may be easily found using equation (2.93).
One gets q
n n
Z ⋅ = ω
max ,
&
.
Therefore the maximum kinetic energy may be written in the form:
∑
=
⋅ ⋅ ⋅ =
N
i
n
2
n
n
Z
g
W
T
1
2
max
2
1
ω (2.95)
Thus the frequency in a MDOF with N degrees of freedom determined by equating
the maximum values for the strain and kinetic energies, respectively, is:
∑
∑
=
=
⋅
⋅
⋅ =
N
i
n n
N
i
n n
Z W
Z W
g
1
2
1
ω (2.96)
29
Standard Methods for Seismic Analyses
30
Standard Methods for Seismic Analyses
3. Seismic Analysis by Response Spectra
Response spectrum analysis is perhaps the most common method used in design to
evaluate the maximum structural response due to the seismic action. This is a linear
approximate method based on modal analysis and on a response spectrum
definition. According to the analogy between SDOF and MDOF systems, the
maximum modal response of the nth mode, Y , is the same as for a SDOF system
having
max
n
n
ω ω = and
n
ξ ξ = (see equation (2.75))
It should be emphasized that this procedure only leads to the maximum response,
instead of fully describing the response. This saves up a lot of calculation effort with
evident consequences in the time consumed and CPU requirements. The maximum
response is established for each mode by means of the adequate response
spectrum. Therefore the response spectrum analysis is often considered to be the
most attractive method for the seismic design of a given structural system.
3. 1 Response Spectrum Concept
To explain the response spectrum concept, one considers a SDOF system submitted
to an external action that may be either an applied force or a support displacement.
The procedures used to formulate and solve the equation of motion, ( ) t q , and
therefore to achieve the time dependent response of the referred SDOF system,
were already discussed in paragraphs 2.1 to 2.5. For the response spectrum
definition, it is necessary to evaluate the value of the maximum response, which may
be easily determined once its equation of motion, ( ) t q , is fully known.
Figure 8 – Typical representation of response spectrum
If the procedure of determining the maximum response is repeated for a sufficient
range of SDOF systems, with a specified critical damping ratio, ξ , and for different
natural vibration frequencies, ω , submitted to the same external action, it is possible
to define a function and represent it in a diagram similar to the one shown in figure 8.
31
Standard Methods for Seismic Analyses
This diagram is generally known as a response spectrum, ( ) ξ ω, S . Usually it is
represented with the xaxis being the natural vibration frequencies or periods of
vibration
8
of the SDOF and the yaxis being the corresponding maximum response
values. Generally, in the same graph different response spectra, corresponding to
the same action and to different damping ratios usually found in common structures
(2%, 5% and 10%) are shown as in figure 8.
)
Figure 8 represents a typical relative displacement response spectrum, ( ) ξ ω,
d
S , for
values of critical damping ratio, ξ , usually found in common structural systems. The
meaning of the relative displacement, q*, was already discussed in paragraph 2.5. It
is worth to analyse the evolution of the response spectrum function:
1) For low values of frequency, close to zero, one
may see that the maximum value for the relative
displacement tends to a certain value, which is
the support displacement, q
s
. This is easily
explained if one remembers the concept of the
natural vibration frequency, ω , in a SDOF
system, described by expression (2.10). In fact a
SDOF with a low value of ω is very flexible and
behaves as shown in figure 9 when submitted to
a support displacement.
f
i
Figure 9
2) After a certain value of frequency, the relative
displacement tends to zero. In fact high values of
frequency correspond to a very stiff system. The
response motion will then be as shown in figure
10 – the relative displacements, q*, tend to zero.
F
F
i
F
Figure 10
It should be noted that the maximum responses, ( ) ξ ω, S may be presented in every
desired form, i.e. for displacements, ( ) ξ ω, S
d
, velocities, ( ξ ω,
v
S , and accelerations,
( ) ξ ω, S
a
, or even in the form of internal forces or bending moments in a given point of
the SDOF system.
8
The period T, in seconds, is the inverse of the cyclic frequency in Hz (cycles per second)
32
Standard Methods for Seismic Analyses
The available response spectra used for design purpose, in most of the Seismic
Design codes, are defined by means of an accelerogram representing a typical
earthquake in the region of the structure.
N.B.: An accelerogram is a record of the ground accelerations either measured in a
certain place or generated artificially.
3. 2 Response Spectrum Anal ysi s Appl i ed to MDOF Systems
It was concluded in chapter 2.11 that the equation of motion for the nth degree of
freedom under a support excitation in direction J for a given MDOF system may be
written as in (3.1):
( ) ( ) t Y P t q
n iJ
N
i
i n J n
′ ⋅ ⋅ =
∑
=1
, ,
φ (3.1)
As mentioned, the term P
iJ
may be omitted, and so the modal coordinate, ( ) t
n
Y′ , may
be found using to the analogy between equations (2.80) and (2.22) for MDOF and
SDOF systems, respectively.
For direction J, the maximum value for the modal coordinate in terms of
displacements,
max , n
Y′ , may be easily achieved if the displacement response
spectrum, ( ) ξ ω, S
d
, is available. Instead of solving mathematically an expression in
the form of (2.80), Y is established from the response spectrum,
max , n
′ (
n n d
, S ) ξ ω , for the
SDOF system with both the same natural vibration frequency,
n
ω and critical
damping ratio,
n
ξ . The procedure is illustrated in figure 11.
Figure 11
After establishing the maximum value for the modal coordinate, Y , the modal
participation factor is recovered as:
d n
S = ′ &
max ,
( )
n n d,J iJ n,max
, S P Y ξ ω ⋅ =& (3.2)
33
Standard Methods for Seismic Analyses
In the same way one may calculate the maximum response in terms of
accelerations, Y , or velocities, Y , if the corresponding spectra,
max , n
& &
max , n
&
( ) ω
a
S or ( ) ω
v
S
are accessible
9
.
( )
n n J a, iJ max n,
, S P Y ξ ω ⋅ =&
& &
(3.3)
( )
n n v,J iJ n,max
, S P Y ξ ω ⋅ =&
&
(3.4)
We now discuss the problem of establishing a reasonable value for the global
maximum response of the system. The assumption behind the reasoning expressed
in (3.1), i.e. to sum the maximum values of each modal coordinate, Y , certainly
will correspond to an upper limit of the global response with a low probability of
occurrence, since is very unlikely for the maximum modal responses to happen
simultaneously. In fact this is the main disadvantage of the response spectra
analyses: The result provided is a set of extreme values that don’t take place at the
same time and therefore do not correspond to an equilibrium state. Thus this method
can’t provide information on the failure mode of the structure, which is an important
information from the engineering point of view.
n,max
To minimize these disadvantages it is necessary to combine the modal responses.
There are several ways of carrying out this and it is out of the purpose of the present
text to discuss them. Therefore only two methods are presented. It should be
mentioned that there is some controversy about which method leads to better results.
In the design codes, usually the first method to be discussed below is suggested.
However is up to the designer to choose more accurate procedures of combining the
modal response if the SRSS method can’t be applied.
1) SRSS (Square Root of Sum of Squares)
This is one of the most frequently used modal combination methods. According to
this rule the maximum response in terms of a given parameter, G,
(displacements, velocities, accelerations or even internal forces) may be
estimated through the square root of the sum of the m modal response squares,
, contributing to the global response, i.e. ( )
2
n
G
( )
∑
=
≈
m
n
n
G G
1
2
(3.5)
This method usually gives good results if the modal frequencies of the modes
contributing for the global response are sufficiently separated to each other.
Otherwise another method, such as the one following, will be more adequate.
9
Alternatively this may be done by means of the socalled pseudoresponsespectra. These are
determined remembering that each vibration mode will have an expression in the form of (2.42) for the
corresponding equation of motion. Therefore one has ( ) ( ) ω ω ω
d v
S S ⋅ =& and S . ( ) ( ) ω ω ω
d
2
a
S ⋅ =&
34
Standard Methods for Seismic Analyses
2) CQC (Compl ete Quadrati c Combi nati on)
The reason why this method is more effective in evaluating the maximum
response when the modal frequencies are close to each other is due to the fact
that it considers the correlation between modal responses, whereas the SRSS
method considers these to be independent. In fact if two vibration modes have
close frequencies their contribution to the global response is not independent.
Usually this method is used if 1.5
n 1 n
≤
+
ω ω . The correlation between modes i and
n is estimated using the parameter,
in
ρ , given by the following expression:
( )
( ) ( )
2
in in
2
2
2
in
2
3
in in
2
in
1 4 1
1 8
β β ξ β
β β ξ
ρ
+ ⋅ ⋅ ⋅ + −
⋅ + ⋅ ⋅
= (3.6)
The parameter
in
β is
n
i
in
ω
ω
β = .
The global response is achieved applying the following expression.
∑∑
= =
⋅ ⋅ ≈
m
n
m
i
n i in
G G G
1 1
ρ (3.7)
3. 3 Ducti l e Behavi our Consi derati on
As may be understood by the discussion so far, earthquake analysis by response
spectra is based on the assumption that the system behaves linearly. This means
that even for the maximum response situation the internal forces on the different
structural elements of the system are assumed to be proportional to the
displacements achieved.
However this hypothesis is far from reality for structural materials as reinforced
concrete or steel. For instance, a sketch of the stressstrain curve for steel, in figure
12, shows that this material will roughly behave linearly until yielding and thereafter
nonlinearly until failure. The symbols
y
ε and
u
ε stand for yielding and ultimate
strains, respectively.
Figure 12 – Typical stressstrain curve for steel in uniaxial tension or compression
35
Standard Methods for Seismic Analyses
The capacity of the material to absorb deformations in a stabilized way is called
ductility. One way of measuring ductility is the ratio of ultimate deformation to the
yielding deformation. The larger this value the more ability of the material to dissipate
energy after yielding, and therefore the more ductile.
The seismic design criteria consider that a structure submitted to an extreme
earthquake should be prevented from collapse but significant damage is expected.
Therefore this type of action must be included among the design load conditions for
the Ultimate Limit State design. Under these conditions, yielding is expected which
will lead to inelastic response of the structure.
Assuming that the deflections, δ , produced by a given earthquake are essentially the
same whether the structure behaves linearly or yields significantly, one can utilize the
nonlinear behaviour and design structures for less values of stresses, σ, or internal
forces, F. This idea is illustrated in figure 13.
Figure 13
Therefore if the response spectra method is used to design a structural system, the
stresses / internal forces corresponding to the maximum deformations previously
achieved may be reduced to take into account the yielding of the material. This is
done by means of the coefficient, η , called the reduction factor or behaviour
coefficient the physical meaning of which is shown in figure 13.
The determination of this coefficient is also a matter of controversy. Usually, the
value given for the behaviour coefficient is much less than the real one as the elastic
response is reduced using further reduction coefficients (see chapter 6). However, it
is accepted that in order to maximize the nonlinear behaviour of the system and thus
its behaviour coefficient, it is desirable to design it in a redundant way i.e. with a
sufficient number of plastic hinges allowed before collapse.
It should be stated that ductility does not depend only on the material characteristics
but also on the system and the direction of loading. Consider, for instance, the MDOF
system in figure 5 b). The horizontal motion of the mass will induce bending moments
on the column whereas the vertical motion of the mass will lead to a compression /
36
Standard Methods for Seismic Analyses
tension. For the first situation the momentrotation curve will show that the element
has capacity to absorb deformations after yielding and so ductile behaviour may be
assumed. On the other hand, the axial force – axial deformation diagram often show
brittle behaviour and so 1 = η is usually adopted. This is the reason why in most of
the analyses, for vertical seismic action, the reduction factor is taken as unity.
37
Standard Methods for Seismic Analyses
38
Standard Methods for Seismic Analyses
4. Seismic Response by TimeHistory Analysis
TimeHistory analysis is a stepbystep procedure where the loading and the
response history are evaluated at successive time increments, ∆t – steps. During
each step the response is evaluated from the initial conditions existing at the
beginning of the step (displacements and velocities) and the loading history in the
interval. With this method the nonlinear behaviour may be easily considered by
changing the structural properties (e.g. stiffness, k) from one step to the next.
Therefore this method is one of the most effective for the solution of nonlinear
response, among the many methods available. Nevertheless, in the present text, a
linear time history analysis is adopted i.e. the structural properties are assumed to
remain constant during the entire loading history and further it is assumed that the
structure behaves linearly. As a consequence the mode superposition method,
already discussed in chapter 2, may be applied.
4. 1 Response of a SDOF System to General Dynami c
Loadi ng; Duhamel ’ s I ntegral
The equilibrium equation for a given general dynamic loading, p(t), may be
expressed in the same form as (2.22) for a damped SDOF system, i.e.:
( ) ( ) ( )
( )
m
t p
t q t q 2 t q
2
= ⋅ + ⋅ ⋅ ⋅ + ω ω ξ
& & &
(2.22)
It should be noted that both the response, q(t), and the dynamical loading, p(t),
depend on time. The purpose of Duhamel’s integral is to achieve the response at any
time, t, due to load applied at another time τ.
The response to general dynamic loading of a SDOF system subjected to initial
conditions q
0
and q is deduced considering first the corresponding free vibration
response as in equation (2.20).
0
&
( ) ( ) (
(
¸
(
¸
⋅ ⋅
⋅ ⋅ +
+ ⋅ ⋅ =
⋅ ⋅ −
t sin
q q
t cos q e t q
d
d
0 0
d 0
t
ω
ω
ω ξ
ω
ω ξ
&
) (2.20)
If the starting time is different from 0, the above expression may be written in a
general form introducing τ as the time corresponding to the initial conditions:
( )
( )
( ) ( ) ( )
( ) ( )
( ) (
(
¸
(
¸
⋅ ⋅
⋅ ⋅ +
+ ⋅ ⋅ =
⋅ ⋅ −
τ ω
ω
ω ξ τ τ
τ ω τ τ
τ ω ξ
 t sin
q q
 t cos q e  t
d
d
d
 t
&
) q
(4.1)
39
Standard Methods for Seismic Analyses
Now we consider the same SDOF system acted upon by a load p(τ). This load
induces into the system a velocity variation, q
&
∆ , in the interval ∆τ given by the
impulsemomentum relationship:
( ) τ τ d p q m
∫
= ∆ ⋅
&
(4.2)
The second term in this equation represents the area of the plot p(τ) in the time
interval ∆τ. For a differential time interval, dτ, this area is simply p(τ)dτ, which allows
to rewrite equation (4.2) as follows:
( ) ( ) τ τ τ d p q d m = ⋅
&
(4.3)
Using the previous relation and noticing that the response after the termination of the
short duration impulse, ( ) τ τ d p , is a free vibration motion subjected to an initial
velocity, , one may write the differential response, dq(t), as follows, for t>τ: ( ) τ q d
&
( )
( )
( )
( ) (
(
¸
(
¸
− ⋅ ⋅
⋅
⋅
=
⋅ ⋅ −
τ ω
ω
τ τ
τ ω ξ
t sin
m
d p
e t dq
d
d
 t
) (4.4)
The entire loading history may be considered to consist of a succession of such short
impulses, each producing its own differential response according to the expression
above. Because the system is assumed to be linear, the total response may be
established by summing all the differential responses developed during the loading
history. This is the same as saying that the response at time t is given by the integral
of the differential displacements since time t=0 until time t., i.e.:
( ) ( )
( )
( ) ( τ τ ω τ
ω
τ ω ξ
d t sin e p
m
t q
d
t
t
d
− ⋅ ⋅ ⋅
⋅
=
− ⋅ ⋅ −
∫
0
1
) (4.5)
This result is known as Duhamel’s Integral and is one of the most important results in
Structural Dynamics as it may be used to express the response of any damped
SDOF system subjected to any form of dynamical loading, p(τ). There are several
procedures to evaluate this integral and it is out of the purpose of this text to discuss
them here.
10
To take into account initial conditions, the free damped vibration response must be
added to the solution, which leads to the result:
( ) ( ) ( )
( )
( )
( ) ( ) τ τ ω τ
ω
ω
ω
ω ξ
ω
τ ω ξ
ω ξ
d t sin e p
m
t sin
q q
t cos q e t q
d
t
t
d
d
d
0 0
d 0
t
− ⋅ ⋅ ⋅
⋅
+
+
(
¸
(
¸
⋅ ⋅
⋅ ⋅ +
+ ⋅ ⋅ =
− ⋅ ⋅ −
⋅ ⋅ −
∫
0
1
&
(4.6)
10
References 1, chapter 7, and reference 2, section 4, provide useful information about the evaluation
of the Duhamel Integral for SDOF systems.
40
Standard Methods for Seismic Analyses
As one may notice the general response in the form (4.6) for damped SDOF systems
is composed by two terms with the same nature as discussed in paragraph 2.4. The
first term reflects only the influence of the initial conditions and the second term
corresponds to the loading effect on the structural response.
4. 2 Li near Ti me Hi story Anal ysi s for MDOF Systems
It may be inferred from the discussion held in paragraphs 2.6 to 2.11, that the
solution given by the Duhamel Integral may be used to determine the modal
coordinates of a given MDOF system submitted to general dynamic loading. The
mode superposition method is then used to determine the global response of the
system.
The determination of the modal coordinates of a given MDOF systems, Y
n
(t), is
accomplished from equation (2.72) in which the vector ( ) { } t p represents the general
dynamic loading applied in the corresponding degrees of freedom.
( ) ( ) ( ) ( ) { } t p t Y t Y 2 t Y
T
n n
2
n n n n n
⋅ = ⋅ + ⋅ ⋅ ⋅ + φ ω ξ ω
& & &
(4.7)
The modal coordinate Y
n
(t), has the same form as (4.5), assuming that the system
starts from rest, with
n
ξ ξ = and
n
ω ω = , i.e.:
( ) ( ) ( ) ( τ τ ω τ φ
ω
τ ω ξ
ω ξ
d t sin p e
e
t Y
n d
N
i
t
i in
n d,
t
n
n n
n n
− ⋅ ⋅ ⋅ ⋅ ⋅ =
∑
∫
=
⋅ ⋅
⋅ ⋅ −
,
1
0
) (4.8)
Once this procedure is done for all normal coordinates, one applies the expression
(2.60) to obtain the time dependent equation of motion for each degree of freedom in
geometric coordinates. This will lead to the global response of the system at any
desired time t.
If the system is submitted to initial conditions different from zero, then it is obvious
that equation (4.8) would have to be written in the form of (4.6). For this case one
would have to compute the modal initial conditions q
0,n
and as expressed in
(2.61) considering the vectors
n
q
, 0
&
{ }
0
q and { }
0
q
&
.
It should be noticed that, in order to obtain the equation of motion for a given degree
of freedom at a time t in a MDOF system with N degrees of freedom, it is required to
solve the set of N equations as (4.8). To obtain the global response of the system it is
necessary to compute the equation of motion for the N degrees of freedom. This is
done by means of expression (2.60). Therefore one may conclude that to establish
the deformed shape of a structure at a certain time t, NxN equations in the form of
(4.8) must be solved. If one wants to represent the time history of the displacements,
then a set of time intervals must be established taking into account the desired
accuracy of the time history representation. If the time history has m time intervals
then it is obvious that mxNxN equations in the form of (4.8) must be solved. In most
41
Standard Methods for Seismic Analyses
cases less than N modes are considered since, the modes corresponding to high
frequencies have a small contribution for the response of the structure.
Thus, it may be concluded, that the decision about the number of degrees of freedom
and the desired accuracy for the time history representation affect directly the
number of calculations to accomplish and therefore must be chosen carefully taking
into account the time consuming and the CPU requirements to proceed a time history
analysis. However, for some structures or certain types of analysis the number of
degrees of freedom may be very high, which makes the application of this method
impracticable. This is actually one of the main disadvantages of the method.
4. 3 Ti me Hi story Anal ysi s for Earthquakes
As mentioned before, an earthquake action is considered as a base motion
computed on the basis of the support acceleration. Thus all the results in paragraph
2.5 and 2.11 may be applied.
Separating the support acceleration vector, ( ) t q
s
& &
, in its three components, along the
axes X, Y and Z, we have for each degree of freedom a dynamic load given by the
product of the mass, m, and the corresponding acceleration value, . ( ) t q
sJ
& &
As stated in paragraph 2.11, the definition of the modal coordinates Y
n
(t) may be
done for each direction separately, using the mode superposition approach and the
assumed linear behaviour of the system. If expression (2.80) is used, the equation of
motion for the nth mode under direction J is the following:
( ) ( ) ( ) t q P t Y t Y t Y
sJ nJ n n n n n n
& &
& & &
⋅ = ⋅ + ⋅ ⋅ ⋅ +
2
2 ω ξ ω ( ) (4.9)
Remembering the expression of the modal participation factor, P
nJ
, it is obvious that
the second term in (4.9) may be written as:
( )   { } ( ) t q M t q P
sJ J
T
n sJ nJ
& & & &
⋅ ⋅ ⋅ = ⋅ 1 φ (4.10)
The analogy between expression (4.9) and (4.7) is evident expressing the load vector
as: ( ) { t p }
( ) { }   { } ( ) t q M t p
sJ J
& &
& ⋅ ⋅ = 1 (4.11)
A solution is achieved by substituting into equation (4.8) the term ( ) t p { by
which leads to:
}
  { } ( ) t q M
sJ J
& &
⋅ ⋅ 1
( ) ( ) ( ) ( τ τ ω τ
ω
τ ω ξ
ω ξ
d t sin q P e
e
t Y
n d,
t
sJ nJ
n d,
t
n
n n
n n
− ⋅ ⋅ ⋅ ⋅ ⋅ =
∫
⋅ ⋅
⋅ ⋅ −
0
& &
) (4.12)
The problem now consists in solving this expression above for each modal
coordinate. One of the most common techniques is to assume the load subdivided
into a sequence of time intervals, steps, in which the modal coordinates, Y
n
(t), are
42
Standard Methods for Seismic Analyses
calculated. This procedure is called the stepbystep integration method and next we
shall briefly describe one of the many different ways to solve it.
Stepbystep i ntegrati on method wi th l i near vari ati on of the l oad
In order to perform a time history analysis of a given structure, normally, the designer
uses an accelerogram of a certain earthquake considered to be a typical seismic
action. As previously stated an accelerogram may be a record of the ground
accelerations measured in a certain place during the period of an earthquake. A
complete accelerogram contains the record of the acceleration for the three
directions corresponding to the three cartesian axes, X, Y and Z, and therefore
making automatically available to the designer the values of q , and ( ) t
sX
& &
( ) t q
sY
& &
( ) t
sZ
& &
q .
According to the desired accuracy of the time history analysis, the designer decides
the number of time intervals, ∆t, in which each acceleration component should be
divided. The acceleration is assumed to vary linearly within the referred interval
between the initial value, ( ) τ
0 , sJ
q
& &
, and the final value, ( ) t
sJ
q ∆ + τ
& &
, i.e.:
( ) ( )
( ) ( )
τ
τ
τ τ
τ τ ⋅
∆
− ∆ +
+ =
0 ,
0 ,
sJ sJ
sJ sJ
q t q
q q
& & & &
& & & &
(4.13)
Thus, equation (4.12) for the modal coordinate, Y
n
(∆t), becomes:
( ) ( )
( ) ( )
( ) ( τ τ ω τ
τ
τ τ
τ
ω
τ ω ξ
ω ξ
d t sin
q t q
q e P
e
t Y
n d
t
sJ sJ
sJ nJ
n d,
t
n
n n
n n
− ⋅ ⋅


.

\

⋅
∆
− ∆ +
+ ⋅ ⋅ ⋅ = ∆
∫
∆
⋅ ⋅
∆ ⋅ ⋅ −
,
0
0 ,
0 ,
& & & &
& &
) (4.14)
It should be noted that this expression is exact for the first time interval assuming that
the system is at rest until the load is applied. For the next time intervals, ∆t
i
, regarding
the continuity of the response, the initial conditions, Y and Y must be
determined. These parameters are achieved computing the response at the end of
the previous time interval, ∆t
1 , − i n 1 , − i n
&
i1
, in terms of displacements and velocities. Therefore at
the time i the response for the modal coordinate n is of the same form as (4.6),
i.e.:
t ∆ ⋅
( ) ( ) ( )
( )
( ) ( )
( ) ( ) τ τ ω τ
τ
τ τ
τ
ω
ω
ω
ω ξ
ω
τ ω ξ
ω ξ
ω ξ
d t sin
q t q
q e P
e
t sin
Y Y
t cos Y e t i Y
n d
t
sJ sJ
sJ nJ
d,n
t
d,n
d,n
n 1  n,i 1  n,i
d,n 1  n,i
t
n
n n
n n
n
− ⋅ ⋅


.

\

⋅
∆
− ∆ +
+ ⋅ ⋅ ⋅ +
+
(
(
¸
(
¸
⋅ ⋅
⋅ ⋅ +
+ ⋅ ⋅ = ∆ ⋅
∫
∆
⋅ ⋅
∆ ⋅ ⋅ −
∆ ⋅ ⋅ −
,
0
0 ,
0 ,
& & & &
& &
&
(4.15)
Once all the modal coordinates have been determined for the time t i ∆ ⋅ , it is
possible to compute the corresponding global response in terms of geometric
coordinates using superposition. The repetition of this procedure for each time
interval leads to the time history response of the structure.
43
Standard Methods for Seismic Analyses
44
Standard Methods for Seismic Analyses
5. Equivalent Static Method
This method is perhaps the simplest procedure at disposal for a structural engineer
to perform an earthquake analysis and achieve reasonable results. It is prescribed in
any relevant code for earthquake analysis and is widely used especially for buildings
and other common structures meeting certain regularity conditions.
The method is also called The Lateral Forces Method as the effects of an earthquake
are assumed to be the same as the ones resulting from the statical transverse
loadings.
As discussed before, in the Rayleigh method, an inertia loading provides a good
approximation to the natural vibration shape of the structure. If the structural
response is not significantly affected by contributions from higher modes of vibration
it is reasonable to assume that with an appropriate set of inertia forces one may
achieve a good approximation for the response. This is the basic concept of the
Equivalent Static Method.
Each code presents its own procedure to compute and to distribute the static
equivalent forces in order to achieve the earthquake effects on the structure
11
.
Usually an expression is defined to prescribe the minimum lateral seismic force, also
designated the base shear force.
One usual requirement for the structure regarding the application of this method is
that the natural vibration period of the structure should be limited by a maximum
value, which leads to a certain minimum value of frequency/stiffness. This is due to
the fact that often the response is mainly controlled by the first mode of vibration.
Thus, imposing a minimum value of frequency the higher modes contribution may be
neglected.
The structure to be analysed by the equivalent static method should respect certain
criteria regarding its geometrical regularity and stiffness distribution such as
12
:
All lateral load resisting elements (such as columns or walls) should run from
the base to the top without any interruption:
Mass and lateral stiffness should not change abruptly from the base to the top;
Geometrical asymmetries in height or in plan due to setbacks should not
exceed certain values;
11
Regarding the determination and distribution of the static equivalent forces in a given structure, the
chapters 23 and 24 at reference 1 and the section 4.3.3.2 in reference 3 are recommended.
12
A complete set of requirements of this type is presented, for example, in Reference 3 at section
4.3.3.2.1.
45
Standard Methods for Seismic Analyses
46
Standard Methods for Seismic Analyses
6. Case Study
The present chapter presents seismic analyses of a bridge, similar to one designed
for the High Speed Transportation System in Taiwan, using the methods discussed
before.
6. 1 Structural Model of the Bri dge
A sketch of the bridge is shown in figure 14.
Figure 14
The bridge is a threespan bridge with two rail tracks. Each span has a length 40m
and a 13m width. The crosssection is a box girder. The alignment of the main span
axis is straight. The piers are 15.80m and 12.35m tall and both are rigidly connected
to a shear tap element at the top. The shear tap element is a concrete box with 2m
height and of length 5.4m for each side.
Figure 15 – Structural model (See table 1 for detailed information about cross section properties)
Mid Girder Mid Girder Mid Girder End Girder End Girder End Girder End Girder
Pier 1
Pier Tap
Pier 2
Girder
Tracks
Rigid crosssection
Detail
8 m 24 m 16 m 24 m 16 m 24 m 8 m
; 2 m
; 15.8 m
Pier Tap ; 2 m
; 12.35 m
1
2
y
x
2 1 3
47
Standard Methods for Seismic Analyses
The following assumptions are made for the structural model:
i. The three spans are independent and simply supported at the abutments and
shear taps;
ii. The pier supports are assumed to be fixed;
iii. Both abutments allow rotations perpendicular to the bridge plane and restrain
all the others;
iv. The abutments (see figure 15) allow translation in the same direction as the
main span axis;
v. To take into account the torsional effects due to train loads, the tracks are
assumed to be connected to the girder centroid through weightless rigid
members (see Detail in figure 15);
vi. Cracked column section with effective flexural rigidity, (EI)
e
, equal to ½ EI is
used.
N.B.: The support system assumed for the piers and abutments is too much on the
conservative side. In fact it would be more realistic to admit spring systems to
simulate it. However the procedure adopted is considered adequate for the present
purpose.
The global axes X and Y are shown in figure 15. The axis Z is defined applying the
righthand rule. The local coordinate system coincides with the global coordinate
system for horizontal members. For vertical members the local coordinate system is
achieved applying a positive rotation of 90º on the global coordinate system.
Three types of loads are considered:
Self Weight – the weight of the entire structure which is carrying the loads;
Superimposed dead load on the span – 200 kN/m in vertical direction which
includes the weight of the components other than the main structure stated
above;
One train live load – the weight of a train occupying one track. It is computed
as shown in figure 16.
Figure 16 – One train live load
The geometrical parameters of each crosssection shown in Figure 15 are
summarized in Table 1.
48
96.25 kN/m
60 kN/m
Span Length
6.4 m
Standard Methods for Seismic Analyses
Table 1 – Geometrical properties of the crosssections
A (m
2
) I
xy
(m
4
) I
yy
(m
4
) I
zz
(m
4
) e
y
(m)
End Girder 18.7 68.6 109 47.6 0.53
Mid Girder 8.82 31.9 80.7 20.2 0.34
Rigid crosssection 1000 1000 1000 1000 0
Tracks 7.7x10
3
0.1 0.1 0.1 0
Pier Tap 29.2 120 70.9 70.9 0
Pier 11.5 16.6 12.4 9.83 0
Here:
A, Cross section area;
I
xy
, Torsional moment of inertia;
I
yy
, Moment of inertia about local axis y;
I
zz
, Moment of inertia about local axis z;
e
y
, local coordinate y of shear centre with respect to centroid.
Finally, the material assigned for all the sections is concrete of class C25/30.
Exceptions for the Rigid crosssection and Pier crosssections are made considering
assumptions vi. and vii., respectively. Therefore no mass density is considered for
the concrete assigned for the Rigid crosssection and the Young’s Modulus, E, is
reduced to half the standard value for the Pier crosssection.
6. 2 Frequenci es and Vi brati on Mode Shape Determi nati on for
the Bri dge
The first step to accomplish a dynamical analysis is to model the structure as a
MDOF system. This means to define the degrees of freedom of the structure. The
model definition must represent the real behaviour of the system and plays a
fundamental role in the accuracy of the results.
Regarding the geometry of the bridge, the use of uniaxial finiteelements, called
members, for all elements (piers, spans or pier tap’s) less than 4m long is considered
adequate. The drawings A.1 in the Appendix represent the identification of the
members and joints adopted in this study
The establishment of the degrees of freedom is done according to the mass
distribution and the static loads applied:
49
Standard Methods for Seismic Analyses
Half of the mass of each member is considered to be concentrated in the
nearest joint.
The sum of the loads applied on each member is concentrated at the middle
and “transformed” to a mass dividing by the acceleration of gravity, 9.81 m/s
2
.
Thus it is possible to define the mass properties of the structure assuming that the
entire mass is concentrated at the nodes at which the translational displacements are
specified. This procedure leads to a lumpedmass matrix with null offdiagonal terms
and it represents the simplest form of defining the mass properties of a given
structure.
In this study 640 degrees of freedom were computed.
The procedures leading to the definition of the stiffness matrix, [K], may be found in
any publication about finiteelements and it is out of the purpose to expose them
here.
Once the mass and stiffness matrix are computed, each frequency and the
corresponding vibration mode shape of the system may be determined using
equations (2.44) and (2.43). As one may remember there will be as many modes as
degrees of freedom. This means that the above procedure will be repeated
successively as many times as the number of the degrees of freedom to achieve all
the mode frequencies and vibration shapes.
For the present simplified model, the computation of the 639 frequencies and
vibration mode shapes considered does not represent a significant computational
effort, regarding the automatic calculation systems available nowadays. More
accuracy in the results means bigger refinement of the model, which leads to more
degrees of freedom and therefore larger calculation requirements to solve the eigen
value problem.
One of the most common ways to overcome this situation is the mass participating
criterion. Under this criterion, the response determined by considering only a few
modes is a good approximation as long as the mass participating in it exceeds a
certain value. Of course, the larger this value the more accurate the results. Usually
this value should be bigger than 70%
For this study it is decided to use the first 125 modes, ordered by ascending
frequency values. Generally the modes with lower frequencies contribute more for
the global response. Table 2 shows, for each direction, the mass participation in
terms of percentage of all the mass of the system.
Table 2 – Total mass participation factors for the first 125 modes
Transverse direction Longitudinal direction Vertical direction
Mass Participation % 97.8 99.89 99.30
50
Standard Methods for Seismic Analyses
As it may be seen, with only 20% (125/640) of the modes, practically all the mass
participates in the response. According to the mass participation criterion, the
response results determined considering only 20% of the modes are practically the
same as considering all the modes, with an evident decrease of the time consumed
to perform the calculations.
In the Appendix one may find the results for the eigenvalue problem solution. Table
A.2.1 summarizes the frequencies / period and each mass participation factor for
each mode in the three directions. The following figures refer to the deformed shape
of the first three modes, each one with the corresponding displacements in one
predominant direction.
Finally, it is worth to compare the results for the first mode shape and frequency
given by the eigenvalue solution with the solution provided by the application of the
Rayleigh method. As one may see in table A.2.1 the solution of the eigenvalue
problem leads to a frequency of 1.11 Hz for the first mode. The application of the
Rayleigh method assumes that the first mode shape will have displacements mainly
in the longitudinal direction. Therefore the weight load is applied on this direction in
order to compute the natural frequency as in (2.96). The result is 1.53 Hz. As
expected, is an upper value of the real frequency. However this is considered to be a
good approximation if one compares the figures in Appendix A.2 of the first vibration
mode shape given by the eigenvalue problem solution with the one given by the
Rayleigh method.
6. 3 Response Spectrum Anal ysi s of the Bri dge
Once the modal frequencies and the vibration mode shapes are computed, a
response spectrum analysis may be done.
For the present analysis a response spectrum in terms of accelerations vs period, T
is assumed. This spectrum is computed from the NorthSouth component of El
Centro earthquake scaled by a factor of 2 and is shown in table 3.
Table 3 – The EL Centro’s NS component acceleration response
spectrum scaled up by a factor of 2 for critical damping ratio, % 5 = ξ
a (m/s
2
) 6.26 9.58 14.5 13.8 15.4 15.3 18.1 14.7 9.92 10.3 10.0
T (sec) 0 0.01 0.11 0.21 0.31 0.41 0.51 0.61 0.71 0.81 0.91
It should be noticed that the response spectrum used is only considered for periods
up to 0.91 sec. In fact, as shown in table A.2.1, the period of the first mode is 0.90
seconds making it pointless to compute the response analysis with spectrum values
for periods greater than this as all the other modes will have lower periods.
51
Standard Methods for Seismic Analyses
Because the present analysis is merely an example, simplifications are assumed. For
instance the response spectrum above is used regardless the soil nature. It is known
that the soil characteristics influence a great deal the way the seismic waves reach a
structure and affect its dynamical behaviour. A correct analysis would require the
consideration of the response spectrum corresponding to the soil conditions of the
area where the bridge is built. Usually the soils are classified for earthquake analysis
according to their consistence as soft or hard and/or according to the soil being
sandy or argillaceous.
Another factor the designer needs to take into account is the geographical
localization of the structure. In fact, depending on many factors, there are regions
with a seismic intensity higher than others. In most of the Seismic Codes, this fact is
taken into account by scaling up or down the given response spectra by means of
regional coefficients. For the present case, a correct analysis would require the use
of a response spectrum typical for Taiwan instead of El Centro’s NS component.
Still, given the exemplificative nature of this text, it is decided to use the set of four
regional coefficients, Z, in the Code of Taiwan. The maximum value is 0.4 and the
minimum is 0.22. The structure analysed will be in a region for which the regional
coefficient, Z, is 0.34, i.e. the expected earthquake intensity is scaled down to 85% of
the one expected in the most sensitive region.
For each direction, transverse, longitudinal and vertical, response spectrum loads are
created from the response spectrum shown in table 3. It is not usual to use the same
response spectrum to compute the vertical loading as done in this example. In fact
the vertical motions are generally of a lower intensity than horizontal. For the present
analysis, this is taken into account by reducing the vertical action using a coefficient,
α
v
, equal to 2/3.
Since the mode frequencies were very close, it is decided to adopt the CQC modal
combination.
The behaviour coefficient assumed, η, is 2, i.e. internal forces evaluated by means of
linear analysis are reduced to 50%.
It should be noted that since the earthquake action is in the form of an excitation, the
analysis using response spectra provides an envelope of the response, Therefore the
results are presented regardless of the sign. Thus the designer is requested a critical
attitude when analysing the results attained.
Before discussing the results for the present bridge it is worth to make the following
consideration with respect to the modal participation factors of the modes shown in
the figures of Appendix A.2.
52
Standard Methods for Seismic Analyses
Table 4 – Modal Participation Factors for modes 1, 2 and 3.
Transverse direction Longitudinal direction Vertical direction
Mode 1 0.01 201.73 0.94
Mode 2 166.02 0.22 5.46
Mode 3 2.84 5.87 124.2
As referred in the end of paragraph 2.11 the modal participation factor, P
nJ
, is a good
measure of the contribution of the nth mode for the global response in J direction.
This may easily be confirmed by comparing the figures in the Appendix with the
results in table 4. In fact it is observed that for each mode the largest modal
participation factor is achieved precisely for the predominant direction of the
displacements.
6. 4 Resul ts of the Response Spectrum Anal ysi s
The results obtained are processed in a different manner according to the direction of
the loading and its type (displacements or forces in the members). In the following
these results are presented separately.
The results to be presented correspond to the members and joints identified in the
figures in Appendix A.1.
I nternal forces due to earthquake l oadi ng i n hori zontal di recti on
For this type of results, nonlinear behaviour is allowed and therefore the reduction
factor, η, is used as discussed in paragraph 4.3.
As is defined in some modern seismic codes for earthquake analysis, the Design
Specifications elaborated by the Taiwan High Speed Rail Corporation, allow the
designer to reduce the member forces considering to structural type of the system
using a coefficient, α
y
, equal to 1.25. The coefficient α
y
may be defined as the ratio of
the seismic design action used to the seismic design action leading to formation of a
sufficient number of plastic hinges for overall structural instability.
Therefore the reduction coefficient adopted for multiplying the internal member forces
due to earthquake loading in the horizontal direction is given by:
136 , 0
2 25 , 1
34 , 0
=
×
=
⋅η α
y
Z
(6.1)
Tables A.3.1 to A.3.4 in the Appendix A.3 present the reduced internal member
forces for both earthquake loadings in transverse and longitudinal direction.
53
Standard Methods for Seismic Analyses
I nternal forces due to earthquake l oadi ng i n verti cal di recti on
Usually ductility is not taken into account to compute the internal forces obtained
when the earthquake acts under vertical direction. This is due to the fact that the
response is mainly influenced by the vertical vibration modes, which are normally
associated with brittle behaviour. So the reduction coefficient will be of the same form
as in (6.1) assuming η equals to the unity.
181 , 0
25 , 1
34 , 0 3 2
=
×
=
⋅
y
v
Z
α
α
(6.2)
Tables A.3.5 and A.3.6 in Appendix A.3 present the reduced internal member forces
for earthquake loading in vertical direction.
Di spl acements
The displacements were reduced by simply using the same coefficients for Z and α
v
.
Therefore the following reduction coefficients are considered:
Displacements due to earthquake loading in horizontal direction:
Z = 0.34 (6.3)
Displacements due to earthquake loading in the vertical direction:
227 . 0 34 . 0 3 2 = ⋅ = ⋅ α Z
v
(6.4)
Tables A.3.7 to A.3.9 in Appendix A.3 present the reduced displacements.
Combi nati on of Orthogonal Sei smi c Effects
To account for the directional uncertainty of earthquake motions and the
simultaneous occurrences of the corresponding internal forces in three perpendicular
directions, the results achieved are usually combined. For the present case the
Taiwan High Speed Rail Corporation suggests the combination according to (6.5).
The maximum displacement, internal force or moment, S
max
is given by:
¦
)
¦
`
¹
¦
¹
¦
´
¦
⋅ + ⋅ +
⋅ + ⋅ +
⋅ + ⋅ +
=
y x z
z x y
z y x
S S S
S S S
S S S
S
3 , 0 3 , 0
3 , 0 3 , 0
3 , 0 3 , 0
max
max
(6.5)
6. 5 Ti meHi story Response Anal ysi s of the Bri dge
Once the modal frequencies and the vibration mode shapes are computed, a time
history analysis may be performed.
54
Standard Methods for Seismic Analyses
Considering the exemplificative nature of this analysis and, and the simplification
used previously for the response spectra analysis, only the North – South component
of the El Centro’s accelerogram scaled up by a factor of 2 is used to compute the
three earthquake loadings. Each load corresponds to the application of El Centro’s
NS component in one direction of the bridge. Figure 17 represents the acceleration
plot of the NS component of the El Centro earthquake.
Figure 17 – Accelerogram of the NS component of the El Centro earthquake scaled up by a factor of 2
A complete time history analysis of this bridge would require the use of three
components of the acceleration vector. In this case two situations would have to be
considered corresponding to the application of each horizontal acceleration
component for both transverse and longitudinal direction of the bridge.
As in chapter 6.3 the considerations about the coefficients related to the soil nature,
to the regional coefficients and to the vertical direction apply here.
8
6
4
2
0
2
4
6
8
0 5 10 15 20 25 30
t (sec)
a (m/s
2
)
6. 6 Resul ts of the Ti meHi story Response Anal ysi s
The results are computed in the same way as in the response spectrum analysis.
This is due to the fact that both analyses rest on the mode superposition method
based on the assumption that the system behaves linearly.
Therefore, in this analysis the same values for the reduction coefficients as adopted
for the response spectrum analysis are used.
55
Standard Methods for Seismic Analyses
Tables A.4.1 to A.4.9 present the results similar to the ones presented in tables A.3.1
to A.3.9. Each result refers to the maximum value during the whole time history
reduced by applying the reduction coefficients summarized here:
Member forces due to horizontal earthquake loading: 0.136
Member forces due to vertical earthquake loading: 0.181
Displacements due to horizontal earthquake loading: 0.340
Displacements due to vertical earthquake loading: 0.227
Because the same accelerogram is used to define the support acceleration in the
three directions, X, Y and Z, to account for the directional uncertainty of the
earthquake motions and the low probability of simultaneous occurrence of the
maximum response for each direction, the rule presented in (6.5) may be applied
again. It should be noted that the value S is to be inserted regardless of the sign.
As mentioned the timehistory method allows a much more complete analysis
because it provides the time evolution of any kind of result. The graphs shown in the
following provide some examples of time variation of certain results.
24000
18000
12000
6000
0
6000
12000
18000
24000
0 5 10 15 20 25 30 35
t(sec)
M
y
(kNm)
Figure 18 – Time variation of the moment at the base of pier 1 due to earthquake loading in the
transverse direction
56
Standard Methods for Seismic Analyses
Figure 19 – Time variation of the shear force in longitudinal direction at base of pier 2
due to earthquake loading in the longitudinal direction
Figure 20 – Time variation of the transverse displacement at midspan section of the middle span
due to earthquake loading in the transverse direction
8000
6000
4000
2000
0
2000
4000
6000
8000
0 5 10 15 20 25 30 35
V (kN)
t (sec)
32
24
16
8
0
8
16
24
32
0 5 10 15 20 25 30 35
δ
(mm)
t(sec)
57
Standard Methods for Seismic Analyses
6. 7 Equi val ent Stati c Anal ysi s of the Bri dge
As discussed in chapter 5, this method provides good results when applied to
structures meeting certain “regularity” conditions with respect to geometry, stiffness
and mass distribution. Bridges are not usually part of this group of structures as they
are normally rather complex. Moreover, bridges are often important infrastructures in
social and economic terms, which implies a more careful analysis used in the design.
Therefore equivalent static analysis is normally used only in the predesign phase for
this type of structures. However in this paragraph we illustrate the application of this
method by computing the base shear force when the earthquake is in the transverse
direction.
The Design Specifications of the Taiwan High Speed Rail Corporation prescribe that
the bridge shall be designed and constructed to resist a minimum lateral seismic
force, V, given by the expression:
( )
g
W
T S
Z
V
tot
a
y
⋅ ⋅
⋅
=
η α
(6.6)
where
T is the fundamental period in the direction under consideration. Since we use
a simplified method, this parameter may be determined by the Rayleigh
method which gives T=0.65 sec for the longitudinal direction (cf. paragraph
6.2);
S
a
(T) is the acceleration corresponding to the fundamental period determined
by means of a typical response spectrum. In this case, as in the previous
analyses, we use the NS component of the El Centro earthquake scaled up
by a factor of 2. The value for the acceleration is computed from table 3 by
linear interpolation and is equal to 12.6 m/s
2
;
W
tot
is the total weight of the structure accounting for the train loads.
and kN W
tot
894 72 =
Z, α
y
and η have the same meaning as in the previous analyses.
The base shear force for the earthquake acting in the transverse direction, V
z
, is:
kN V
z
734 12
81 . 9
894 72
6 . 12 136 . 0 = × × = (6.7)
As expected, this value is higher than the ones obtained using the previous analyses.
For instance, consider the results from the response spectrum analysis for the shear
force in the longitudinal direction in the support joints (2 and 5) when the earthquake
acts in the same direction (see table A.3.4). The sum of these internal forces equals
10 019 kN. In fact, as discussed in paragraph 2.12, the deformed shape from the
inertia loading is an approximation to the natural vibration shape and therefore
introduces additional stiffness/frequency and consequently higher internal forces in
the structure.
58
Standard Methods for Seismic Analyses
It should be noted that expression (6.7) is to be applied for each direction of the
bridge so that a complete set of internal forces and displacements may be obtained.
This implies the determination of the fundamental period for the three directions.
As in the previous analyses a combination rule such as in (6.5) should be applied to
obtain the maximum design values in terms of displacements and internal forces.
59
Standard Methods for Seismic Analyses
References
[1] R.W. Clough and Joseph Penzien, Dynamics of Structures, McGrawHill,
1975
[2] Mario Paz, Structural Dynamics: Theory and Computation – third edition, Van
Nostrand Reinhold, 1991.
[3] CEN  European Committee for Standardization, Eurocode 8: Design of
structures for earthquake resistance  Part 1, Draft No 6, 2003.
[4] Luís Guerreiro, Revisões de análise modal e análise sísmica por espectros
de resposta, Reprografia DECivil – Instituto Superior Técnico, 1999.
[5] João Azevedo and Jorge Proença, Dinâmica de estruturas, Reprografia
DECivil – Instituto Superior Técnico, 1991.
[6] Design Specifications, Taiwan High Speed Rail Corporation, 2000.
60
Appendix
A.1 Model Identification
Members Identification
Joints Identification
A.2 Eigen – Value Solution
Table A.2.1
Deformed Shapes
i. Mode 1
ii. Mode 2
iii. Mode 3
iv. Mode 1 applying the Rayleigh Method
X Y Z
1 4,83E+01 1,11 0,904 94,747 0,002 0,000
2 2,23E+02 2,37 0,421 0,000 0,072 66,490
3 5,65E+02 3,78 0,264 0,080 37,217 0,019
4 5,70E+02 3,80 0,263 0,514 10,163 0,005
5 6,68E+02 4,11 0,243 0,007 21,406 0,039
6 1,09E+03 5,26 0,190 0,000 0,002 0,087
7 3,42E+03 9,31 0,107 0,000 0,000 9,956
8 4,21E+03 10,32 0,097 0,001 0,022 0,005
9 5,44E+03 11,74 0,085 0,000 0,001 4,589
10 7,25E+03 13,55 0,074 0,876 0,039 0,078
11 7,45E+03 13,73 0,073 0,259 0,632 0,120
12 7,54E+03 13,82 0,072 0,083 0,000 0,132
13 8,02E+03 14,25 0,070 0,369 1,018 0,000
14 1,04E+04 16,26 0,061 0,000 0,002 0,003
15 1,15E+04 17,05 0,059 0,000 0,003 0,080
16 1,29E+04 18,06 0,055 0,061 0,343 0,001
17 1,32E+04 18,30 0,055 0,236 0,647 0,004
18 1,56E+04 19,86 0,050 0,119 1,224 0,041
19 1,63E+04 20,34 0,049 0,011 5,519 0,023
20 1,77E+04 21,19 0,047 0,067 2,082 0,013
21 1,81E+04 21,42 0,047 0,011 0,011 3,106
22 1,84E+04 21,59 0,046 0,106 0,100 0,061
23 1,93E+04 22,10 0,045 0,070 0,244 0,637
24 1,99E+04 22,48 0,044 0,122 1,384 0,492
25 2,01E+04 22,56 0,044 0,016 0,856 1,451
26 2,08E+04 22,94 0,044 0,054 3,189 0,033
27 2,25E+04 23,89 0,042 0,294 0,265 0,002
28 2,75E+04 26,38 0,038 0,516 0,059 0,016
29 2,92E+04 27,19 0,037 0,001 21,406 0,000
30 3,11E+04 28,05 0,036 0,008 0,036 0,136
31 3,30E+04 28,89 0,035 0,001 0,036 0,082
32 3,94E+04 31,57 0,032 0,000 0,016 0,389
33 4,81E+04 34,91 0,029 0,003 0,297 0,110
34 5,01E+04 35,62 0,028 0,053 0,068 0,070
35 5,49E+04 37,31 0,027 0,067 0,632 0,001
36 5,69E+04 37,97 0,026 0,000 0,064 0,616
37 6,48E+04 40,52 0,025 0,029 0,012 0,408
38 6,59E+04 40,85 0,024 0,000 0,032 0,003
39 6,89E+04 41,79 0,024 0,001 0,000 0,667
40 7,66E+04 44,04 0,023 0,012 0,000 0,890
41 7,72E+04 44,22 0,023 0,014 0,647 0,130
42 8,72E+04 47,01 0,021 0,355 0,007 0,183
43 9,01E+04 47,77 0,021 0,027 0,006 0,386
44 9,91E+04 50,11 0,020 0,020 0,027 0,961
45 1,06E+05 51,72 0,019 0,005 0,024 0,022
46 1,09E+05 52,61 0,019 0,001 0,000 0,004
47 1,10E+05 52,77 0,019 0,016 0,244 0,053
48 1,25E+05 56,34 0,018 0,015 0,003 0,001
49 1,29E+05 57,18 0,017 0,000 0,077 0,116
50 1,34E+05 58,30 0,017 0,045 0,095 0,001
51 1,37E+05 58,95 0,017 0,087 0,099 0,061
Mass participation (%)
Mode
Eigenvalue
( d/ 2)
Frequency
(H )
Period
( )
Mass participation (%)
Mode
Eigenvalue
(rad/sec2)
Frequency
(Hz)
Period
(sec)
X Y Z
52 1,53E+05 62,15 0,016 0,003 0,109 0,064
53 1,55E+05 62,69 0,016 0,032 0,059 0,017
54 1,60E+05 63,73 0,016 0,007 0,132 0,000
55 1,74E+05 66,43 0,015 0,035 0,009 0,015
56 1,77E+05 66,99 0,015 0,024 0,027 0,047
57 1,86E+05 68,57 0,015 0,001 0,018 0,003
58 1,90E+05 69,46 0,014 0,026 0,039 0,008
59 1,95E+05 70,34 0,014 0,083 0,068 0,015
60 1,98E+05 70,78 0,014 0,049 0,007 0,051
61 2,12E+05 73,31 0,014 0,001 0,005 0,351
62 2,19E+05 74,56 0,013 0,001 0,000 0,004
63 2,27E+05 75,81 0,013 0,001 0,004 0,000
64 2,28E+05 76,00 0,013 0,004 0,008 0,145
65 2,32E+05 76,70 0,013 0,006 0,000 0,136
66 2,37E+05 77,41 0,013 0,002 0,002 0,747
67 2,47E+05 79,05 0,013 0,026 0,002 0,011
68 2,61E+05 81,36 0,012 0,000 0,006 0,002
69 2,72E+05 82,99 0,012 0,002 0,000 0,169
70 2,90E+05 85,71 0,012 0,037 0,030 0,043
71 2,99E+05 87,06 0,011 0,087 0,000 0,097
72 3,06E+05 88,03 0,011 0,002 0,017 0,216
73 3,10E+05 88,66 0,011 0,002 0,004 0,183
74 3,13E+05 89,03 0,011 0,000 0,022 0,105
75 3,20E+05 90,02 0,011 0,015 0,379 0,143
76 3,26E+05 90,90 0,011 0,007 0,536 0,072
77 3,45E+05 93,47 0,011 0,024 0,109 0,000
78 3,61E+05 95,62 0,010 0,000 0,002 0,004
79 3,66E+05 96,26 0,010 0,000 0,004 0,036
80 3,80E+05 98,10 0,010 0,007 0,044 0,000
81 3,87E+05 99,03 0,010 0,053 0,003 0,000
82 4,20E+05 103,13 0,010 0,000 0,032 0,015
83 4,34E+05 104,79 0,010 0,012 0,039 0,001
84 4,74E+05 109,53 0,009 0,000 0,004 0,000
85 4,87E+05 111,09 0,009 0,000 0,000 0,119
86 4,94E+05 111,83 0,009 0,000 0,193 0,029
87 5,20E+05 114,80 0,009 0,008 0,405 0,174
88 5,37E+05 116,59 0,009 0,017 0,018 0,034
89 5,39E+05 116,85 0,009 0,019 0,008 0,072
90 5,60E+05 119,14 0,008 0,009 0,001 0,179
91 5,68E+05 119,94 0,008 0,000 0,028 0,000
92 5,74E+05 120,55 0,008 0,007 0,027 0,011
93 5,79E+05 121,16 0,008 0,018 0,019 0,000
94 5,92E+05 122,43 0,008 0,001 0,025 0,178
95 5,97E+05 123,02 0,008 0,001 0,030 0,402
96 6,15E+05 124,79 0,008 0,000 0,000 0,002
97 6,27E+05 126,06 0,008 0,000 0,000 0,018
98 6,59E+05 129,24 0,008 0,000 0,001 0,003
99 6,85E+05 131,68 0,008 0,000 0,152 0,000
100 7,07E+05 133,81 0,007 0,000 0,060 0,003
101 7,10E+05 134,13 0,007 0,000 0,536 0,000
102 7,28E+05 135,78 0,007 0,000 0,000 0,041
X Y Z
Mass participation (%)
Mode
Eigenvalue
(rad/sec2)
Frequency
(Hz)
Period
(sec)
Mode
(rad/sec2) (Hz) (sec)
103 7,58E+05 138,57 0,007 0,000 0,003 0,000
104 7,68E+05 139,52 0,007 0,000 0,000 0,002
105 8,12E+05 143,41 0,007 0,000 0,001 0,031
106 8,23E+05 144,41 0,007 0,000 0,000 0,065
107 8,29E+05 144,88 0,007 0,000 0,032 0,021
108 9,08E+05 151,66 0,007 0,000 0,011 0,000
109 9,65E+05 156,33 0,006 0,000 0,003 0,010
110 9,83E+05 157,79 0,006 0,000 0,022 0,306
111 9,97E+05 158,96 0,006 0,000 0,074 0,038
112 1,05E+06 163,00 0,006 0,000 0,000 0,005
113 1,07E+06 164,30 0,006 0,000 0,018 0,000
114 1,08E+06 165,74 0,006 0,000 0,017 0,018
115 1,09E+06 166,30 0,006 0,000 0,046 0,017
116 1,12E+06 168,73 0,006 0,000 0,000 0,002
117 1,15E+06 170,54 0,006 0,000 0,000 0,000
118 1,15E+06 170,61 0,006 0,000 0,002 0,021
119 1,17E+06 171,89 0,006 0,000 0,025 0,000
120 1,22E+06 176,01 0,006 0,000 0,000 0,000
121 1,29E+06 181,10 0,006 0,000 0,000 0,002
122 1,33E+06 183,64 0,005 0,000 0,039 0,036
123 1,42E+06 189,47 0,005 0,000 0,020 0,031
124 1,44E+06 190,78 0,005 0,000 0,002 0,079
125 1,45E+06 191,76 0,005 0,000 0,060 0,438
Deformed Shape – Mode 1
i. Isometric view
ii. XY Plane view
Deformed Shape – Mode 2
i. Isometric view
ii. XZ Plane view
Deformed Shape – Mode 3
i. Isometric view
ii. XY Plane view
Deformed Shape – Mode 1 –
using the Rayleigh method
i. Isometric view
ii. XY Plane view
A.3 Tables of Results for
Response Spectra Analysis
i. Internal Forces in the Girder – Transverse Loading
ii. Internal Forces in the Piers – Transverse Loading
iii. Internal Forces in the Girder – Longitudinal Loading
iv. Internal Forces in the Piers – Longitudinal Loading
v. Internal Forces in the Girder – Vertical Loading
vi. Internal Forces in the Piers – Vertical Loading
vii. Displacements – Transverse Loading
viii. Displacements – Longitudinal Loading
ix. Displacements – Vertical Loading
Span ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
1 0 57 3230 8454 74744 0
14 10 57 3230 8454 61962 227
14 10 57 3230 8454 61962 227
15 17 47 3203 8397 49224 442
15 17 47 3203 8397 49224 442
16 20 40 3081 7978 36822 583
16 20 40 3081 7978 36822 583
17 22 32 2994 7878 24838 672
17 22 32 2994 7878 24838 672
18 27 21 2840 7657 13773 691
18 27 21 2840 7657 13773 691
19 32 21 2632 7369 5594 634
19 32 21 2632 7369 5594 634
20 35 32 2404 7139 9307 527
20 35 32 2404 7139 9307 527
21 39 42 2132 6867 17166 366
21 39 42 2132 6867 17166 366
22 42 56 1787 6742 24033 142
22 42 56 1787 6742 24033 142
23 44 63 1513 6469 26992 0
24 47 96 2052 3413 31121 0
25 47 96 2052 3413 35119 194
25 47 96 2052 3413 35119 194
26 51 87 1529 3115 40991 535
26 51 87 1529 3115 40991 535
27 55 68 914 2520 44393 790
27 55 68 914 2520 44393 790
28 56 47 419 2027 45763 963
28 56 47 419 2027 45763 963
29 58 25 284 1524 44940 1039
29 58 25 284 1524 44940 1039
30 57 20 848 850 41618 982
30 57 20 848 850 41618 982
31 55 51 1453 762 35850 801
31 55 51 1453 762 35850 801
32 52 70 1965 1061 28047 536
32 52 70 1965 1061 28047 536
33 49 91 2552 1470 17987 191
33 49 91 2552 1470 17987 191
34 46 98 2872 1713 12311 0
35 44 65 715 7657 9704 0
36 44 65 715 7657 8884 132
36 44 65 715 7657 8884 132
37 41 57 1023 7945 6284 374
37 41 57 1023 7945 6284 374
38 38 43 1292 8044 5421 542
38 38 43 1292 8044 5421 542
39 35 31 1521 8313 9275 655
39 35 31 1521 8313 9275 655
40 31 17 1719 8539 15527 705
40 31 17 1719 8539 15527 705
41 27 16 1908 8809 22930 680
41 27 16 1908 8809 22930 680
42 23 30 2053 9012 31072 582
42 23 30 2053 9012 31072 582
43 20 41 2137 9103 39659 431
43 20 41 2137 9103 39659 431
44 15 51 2243 9448 48680 236
44 15 51 2243 9448 48680 236
8 0 59 2265 9476 57721 0
25
26
Table A.3.1  Member Forces in the Girder  Earthquake
loading in transverse direction (kNm)
30
27
28
29
19
20
3
17
18
23
24
21
22
8
9
10
2
11
12
13
14
15
16
Forces Moments
1
4
5
6
1
2
3
7
Pier ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
2 172 19 1737 2876 22935 77
9 172 19 1737 2876 16897 20
9 172 19 1737 2876 16897 20
10 172 15 1717 2874 11125 38
10 172 15 1717 2874 11125 38
11 171 14 1658 2871 5630 76
11 171 14 1658 2871 5630 76
3 169 15 1547 2865 1583 118
3 169 15 1547 2865 1583 118
4 167 15 1338 2848 3085 142
5 174 19 3242 4065 33124 65
12 174 19 3242 4065 21957 11
12 174 19 3242 4065 21957 11
13 174 16 3218 4062 11058 53
13 174 16 3218 4062 11058 53
6 173 16 3143 4057 1276 103
6 173 16 3143 4057 1276 103
7 171 16 2967 4039 6077 133
Table A.3.2  Member forces in the Piers  Earthquake
loading in transverse direction (kNm)
2
35
36
37
39
Forces Moments
32
1
31
33
34
38
Span ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
1 0 249 24 100 145 0
14 254 249 24 100 117 997
14 254 249 24 100 117 997
15 669 217 23 93 89 1780
15 669 217 23 93 89 1780
16 1011 171 21 75 88 2334
16 1011 171 21 75 88 2334
17 1283 132 21 58 93 2695
17 1283 132 21 58 93 2695
18 1582 99 20 32 58 2751
18 1582 99 20 32 58 2751
19 1881 92 20 28 55 2546
19 1881 92 20 28 55 2546
20 2152 114 20 37 68 2157
20 2152 114 20 37 68 2157
21 2422 156 20 49 82 1533
21 2422 156 20 49 82 1533
22 2760 215 20 64 97 643
22 2760 215 20 64 97 643
23 2994 254 20 70 105 0
24 408 162 21 42 126 0
25 408 162 21 42 111 329
25 408 162 21 42 111 329
26 754 92 21 38 99 678
26 754 92 21 38 99 678
27 1090 73 22 34 84 663
27 1090 73 22 34 84 663
28 1359 112 22 42 77 498
28 1359 112 22 42 77 498
29 1628 139 22 49 83 567
29 1628 139 22 49 83 567
30 1924 126 22 43 63 809
30 1924 126 22 43 63 809
31 2218 80 22 30 73 929
31 2218 80 22 30 73 929
32 2487 49 22 35 97 849
32 2487 49 22 35 97 849
33 2823 97 21 53 112 432
33 2823 97 21 53 112 432
34 3055 143 21 60 117 0
35 2972 328 20 94 103 0
36 2972 328 20 94 103 665
36 2972 328 20 94 103 665
37 2625 263 20 86 109 1816
37 2625 263 20 86 109 1816
38 2290 192 20 63 101 2582
38 2290 192 20 63 101 2582
39 2022 135 20 42 96 3074
39 2022 135 20 42 96 3074
40 1753 96 21 29 98 3266
40 1753 96 21 29 98 3266
41 1457 105 21 40 63 3175
41 1457 105 21 40 63 3175
42 1160 153 21 68 49 2766
42 1160 153 21 68 49 2766
43 889 205 21 84 57 2057
43 889 205 21 84 57 2057
44 551 261 23 100 87 1115
44 551 261 23 100 87 1115
8 0 297 23 110 133 0
18
25
3
21
22
23
24
29
30
27
28
26
2
11
12
13
14
19
20
17
15
16
3
1
1
4
7
8
9
10
Table A.3.3  Member Forces in the Girder 
Earthquake loading in longitudinal direction (kNm)
2
5
6
Forces Moments
Pier ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
2 278 3319 9 2 36 51776
9 278 3319 9 2 12 40166
9 278 3319 9 2 12 40166
10 276 3309 7 2 22 28925
10 276 3309 7 2 22 28925
11 273 3274 5 2 37 17804
11 273 3274 5 2 37 17804
3 269 3206 5 2 48 6614
3 269 3206 5 2 48 6614
4 263 3069 6 2 54 856
5 297 6700 12 3 47 82745
12 297 6700 12 3 16 59634
12 297 6700 12 3 16 59634
13 296 6684 11 3 40 36911
13 296 6684 11 3 40 36911
6 293 6633 9 3 69 13710
6 293 6633 9 3 69 13710
7 287 6506 8 3 83 992
2
35
36
37
39
Table A.3.4  Member forces in the Piers  Earthquake
loading in longitudinal direction (kNm)
Forces Moments
1
31
32
33
34
38
Span ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
1 0 2598 159 945 3583 0
14 31 2598 159 945 2964 10394
14 31 2598 159 945 2964 10394
15 79 2240 157 877 2335 19428
15 79 2240 157 877 2335 19428
16 115 1671 151 732 1723 25828
16 115 1671 151 732 1723 25828
17 142 1046 145 591 1142 29903
17 142 1046 145 591 1142 29903
18 171 297 138 373 604 30744
18 171 297 138 373 604 30744
19 193 669 128 400 232 28168
19 193 669 128 400 232 28168
20 199 1373 116 515 476 22732
20 199 1373 116 515 476 22732
21 196 1960 100 628 848 14952
21 196 1960 100 628 848 14952
22 182 2452 80 725 1145 5341
22 182 2452 80 725 1145 5341
23 152 2592 67 735 1264 0
24 476 2169 119 601 1458 0
25 476 2169 119 601 1667 4392
25 476 2169 119 601 1667 4392
26 490 1973 87 589 1980 12353
26 490 1973 87 589 1980 12353
27 492 1538 55 495 2167 18336
27 492 1538 55 495 2167 18336
28 497 1020 30 368 2235 22364
28 497 1020 30 368 2235 22364
29 501 401 29 211 2176 23880
29 501 401 29 211 2176 23880
30 505 384 60 149 1974 22472
30 505 384 60 149 1974 22472
31 506 1052 95 406 1644 18305
31 506 1052 95 406 1644 18305
32 507 1556 123 527 1253 12140
32 507 1556 123 527 1253 12140
33 512 1985 151 622 887 4351
33 512 1985 151 622 887 4351
34 508 2111 168 639 816 0
35 148 2648 46 787 821 0
36 148 2648 46 787 743 5362
36 148 2648 46 787 743 5362
37 162 2446 55 778 560 15222
37 162 2446 55 778 560 15222
38 168 1962 68 679 415 22858
38 168 1962 68 679 415 22858
39 171 1377 77 564 496 28298
39 171 1377 77 564 496 28298
40 165 668 81 450 750 30866
40 165 668 81 450 750 30866
41 146 281 85 414 1060 30035
41 146 281 85 414 1060 30035
42 118 1046 91 610 1400 25898
42 118 1046 91 610 1400 25898
43 92 1676 99 743 1767 19249
43 92 1676 99 743 1767 19249
44 65 2250 109 873 2169 10489
44 65 2250 109 873 2169 10489
8 0 2610 114 932 2590 0
Table A.3.5  Member Forces in the Girder  Earthquake
loading in vertical direction (kNm)
6
Forces Moments
1
5
1
2
9
10
3
4
18
19
7
8
30
26
2
11
12
13
14
15
16
17
27
28
20
3
21
22
23
24
25
29
Pier ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
2 4507 424 156 133 1333 1997
9 4507 424 156 133 842 554
9 4507 424 156 133 842 554
10 4495 414 153 133 492 935
10 4495 414 153 133 492 935
11 4472 398 147 132 548 2261
11 4472 398 147 132 548 2261
3 4439 385 138 132 910 3584
3 4439 385 138 132 910 3584
4 4388 382 124 131 1133 4332
5 4446 522 224 192 1676 1938
12 4446 522 224 192 993 380
12 4446 522 224 192 993 380
13 4434 511 221 192 591 1672
13 4434 511 221 192 591 1672
6 4412 493 215 192 917 3369
6 4412 493 215 192 917 3369
7 4373 484 205 191 1260 4322
2
35
36
37
39
Table A.3.6  Member forces in the Piers  Earthquake
loading in vertical direction (kNm)
Forces Moments
1
31
32
33
34
38
X Y Z X Y Z
1 0,0 0,0 0,0 0,0 0,0 0,0
8 0,0 0,0 0,0 0,0 0,0 0,0
18 0,0 0,4 9,2 0,1 0,0 0,0
4 0,0 0,0 23,8 0,1 0,0 0,0
29 0,0 0,6 30,4 0,1 0,0 0,0
7 0,0 0,0 21,2 0,1 0,0 0,0
40 0,0 0,4 7,5 0,1 0,0 0,0
Translations (mm) Rotations (rad)
Table A.3.7  Displacements  Earthquake
loading in transverse direction
Joint
X Y Z X Y Z
1 73,5 0,0 0,0 0,0 0,0 0,0
8 73,0 0,0 0,0 0,0 0,0 0,0
18 73,3 7,8 0,0 0,0 0,0 0,0
4 73,0 0,1 0,0 0,0 0,0 0,4
29 72,9 1,8 0,0 0,0 0,0 0,0
7 72,5 0,0 0,0 0,0 0,0 0,5
40 72,8 9,8 0,0 0,0 0,0 0,0
Translations (mm) Rotations (rad)
Table A.3.8  Displacements  Earthquake
loading in longitudinal direction
Joint
X Y Z X Y Z
1 0,2 0,0 0,0 0,0 0,0 0,0
8 0,2 0,0 0,0 0,0 0,0 0,0
18 0,2 6,0 0,1 0,0 0,0 0,0
4 0,2 0,3 0,4 0,0 0,0 0,2
29 0,2 4,3 0,5 0,0 0,0 0,0
7 0,2 0,2 0,3 0,0 0,0 0,2
40 0,2 6,0 0,1 0,0 0,0 0,0
Translations (mm) Rotations (rad)
Table A.3.9  Displacements  Earthquake
loading in vertical direction
Joint
A.4 Tables of Results for
TimeHistory Analysis
i. Internal Forces in the Girder – Transverse Loading
ii. Internal Forces in the Piers – Transverse Loading
iii. Internal Forces in the Girder – Longitudinal Loading
iv. Internal Forces in the Piers – Longitudinal Loading
v. Internal Forces in the Girder – Vertical Loading
vi. Internal Forces in the Piers – Vertical Loading
vii. Displacements – Transverse Loading
viii. Displacements – Longitudinal Loading
ix. Displacements – Vertical Loading
Span ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
1 0 66 3349 8163 74498 0
14 8 66 3349 8163 61289 264
14 8 66 3349 8163 61289 264
15 8 57 3278 8074 48635 541
15 8 57 3278 8074 48635 541
16 7 47 3095 7660 36368 718
16 7 47 3095 7660 36368 718
17 10 34 2971 7557 24420 819
17 10 34 2971 7557 24420 819
18 16 18 2791 7399 14410 812
18 16 18 2791 7399 14410 812
19 22 17 2578 7208 6205 739
19 22 17 2578 7208 6205 739
20 26 33 2357 7058 9452 608
20 26 33 2357 7058 9452 608
21 29 49 2089 6862 16767 410
21 29 49 2089 6862 16767 410
22 31 66 1731 6812 23320 149
22 31 66 1731 6812 23320 149
23 36 73 1547 6591 26305 0
24 40 121 2060 3388 30514 0
25 40 121 2060 3388 34527 244
25 40 121 2060 3388 34527 244
26 44 108 1591 3068 40359 672
26 44 108 1591 3068 40359 672
27 51 85 985 2403 43701 1006
27 51 85 985 2403 43701 1006
28 55 57 449 1973 45028 1237
28 55 57 449 1973 45028 1237
29 59 24 311 1580 44222 1335
29 59 24 311 1580 44222 1335
30 58 20 812 922 41025 1259
30 58 20 812 922 41025 1259
31 53 61 1425 700 35478 1019
31 53 61 1425 700 35478 1019
32 48 87 1940 1252 27925 673
32 48 87 1940 1252 27925 673
33 42 111 2498 1836 18067 242
33 42 111 2498 1836 18067 242
34 40 120 2791 2135 12778 0
35 36 79 865 7638 10553 0
36 36 79 865 7638 9764 160
36 36 79 865 7638 9764 160
37 34 70 1081 7900 6307 440
37 34 70 1081 7900 6307 440
38 33 53 1252 7974 6207 646
38 33 53 1252 7974 6207 646
39 29 36 1506 8223 11357 783
39 29 36 1506 8223 11357 783
40 25 17 1766 8430 16383 848
40 25 17 1766 8430 16383 848
41 19 14 2006 8693 22300 826
41 19 14 2006 8693 22300 826
42 12 36 2188 8899 30906 704
42 12 36 2188 8899 30906 704
43 8 51 2314 8991 39893 514
43 8 51 2314 8991 39893 514
44 5 63 2484 9351 50096 279
44 5 63 2484 9351 50096 279
8 3 71 2551 9425 60519 0
Table A.4.1  Member Forces in the Girder  Earthquake loading in transverse
direction (kNm)
5
6
Forces Moments
2
15
16
3
1
1
4
7
8
9
10
28
26
2
11
12
13
14
19
20
17
18
25
3
21
22
23
24
29
30
27
Pier ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
2 198 15 1814 2852 23193 71
9 198 15 1814 2852 16845 18
9 198 15 1814 2852 16845 18
10 198 14 1775 2851 10811 32
10 198 14 1775 2851 10811 32
11 197 13 1694 2848 5215 75
11 197 13 1694 2848 5215 75
3 196 12 1561 2843 1058 118
3 196 12 1561 2843 1058 118
4 194 12 1339 2830 3089 141
5 209 16 3325 3999 33436 61
12 209 16 3325 3999 21964 9
12 209 16 3325 3999 21964 9
13 209 16 3283 3997 10800 48
13 209 16 3283 3997 10800 48
6 208 15 3191 3993 1011 101
6 208 15 3191 3993 1011 101
7 208 15 3001 3980 6371 131
Table A.4.2  Member forces in the Piers  Earthquake loading
in transverse direction (kNm)
32
33
34
Forces Moments
38
2
35
36
37
39
1
31
Span ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
1 0 209 22 77 115 0
14 253 209 22 77 109 835
14 253 209 22 77 109 835
15 667 175 21 72 88 1545
15 667 175 21 72 88 1545
16 1009 141 21 58 82 2001
16 1009 141 21 58 82 2001
17 1281 122 21 44 84 2342
17 1281 122 21 44 84 2342
18 1579 89 20 27 45 2500
18 1579 89 20 27 45 2500
19 1877 80 20 17 50 2381
19 1877 80 20 17 50 2381
20 2147 117 20 28 53 1996
20 2147 117 20 28 53 1996
21 2418 159 19 47 68 1363
21 2418 159 19 47 68 1363
22 2756 218 19 62 85 608
22 2756 218 19 62 85 608
23 2991 239 19 66 90 0
24 405 143 20 25 105 0
25 405 143 20 25 87 289
25 405 143 20 25 87 289
26 753 73 20 22 93 533
26 753 73 20 22 93 533
27 1089 75 20 23 82 490
27 1089 75 20 23 82 490
28 1359 106 20 31 73 488
28 1359 106 20 31 73 488
29 1628 119 20 35 76 620
29 1628 119 20 35 76 620
30 1924 110 20 33 54 830
30 1924 110 20 33 54 830
31 2218 75 20 25 58 892
31 2218 75 20 25 58 892
32 2487 42 20 32 77 814
32 2487 42 20 32 77 814
33 2823 82 20 47 94 424
33 2823 82 20 47 94 424
34 3056 131 20 56 99 0
35 2970 294 19 84 96 0
36 2970 294 19 84 95 596
36 2970 294 19 84 95 596
37 2623 247 19 78 105 1600
37 2623 247 19 78 105 1600
38 2288 195 20 59 99 2308
38 2288 195 20 59 99 2308
39 2019 136 20 36 90 2842
39 2019 136 20 36 90 2842
40 1751 87 21 20 92 3083
40 1751 87 21 20 92 3083
41 1455 88 21 28 48 2988
41 1455 88 21 28 48 2988
42 1158 134 21 59 39 2570
42 1158 134 21 59 39 2570
43 888 170 22 71 51 1905
43 888 170 22 71 51 1905
44 550 224 23 81 79 1036
44 550 224 23 81 79 1036
8 137 259 23 85 115 0
Table A.4.3  Member Forces in the Girder  Earthquake
loading in longitudinal direction (kNm)
5
6
Forces Moments
2
15
16
3
1
1
4
7
8
9
10
28
26
2
11
12
13
14
19
20
17
18
25
3
21
22
23
24
29
30
27
Pier ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
2 238 3306 6 2 29 51731
9 238 3306 6 2 8 40159
9 238 3306 6 2 8 40159
10 237 3301 6 2 15 28934
10 237 3301 6 2 15 28934
11 236 3271 5 2 28 17812
11 236 3271 5 2 28 17812
3 234 3205 4 2 39 6594
3 234 3205 4 2 39 6594
4 232 3069 4 2 44 726
5 232 6692 7 3 25 82717
12 232 6692 7 3 17 59629
12 232 6692 7 3 17 59629
13 231 6682 6 3 29 36911
13 231 6682 6 3 29 36911
6 231 6633 5 3 44 13696
6 231 6633 5 3 44 13696
7 229 6507 4 3 52 992
Table A.4.4  Member forces in the Piers  Earthquake
loading in longitudinal direction (kNm)
32
33
34
Forces Moments
38
2
35
36
37
39
1
31
Span ID Element Joint Axial Shear Y Shear Z Torsion Bending Y Bending Z
1 0 2638 165 905 3951 0
14 15 2638 165 905 3293 10551
14 15 2638 165 905 3293 10551
15 48 2301 163 841 2598 19825
15 48 2301 163 841 2598 19825
16 47 1730 161 714 1911 26486
16 47 1730 161 714 1911 26486
17 55 1076 155 592 1260 30755
17 55 1076 155 592 1260 30755
18 76 255 146 369 630 31693
18 76 255 146 369 630 31693
19 89 651 134 440 229 29094
19 89 651 134 440 229 29094
20 94 1398 121 615 532 23535
20 94 1398 121 615 532 23535
21 93 2016 104 764 943 15524
21 93 2016 104 764 943 15524
22 104 2548 83 891 1253 5539
22 104 2548 83 891 1253 5539
23 80 2708 68 909 1369 0
24 397 2188 162 618 1562 0
25 397 2188 162 618 1774 4430
25 397 2188 162 618 1774 4430
26 415 1997 118 586 2260 12527
26 415 1997 118 586 2260 12527
27 415 1539 71 473 2581 18538
27 415 1539 71 473 2581 18538
28 421 995 28 355 2715 22489
28 421 995 28 355 2715 22489
29 424 380 31 211 2659 23841
29 424 380 31 211 2659 23841
30 424 397 65 168 2389 22230
30 424 397 65 168 2389 22230
31 418 1070 113 403 1900 17913
31 418 1070 113 403 1900 17913
32 414 1547 154 496 1263 11742
32 414 1547 154 496 1263 11742
33 419 1926 196 557 928 4180
33 419 1926 196 557 928 4180
34 409 2018 221 553 884 0
35 73 2735 44 961 855 0
36 73 2735 44 961 777 5538
36 73 2735 44 961 777 5538
37 91 2543 55 954 574 15772
37 91 2543 55 954 574 15772
38 71 2039 65 840 402 23732
38 71 2039 65 840 402 23732
39 67 1427 72 696 573 29391
39 67 1427 72 696 573 29391
40 60 672 78 514 824 32048
40 60 672 78 514 824 32048
41 54 249 83 412 1120 31133
41 54 249 83 412 1120 31133
42 43 1088 92 614 1457 26757
42 43 1088 92 614 1457 26757
43 33 1750 98 735 1833 19786
43 33 1750 98 735 1833 19786
44 39 2327 100 853 2259 10713
44 39 2327 100 853 2259 10713
8 0 2664 99 911 2695 0
Table A.4.5  Member Forces in the Girder 
Earthquake loading in vertical direction (kNm)
2
5
6
Forces Moments
15
16
3
1
1
4
7
8
9
10
28
26
2
11
12
13
14
19
20
17
18
25
3
21
22
23
24
29
30
27
Pier ID Element Joint Axial Shear Y Shear Z Torsion
Bending
Y
Bending
Z
2 4809 352 201 153 1696 1799
9 4809 352 201 153 1011 566
9 4809 352 201 153 1011 566
10 4811 349 199 152 504 811
10 4811 349 199 152 504 811
11 4805 340 191 152 564 1909
11 4805 340 191 152 564 1909
3 4784 329 181 151 998 2999
3 4784 329 181 151 998 2999
4 4747 323 164 150 1314 3610
5 4711 422 295 224 2084 1654
12 4711 422 295 224 1066 375
12 4711 422 295 224 1066 375
13 4623 419 293 224 653 1419
13 4623 419 293 224 653 1419
6 4578 411 287 223 942 2667
6 4578 411 287 223 942 2667
7 4567 405 275 221 1487 3475
Table A.4.6  Member forces in the Piers  Earthquake
loading in vertical direction (kNm)
32
33
34
Forces Moments
38
2
35
36
37
39
1
31
X Y Z X Y Z
1 0,0 0,0 0,0 0,0 0,0 0,0
8 0,0 0,0 0,0 0,0 0,0 0,0
18 0,0 0,5 9,1 0,0 0,0 0,0
4 0,0 0,0 23,6 0,1 0,0 0,0
29 0,0 0,7 30,2 0,1 0,0 0,0
7 0,0 0,0 21,3 0,1 0,0 0,0
40 0,0 0,5 7,6 0,0 0,0 0,0
Translations (mm) Rotations (rad)
Joint
Table A.4.7  Displacements  Earthquake
loading in transverse direction
X Y Z X Y Z
1 73,5 0,0 0,0 0,0 0,0 0,0
8 73,0 0,0 0,0 0,0 0,0 0,0
18 73,3 0,4 0,0 0,0 0,0 0,0
4 73,0 0,0 0,0 0,0 0,0 0,4
29 72,9 0,6 0,0 0,0 0,0 0,0
7 72,5 0,0 0,0 0,0 0,0 0,5
40 72,8 0,9 0,0 0,0 0,0 0,0
Translations (mm) Rotations (rad)
Joint
Table A.4.8  Displacements  Earthquake
loading in longitudinal direction
X Y Z X Y Z
0 0,2 0,0 0,0 0,0 0,0 0,0
8 0,2 0,0 0,0 0,0 0,0 0,0
18 0,2 9,3 0,2 0,0 0,0 0,0
4 0,2 0,5 0,6 0,0 0,0 0,0
29 0,2 6,5 0,8 0,0 0,0 0,0
7 0,2 0,4 0,5 0,0 0,0 0,0
40 0,2 9,3 0,2 0,0 0,0 0,0
Translations (mm) Rotations (rad)
Joint
Table A.4.9  Displacements  Earthquake
loading in vertical direction
João Luís Domingues Costa
Standard Methods for Seismic Analyses
Report BYG·DTU R064 2003
ISSN 16012917 ISBN 8778771269
Contents
ACKNOWLEDGEMENTS ...........................................................................................3 1. INTRODUCTION.....................................................................................................5 2. STRUCTURAL DYNAMICS FUNDAMENTALS .....................................................7 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 FORMULATION OF EQUATION OF MOTION FOR SDOF SYSTEMS ...........................7 UNDAMPED FREE VIBRATIONS ...........................................................................8 DAMPED FREE VIBRATIONS ................................................................................9 RESONANT RESPONSE....................................................................................11 BASE MOTION FOR SDOF SYSTEMS................................................................14 FORMULATION OF THE EQUATION OF MOTION FOR MDOF SYSTEMS...................15 FREQUENCY AND VIBRATION MODE SHAPE ANALYSIS .......................................17 ORTHOGONALITY CONDITIONS .........................................................................19 Orthogonality with respect to the mass matrix .....................................19 Orthogonality with respect to the stiffness matrix.................................19 MODAL COORDINATES ....................................................................................20 EQUATION OF MOTION IN MODAL COORDINATES ...............................................21 BASE MOTION FOR MDOF SYSTEMS ...............................................................23 VIBRATION ANALYSIS BY THE RAYLEIGH METHOD ..............................................26 Basic concepts.....................................................................................26 Approximate analysis of a general system; Selection of the vibration shape ...................................................................................................26
3. SEISMIC ANALYSIS BY RESPONSE SPECTRA................................................31 3.1 RESPONSE SPECTRUM CONCEPT ....................................................................31 3.2 RESPONSE SPECTRUM ANALYSIS APPLIED TO MDOF SYSTEMS ........................33 1) SRSS (Square Root of Sum of Squares) ......................................34 2) CQC (Complete Quadratic Combination)......................................35 3.3 DUCTILE BEHAVIOUR CONSIDERATION .............................................................35 4. SEISMIC RESPONSE BY TIMEHISTORY ANALYSIS .......................................39 4.1 RESPONSE OF A SDOF SYSTEM TO GENERAL DYNAMIC LOADING; DUHAMEL’S INTEGRAL ......................................................................................................39 4.2 LINEAR TIME HISTORY ANALYSIS FOR MDOF SYSTEMS ....................................41 4.3 TIME HISTORY ANALYSIS FOR EARTHQUAKES ...................................................42 Stepbystep integration method with linear variation of the load.........43
......................................58 REFERENCES........................................53 Internal forces due to earthquake loading in vertical direction .......................5 TIMEHISTORY RESPONSE ANALYSIS OF THE BRIDGE ..........54 6..................................................1 6...............................................55 6.................3 6.....................................................54 6.......4 STRUCTURAL MODEL OF THE BRIDGE.............................47 FREQUENCIES AND VIBRATION MODE SHAPE DETERMINATION FOR THE BRIDGE .............49 RESPONSE SPECTRUM ANALYSIS OF THE BRIDGE .54 Combination of Orthogonal Seismic Effects........................................................ CASE STUDY ......60 2 .......7 EQUIVALENT STATIC ANALYSIS OF THE BRIDGE .................................6 RESULTS OF THE TIMEHISTORY RESPONSE ANALYSIS ................51 RESULTS OF THE RESPONSE SPECTRUM ANALYSIS ................................Standard Methods for Seismic Analyses 5....................................................47 6..... EQUIVALENT STATIC METHOD.............................45 6...................................53 Internal forces due to earthquake loading in horizontal direction ................................................................2 6................................................54 Displacements .........
who has given important and useful comments and suggestions. Rita Bento. Lyngby.Sc. Instituto Superior Técnico.D. Portugal. sponsors the Ph. project under which this report was done.D. for providing the example used in the case study and for assistance in carrying out the study. A word of appreciation should also be addressed to Civil Engineer Ph D Junying Liu.Standard Methods for Seismic Analyses Acknowledgements The work has been carried out at the Department of Structural Engineering and Materials. M. Nielsen. COWI A/S. The Portuguese institution for scientific research Fundação para a Ciência e Tecnologia – FCT. The author would like to thank his supervisor for giving valuable advice and inspiration as well as valuable criticism to the present work. Dr. July 2003 João Luís Domingues Costa 3 . Technical University of Denmark (BYG • DTU) under the supervision of Professor. Lisbon. techn. The author grateful acknowledges this support. Thanks are also due to the author’s cosupervisor M. P. Ph.
Standard Methods for Seismic Analyses 4 .
the reader is introduced to some of the corresponding basic concepts from elementary Structural Dynamics. i. For further study a number of references are given. This document is intended for students or civil engineers who want to have a basic knowledge about earthquake analysis. the static equivalent method. 5 .e. is also introduced. Both methods assume linear behaviour of the structure. For its simplicity. usually used in the predesign phase of regular structures. The first one is widely used as it applies to the major part of a seismic analysis necessary for design purpose.Standard Methods for Seismic Analyses 1. Before discussing seismic analysis in particular. proportionality between deformations and forces. The methods to be discussed are the response spectrum method and the linear timehistory analysis. The timehistory response method provides more detailed information regarding the seismic behaviour of a structure and is therefore used for more specific earthquake analyses. Introduction The following report gives a general survey of the most important methods nowadays at disposal for a structural engineer when performing a seismic design of a given structure. The theoretical information given in this report is complemented with analysis of a bridge similar to one designed for the High Speed Transportation System in Taiwan. It should be noted that this report does not intend to be neither a reference book nor a Structural Dynamics or Earthquake Analysis textbook.
Standard Methods for Seismic Analyses 6 .
1) Referring to Figure 1 b) the resultant force acting on the mass. and the damping forces. are determined using Hooke’s law: fk = k ⋅ q (t ) Chapter 15 and chapter 22 of reference 1 on the formulation of the Equation of Motion is recommended. In dynamical terms. and its energy loss mechanism (damping). 2 1 A dot means differentiation with respect to time 7 . See figure 1 a). b) Dynamical equilibrium of a SDOF system The primary objective in a structural dynamical analysis is to evaluate the time variation of the displacements and to accomplish this the Equation of Motion must be formulated and solved. F(t).1)2. The equilibrium condition may then be written as follows: fi + fc + fk = p(t ) (2. fc. m. Structural Dynamics Fundamentals 2. The dynamical equilibrium condition is given by (2. fi. F (t ) −f i (t ) = 0 ⇔ && ⇔ F (t ) − m ⋅ q (t ) = 0 (2. fk. may be defined as the difference between the external loads p(t) and the sum of the elastic forces.3) Elastic forces. Figure 1 – a) Simplified sketch of a SDOF system. q. a system is called a Single Degree of Freedom (SDOF) system if all these properties may be modelled by a physical element with only one component of displacement. its elastic characteristics (stiffness). fk.1 Formulation of Equation of Motion for SDOF Systems The essential properties of any linearly elastic structural system subjected to dynamical loads include its mass. proportional to its acceleration and opposing the acceleration.Standard Methods for Seismic Analyses 2. k. which implies that the mass develops an inertia force.2) (2. One of the methods to formulate the Equation of Motion1 is direct use of Newton’s second law. characterized by a number c.
Therefore the damping influence is usually quantified on the basis of experience.e.5) is of the form: && m ⋅ q (t ) + k ⋅ q (t ) = 0 (2.7) into (2. These forces are part of a mechanism transforming the mechanical energy of the system to other forms of energy such as heat. the substitution of equation (2.3) and (2. If damping is disregarded the equation of motion (2.6) This is a homogeneous second order linear differential equation with constant coefficients. i.m ⋅ ω 2 + k ⋅ A ⋅ cos(ω ⋅ t ) = 0 ) (2. the term in the first parenthesis must be equal to zero. the superposition of the two solutions above is also a solution. that these are solutions to the differential equation (2. 2. For instance. giving: ω= k m (2. m. Therefore one may write the general solution as: q (t ) = A ⋅ sin (ω ⋅ t ) + B ⋅ cos (ω ⋅ t ) (2.4) Introducing equations (2. fc. k and c and the external dynamical loads as follows: && & m ⋅ q (t ) + c ⋅ q (t ) + k ⋅ q (t ) = p(t ) (2. The mechanism is quite complex and still not completely understood. one may easily verify by direct substitution.2) one may write the equilibrium condition in terms of the coordinate q(t). & fc = c ⋅ q (t ) (2.5) This last expression is known as the Equation of Motion of a SDOF system. of the viscous type are proportional to the velocity.6) is linear and homogeneous.10) Since the differential equation (2.4) into equation (2.3.6).2 Undamped Free Vibrations The motion of a SDOF system free from external action or forces is governed by the initial conditions.Standard Methods for Seismic Analyses Damping forces.9) In order to satisfy this condition at any time t.11) Damping forces are always present in any physical system undergoing motion. the system properties. Considering solutions of the form or q (t ) = A ⋅ cos(ω ⋅ t ) q (t ) = B ⋅ sin(ω ⋅ t ) (2.7) (2.6) leads to: (. 3 8 .8) where A and B are constants.
q (0 ) . The equation of motion (2.e. Consequently the maximum displacement & will be of the same order as the initial displacement. in which the quantity ω is the circular frequency. q(0).14) where C is a constant.3 Damped free vibrations We now discuss a SDOF system vibrating freely but we include the effect of the damping forces.12) This last equation is the equation for the motion of an undamped SDOF system under freevibration conditions.14) into (2.15) Requiring the parenthesis to be zero we get: m ⋅ s2 + c ⋅ s + k = 0 (2. ω ⋅ sin(ω ⋅ t ) + q (0 ) ⋅ cos (ω ⋅ t ) (2. q(0).13) This differential equation is of the same form as before for the undamped case.16) The roots of this quadratic equation are: 9 . Dividing ω by the factor 2 ⋅ π .Standard Methods for Seismic Analyses The constants of integration A and B may be expressed in terms of the initial & conditions. Thus the solution becomes: q (t ) = & q (0 ) . f. the displacement. one obtains the natural frequency of the system. at time t=0. The response frequency is low and the maximum displacement is mainly governed by q (t ) = & ⋅ sin (ω ⋅ t ) . & q (0 ) 2.13) which leads to (m ⋅ s 2 + c ⋅ s + k ⋅ C ⋅ e s ⋅t = 0 ) (2. Expression (2. this parameter only depends on the system properties k and m. This is a simple harmonic motion. A very flexible (or very “heavy”) SDOF system has a large value for m (or low value of k). and so the response frequency is high and the displacements are mainly given by q (t ) = q (0 ) ⋅ cos(ω ⋅ t ) . This is proved substituting (2.5) then has the form: && & m ⋅ q (t ) + c ⋅ q (t ) + k ⋅ q (t ) = 0 (2. i. and the velocity. but the solution now is: q (t ) = C ⋅ e s⋅t (2. This implies that the maximum ω displacements may be larger than the initial displacement.10). expressed in Hz (cycles per second).12) may be used qualitatively to understand how the response is influenced by the stiffness and inertia properties of the system as well as the initial conditions: A very stiff (or very “light”) SDOF system has a large value of k (or low value for m). As shown by expression (2. q(0).
systems with damping below the critical value. Structural systems under normal conditions do not have values of damping above this critical value.10). the damped vibration frequency. and velocity.Standard Methods for Seismic Analyses S1 c c 2 ± −ω =− S2 2⋅m 2⋅m 2 (2. is given approximately by: q n +1 ≅ e −2⋅π ⋅ξ qn (2. only the situation for underdamped systems will be discussed.18) Depending on the value of c.17) As in the previous paragraph. in the following. when the initial condition of displacement.e. the general solution is given by superposition of the two possible solutions: q (t ) = C1 ⋅ e s1⋅t + C 2 ⋅ e s2 ⋅t (2. Under these conditions. So. one gets three types of motion. The value making the squareroot quantity zero is called the critical damping value. so it may be inferred that. c c = 2 ⋅ m ⋅ ω . For common structural systems ( ξ <20%) this value differs very little from the undamped frequency as shown by equation (2.20) The term in parenthesis represents simple harmonic motion. equation (2. introducing the parameters: ξ . as it is of the same form as equation (2. It may be shown that the ratio between two successive peaks. negative or zero. the constant of integration A and B can be evaluated and substituted into equation (2. are introduced. damping will not have any significant influence on the frequency of motion.19) & Finally. q 0 . It is of interest to note that the frequency for this harmonic motion is now given by ω d with the expression as above. and it may be shown that this value represents the largest value of damping that leads to oscillatory motion in free response. i. i. giving: & q + q0 ⋅ ξ ⋅ ω q (t ) = e −ξ ⋅ω ⋅t q 0 ⋅ cos (ω d ⋅ t ) + 0 ⋅ sin (ω d ⋅ t ) ωd (2. ω d = ω ⋅ 1 − ξ 2 q (t ) = e −ξ ⋅ω ⋅t [A ⋅ (ω d ⋅ t ) + B ⋅ sin (ω d ⋅ t )] c 2 ⋅ m ⋅ω (2. ξ = ω d . The effect of damping is more evident when considering the successive peak responses (see figure 2). according to the quantity under the squareroot sign being positive.e.18) may be written in a more convenient form.21) 10 . q 0 .19). qn and qn+1.e. for normal conditions.12). which is the damping ratio to the critical damping value i.
Standard Methods for Seismic Analyses q(t) qn q n+1 t Figure 2 – Plot of a freevibration response equation of motion for underdamped SDOF systems We may now formulate the equation of motion for SDOF systems by introducing the damping ratio.4 Resonant Response 4 To explain this important phenomenon.24) The study of SDOF systems cannot be completed without discussing the equations of motion for harmonic and periodic loading.23) where p0 is the maximum value and ω its frequency The equation of motion (2. for example. taking place when a structure is submitted to dynamical loading. The simplest load of this type is of the form: p(t ) = p0 ⋅ sin (ω ⋅ t ) (2.5) may now be written as follows: && & m ⋅ q (t ) + c ⋅ q (t ) + k ⋅ q (t ) = p0 ⋅ sin (ω ⋅ t ) (2. ω : p(t ) && & q (t ) + 2 ⋅ ξ ⋅ ω ⋅ q (t ) + ω 2 ⋅ q (t ) = m (2. However these subjects are not directly related to the standard methods for seismic design to be presented in this document. 4 11 . chapters 4 and 5 of reference 1 or chapter 3 of reference 2. response to harmonic loading will be considered.22) 2. The reader is referred to. and the natural vibration frequency. ξ .
The second term is called the steadystate response and it may be written in a more convenient form: q (t ) = ρ ⋅ sin (ω ⋅ t − θ ) (2. D. i.25) is called the transient response and because of its dependence on the factor e −ξ ⋅ω ⋅t . usually found in common structures. Therefore its evaluation is of little interest for the present discussion. It may be shown that this value is given in terms of the static displacement factor. ξ .29) θ = tan 1 1− β2 Several plots of the dynamical magnification factor with respect to β are shown in figure 3 for values of damping. which is called dynamical magnification factor: p0 ⋅D k 1 + (2 ⋅ ξ ⋅ β ) 2 p0 multiplied by the k ρ= with D expressed as: D= (2. ω The parameter β is defined as the ratio β = . ω The first term in (2. they depend on the initial conditions. As it may be seen the peak values of D are reached when β is very close to 1 (in fact.Standard Methods for Seismic Analyses One has now a nonhomogenous differential equation which solution is of the form: q (t ) = e −ξ ⋅ω ⋅t ⋅ [A ⋅ sin(ω d ⋅ t ) + B ⋅ cos(ω d ⋅ t )] + + p0 1 − β 2 ⋅ sin(ω ⋅ t ) − 2 ⋅ ξ ⋅ β ⋅ cos(ω ⋅ t ) ⋅ 2 2 k 1 − β 2 + (2 ⋅ ξ ⋅ β ) [( ) ( ) ] (2. the maximum value of the displacement.it damps out quickly. This means that when the load frequency approaches the natural vibration frequency of the SDOF system.26) The term ρ is the amplitude. 12 .e. when β = 1 − 2 ⋅ ξ 2 ).28) (1 − β ) 2 2 The value θ is called the phase angle and describes how the response lags behind the applied load: 2 ⋅ξ ⋅ β (2. the response will increase more and more.27) (2. This phenomenon is called resonance.e.25) Here: A and B have the same meaning as before i.
Standard Methods for Seismic Analyses
D 12
ξ=2% ξ=5%
8
ξ=10%
4
ξ=20%
0 0 1 2 3
β
Figure 3 – The dynamical magnification factor D as a function of β
Substituting the value of β for which D is maximum, one has the following expression for the maximum response, qmax:
qmax = 1 2 ⋅ξ ⋅ 1 − ξ
2
⋅
p0 k
(2.30)
The effect of damping on the resonant response is seen clearly: The lower is the damping value, ξ , the bigger the response. Theoretically for undamped conditions the value is infinite. The physical explanation for resonance is of course that both load frequency and natural vibration frequency of the system are so close that most part of the time the response and the load signals are in the same phase. This means that when the system is moving in a certain direction the load is in the same direction. This will lead to a consecutive amplification of the response in each cycle until the limit given by expression (2.30) is reached. For undamped conditions the response will grow indefinitely. It should be also noticed that for values of β near 0, i.e. when the natural vibration frequency of the system is much higher than the load frequency, D approaches unity. This means that the response will be closer to the static response. In fact, for highly stiff systems the quantity k ⋅ q (t ) is expected to play an important role in the final response.
13
Standard Methods for Seismic Analyses
2.5 Base Motion for SDOF Systems
Figure 4 shows a sketch of a SDOF system when submitted to base motion.
Figure 4 – SDOF system submitted to base motion
When a SDOF system is submitted to base motion, one may write the absolute displacement, q, in terms of the sum of the relative displacement, q*, and the support displacement, qs (figure 4). q = q * + qs (2.31) The formulation of the equation of motion leads to the same form as (2.2). However it should be noted that no load is acting on the system. The only action able to induce deformation on the system is the support displacement, qs . Therefore, as in (2.2), one may write the dynamical equilibrium condition:
fi + fc + fk = 0
(2.32)
Here:
&& Inertia forces, fk, are in terms of absolute coordinates, q .
Elastic forces, fk, and damping forces, fc, are in terms of relative coordinates, & q * and q * , respectively.
&& & m ⋅ q (t ) + c ⋅ q * (t ) + k ⋅ q * (t ) = 0
(2.33)
By means of (2.31) it’s possible to write the previous equation in terms of relative coordinates. This is more convenient for the purpose of achieving the effects on the system due to base motion:
&& & && m ⋅ q * (t ) + c ⋅ q * (t ) + k ⋅ q * (t ) = −m ⋅ qs (t )
(2.34)
Equation (2.34) is of the same form as (2.5). Therefore the response analysis of a SDOF system submitted to ground motion, in terms of relative coordinates, may be && treated assuming a load applied on the system equal to p(t ) = −m ⋅ qs (t ) .
14
Standard Methods for Seismic Analyses
Equation (2.34) may also be formulated in the same way as (2.22):
&& & && q * (t ) + 2 ⋅ ξ ⋅ ω ⋅ q * (t ) + ω 2 ⋅ q * (t ) = −qs (t )
(2.35)
Again we have a nonhomogeneous differential equation and so it is necessary to && find a particular solution, which depends on the form of qs (t ) . In chapter 4, the solution for base acceleration of general form will be discussed.
2.6 Formulation of the Equation of Motion for MDOF systems
From the discussion in the previous paragraphs, a degree of freedom is defined as an independent coordinate, necessary to specify the configuration or position of a system at any time, q(t).
a) Figure 5 – Examples of MDOF systems
b)
A structural system composed by more than one degree of freedom is called a MultiDegree of Freedom system (MDOF). Figure 5 shows two examples of MDOF systems. The establishment of the equations of motion for several degrees of freedom proceeds analogously as for the SDOF systems, which leads to a dynamical equilibrium condition of the same form as (2.2) for each degree of freedom. The result is a system of N differential equations, in which N is the number of degrees of freedom.
fi,1 + fc,1 + fk,1 = p1 (t ) fi,2 + fc,2 + fk,2 = p2 (t ) .......... .......... .......... .. fi,N + fc,N + fk,N = pN (t )
(2.36)
Each of the resisting forces, fi,i, fc,i or fk,i developed for a certain degree of freedom, i, is due to the motion of one degree of freedom. For example the elastic force produced for the degree of freedom 1, fk,1, is the sum of the different elastic forces acting at point 1, each one due to the displacement of each of any of the other degrees of freedom.
15
38) in which c1i are called the damping influence coefficients. bending stiffness of the columns EI and lengths of the columns L. 16 . It may be defined as the force at degree of freedom 1 due to a unit displacement corresponding to degree of freedom i. In figure 6 is illustrated the analysis of the stiffness coefficients in a twostorey frame with masses M1 and M2. Considering again the example above one has: fk.Standard Methods for Seismic Analyses Most conveniently the resisting forces may be expressed by means of a set of influence coefficients. They may be defined as the force at degree of freedom 1 due to unit velocity of the degree of freedom i.1 = ∑c i =1 N 1i & ⋅ qi (t ) (2.1 = ∑k i =1 N 1i ⋅ qi (t ) (2. La M2 (EI)b. Lb q1=1 k11 k12 k21 q2=1 k22 k11 = 2 ⋅ 12 ⋅ (EI )a L3 a k 21 = K12 = −2 ⋅ 12 ⋅ (EI )a L3 a k22 = 2 ⋅ 12 ⋅ (EI )a 12 ⋅ (EI )b + 2⋅ 3 La L3 b Figure 6 – Analysis of frame stiffness coefficients Analogously one may define the damping forces produced for degree of freedom 1: fc.37) in which k1i is called the stiffness influence coefficient. M1 (EI)a.
41) and & {q&(t )} = −ω 2 {q }⋅ sin(ω ⋅ t + θ ) one has (after omitting the sine term): observing that (2.5) for a given MDOF system as it expresses the N equations of motion defining its response5. 2. (constant in time) ω is the vibration frequency and θ the phase angle. 5 17 .e.12): {q (t )} = {q } ⋅ sin(ω ⋅ t + θ ) Here (2.36) may be written in matrix form: & & [M ]⋅ {q&(t )} + [C ] ⋅ {q (t )} + [K ] ⋅ {q (t )} = {p(t )} (2.41) By analogy with the behaviour of SDOF systems.1 = ∑m i =1 N 1i && ⋅ qi (t ) (2. assuming undamped conditions and no loads applied.7 Frequency and Vibration Mode Shape Analysis The problem of determining the vibration frequencies in MDOF systems is solved as for SDOF systems.40) is written as follows: & [M ]⋅ {q&(t )} + [K ]⋅ {q (t )} = 0 (2. i.11 the procedures leading to the solution of this system will be discussed.42) {q } represents the vibration shape of system. [C] and [K] .40) This equation is equivalent to (2. it is assumed that the freevibration motion response is simple harmonic. of the form (2.43) [K − ω 2 ⋅ M ⋅ {q } = 0 ] For further study of the formulation of the equations of motion for MDOF systems. the reader is referred to chapter 11 in reference 1 regarding the evaluation of the matrices [M].39) in which m1i are called the mass influence coefficients and may be defined as the force at degree of freedom 1 due to unit acceleration corresponding to degree of freedom i. Introducing the equation of motion into (2. In the following paragraphs until 2. i. For this situation equation (2.e. It is important to notice that the principle of superposition may be applied only if linear behaviour is assumed.Standard Methods for Seismic Analyses Finally the inertia forces produced for degree of freedom 1: fi. The set of equations in (2.
.. as for SDOF systems... [M ] .. Only ratios between these amplitudes may be established. is determined substituting the corresponding frequency. Another way is to assign the same value for a given degree of freedom in each vibration mode vector. It is of interest to notice that. The matrix. Either way it is convenient to express the vibration mode shapes in the normalized form...... [K ] . into equation (2. i. Thus the problem of determining the frequencies in a MDOF system results in an eigenvalue problem of the nonstandard form.e. into equation (2. etc.43)..45) . The eigenvalues are the squares of the frequencies and the eigenvectors are the vibration modes associated with the frequencies. ω N ) . ....43). (2..43) is homogeneous.. ω i .. ω i . of the system. the frequencies and the corresponding vibration mode shape depend only on the mass.Standard Methods for Seismic Analyses The only nontrivial solution of this equation is the one making the determinant of the matrix K − ω 2 ⋅ M equal to 0. φN1 φN 2 ..44) is called the frequency equation for MDOF systems. The lowest frequency (also called the natural frequency) corresponds to the first vibration mode. [Φ ] . each one representing a possible vibration mode.... ω 2 . This means that it’s impossible to determine the amplitudes of each degree of freedom in the corresponding vibration shape by simply resorting to equation (2.... φNN 18 .44). in dimensionless terms by dividing all the components by one reference component. the next higher corresponds to the second vibration mode. The resulting vector is called the nth mode shape φn.. {q } .. φ2N [Φ ] = 21 22 (2. and the stiffness. Therefore one has a 2 2 set of N solutions (ω12 .. with linear dependent equations and therefore indeterminate. It is usual to do it so to obtain an easy interpretation and comparison of the several vibration modes. It should also be noticed that the system resulting from substituting a given frequency. Expanding the determinant gives a polynomial expression of the Nth degree. It is obvious that there are infinitely many ways of computing the relations between the values of each vibration mode shape. φ1N φ φ .. Each shape vector.. [ ] K − ω2 ⋅ M = 0 (2.44) Equation (2. . assembling each of the mode shapes in a column is called the mode shape matrix and may be written as follows: φ11 φ12 . One of these ways is to normalize the vectors so that the largest value corresponds to unity..
which are very useful in structural dynamical analysis.50) Subtracting equation (2.e. Orthogonality w ith respect to the stiffness matrix 2 2 Dividing equations (2. the following expression is achieved: 2 T T φm ⋅ [K ] ⋅ φn = ω m ⋅ φm ⋅ [M ] ⋅ φn (2.46) by φ m one has: (2.48) Transposing equation (2.: [M ] = [M ]T and [K ] = [K ] . respectively.47) 2 T T φm ⋅ [K ] ⋅ φn = ω n ⋅ φm ⋅ [M ] ⋅ φn (2.8 Orthogonality Conditions The free vibration mode shape vectors.50) by ω n and ω m .51) It is evident from the discussion in paragraph 2.49) If equation (2.49) is multiplied on the righthand side of each member by φ n .52) This condition shows that the vibration mode shapes are orthogonal with respect to the mass matrix.7 that if m ≠ n.50) from equation (2. have certain special properties called orthogonality conditions.54) 1 2 ωm 19 . one has: 1 2 ωn T T ⋅ φm ⋅ [K ] ⋅ φn = φm ⋅ [M ] ⋅ φn T T ⋅ φm ⋅ [K ] ⋅ φn = φm ⋅ [M ] ⋅ φn (2.48) and (2. one has: T T 2 T φm ⋅ [K ] = ω m ⋅ φm ⋅ [M ] (2. [K ] ⋅ φn = ω n2 ⋅ [M ] ⋅ φn (2.Standard Methods for Seismic Analyses 2.43) may be written for the modes n and m as follows.46) 2 [K ] ⋅ φm = ω m ⋅ [M ] ⋅ φm T Multiplying equation (2.48) results in: (ω 2 n 2 T − ω m ⋅ φm ⋅ [M ] ⋅ φn = 0 ) (2.52) valid: T φm ⋅ [M ] ⋅ φn = 0 (2. the corresponding frequencies will be different.53) (2.47) and noticing that [M ] and [K ] are symmetrical. making the following equation (2. Orthogonality w ith respect to the mass matrix The dynamical equilibrium equation in the form (2. φn. i.
is composed by N independent modal vectors and therefore it is nonsingular and may be inverted.56) lead to: [Φ ]T ⋅ [M ]⋅ [Φ ] = [M ]G [Φ ]T ⋅ [K ] ⋅ [Φ ] = [K ]G in which the matrices [M ]G and [K ]G are of diagonal form. {q} using (2. Y. q.52) and (2. the amplitudes of which may serve as generalized coordinates to express any form of displacement. Y. may be written by superimposing suitable amplitudes. The procedure of determining the displacement vector.54) from (2. + φN ⋅ YN ⇔ ⇔ {q} = ∑φ n =1 N n ⋅ Yn (2.9 Modal Coordinates For dynamical analysis of linear systems with any kind of property (damped or undamped. {q} .57) (2.. [Φ ] . with or without loading) it is assumed that the displacements are represented in terms of the free vibration mode shapes.56) The results (2. {q} = φ1 ⋅ Y1 + φ2 ⋅ Y2 + .55) Thus for different vibration mode shapes the following orthogonality condition with respect to the stiffness matrix is valid: T φ m ⋅ [K ] ⋅ φ n = 0 (2. It should be noted that the modeshape matrix. These generalized modeamplitude coordinates are called modal coordinates. to the geometric coordinates.59) It is evident that the modeshape matrix serves to transform from the generalised coordinates. These shapes constitute N independent displacement patterns. so that it may Y be used in equation (2. { } .60) The problem lies now in determining the modal coordinates vector.Standard Methods for Seismic Analyses Subtracting equation (2. (2. {q} = [Φ ] ⋅ {Y } (2.60) is called mode superposition method.58) 2.60) in order to determinate the response of the system in geometrical coordinates.. This is the same to say that any displacement vector.53) gives the following condition: 1 1 T 2 − 2 ⋅ φm ⋅ [K ] ⋅ φn = 0 ω n ωm (2. of the N modes of vibration. This means that 20 . φ n .
Solving each of these equations and applying the mode superposition method leads to the establishment of the dynamical response of the system. should be divided is: {q }T ⋅ [M ] ⋅ {q }n n Finally the normalized vibration mode shape vector φn : φn = (2.n = ω n (2. {Y } = [Φ ]−1 ⋅ {q} 2.Standard Methods for Seismic Analyses it may always be solved directly for the modal coordinates amplitude. T 2 φn ⋅ [K ] ⋅ φn = ω n ⇒ 2 K G. It will now be shown that. The normalising procedure is called normalization with respect to the mass matrix.66) 2) Using the result expressed in (2.40) represents a set of N simultaneous differential equations coupled by the offdiagonal terms in the mass and stiffness matrices.64) As a consequence of this normalization. [M ] .10 Equation of Motion in Modal Coordinates (2. and may consist in writing the vibration mode shape vector. φ n . the reference component by which the nth vibration mode shape.61) The equation of motion (2. associated with any given displacement vector. with an appropriate normalizing procedure for the vectors φ n and regarding the orthogonality conditions observed previously. one has: [Φ ]T ⋅ [M ]⋅ [Φ ] = [I ] with [I ] as the N x N identity matrix.62) In order to determine.67) 21 . Y.46) by φn . so that the following condition will be valid: T φn ⋅ [M ] ⋅ φn = 1 (2.65) Another important result deriving from this type of normalization may also be shown: T 1) Multiplying both members of equation (2. {q }n . it is possible to transform the equation of motion into a set of N independent modal coordinate equations. one obtains: 2 T T φn ⋅ [K ] ⋅ φn = ω n ⋅ φn ⋅ [M ] ⋅ φn (2.63) {q }n {q }T ⋅ [M ]⋅ {q }n n (2.62).65) and remembering (2. {q} . (2.58). φn . using (2.
[Φ ] . This parameter may be interpreted as an energy loss mechanism associated with the corresponding vibration mode6.69).64). considering the transformation expressed in (2.40) 2) Multiplication of both members by [Φ ]T and introduction of the neutral element [Φ ]⋅ [Φ]−1 = [I ] in the first member.68). & & [M ]⋅ {q&(t )} + [C ] ⋅ {q (t )} + [K ] ⋅ {q (t )} = {p(t )} (2. Regarding damping. 22 . if the mode shape matrix. equals the square of the nth vibration mode frequency.nn as: T cG. as for the mass and stiffness matrices.61) and simplifying by means of (2.72) Two comments should be made about this equation: i.58) and (2. 1) Equation of motion in terms of the geometrical coordinates. The mode shape matrix. 2 T & & Y&n + 2 ⋅ ω n ⋅ ξ n ⋅ Yn + ω n ⋅ Yn = φn ⋅ {p(t )} (2. & & [Φ ]−1 ⋅ {q&(t )} + [C ]G ⋅ [Φ ]−1 ⋅ {q (t )} + [K ]G ⋅ [Φ ]−1 ⋅ {q (t )} = [Φ ]T ⋅ {p(t )} (2.67) and (2. [Φ ] .nn = φn ⋅ [C ] ⋅ φn = 2 ⋅ ω n ⋅ ξ n (2.69) where ξ n represents the nth mode damping ratio.68) It may be shown that. the damping matrix is written in a way that the orthogonality conditions are satisfied: [Φ ]T ⋅ [C ] ⋅ [Φ ] = [C ]G (2. then the matrix [C ]G is a diagonal matrix with each diagonal element cG. it will be assumed that. Yn.3 of reference 2. does not change with time which implies: {Y& } = [Φ] 6 −1 & ⋅ {q (t )} (2. In the following the steps that allow writing equation (2. [K ]G .65). the diagonal element at line n of the stiffness matrix. (2. is normalized according to (2.70) 3) Simplification considering the results (2.40) in terms of modal coordinates and therefore as a set of independent equations are described.71) It is evident now that one may write the previous equation for the modal coordinate.Standard Methods for Seismic Analyses Therefore. in the following form. & & [Φ ]T ⋅ [M ] ⋅ [Φ ] ⋅ [Φ ]−1 ⋅ {q&(t )} + [Φ ]T ⋅ [C ] ⋅ [Φ ] ⋅ [Φ ]−1 ⋅ {q (t )} + T −1 T + [Φ ] ⋅ [K ] ⋅ [Φ ] ⋅ [Φ ] ⋅ {q (t )} = [Φ ] ⋅ {p(t )} (2.73) The conditions regarding damping orthogonality are discussed in detail in chapter 133 of reference 1 and section 12.
once the vector { } is determined. in reinforced concrete structures submitted to dynamic loading. However.11 Base Motion for MDOF Systems The establishment of the equations of motion for several degrees of freedom MDOF systems follows the reasoning described above. Therefore no information beyond the elastic limit is provided such as the inelastic energy dissipation. application of the transformation (2. the stiffness distribution successively changes.n . In fact it is assumed that the motion response for the mode n (modal coordinate Yn) is the same as the motion response computed for a SDOF system with the properties m. presented in (2. it may be rather unrealistic to assume linear behaviour. q*.57) and (2. This similarity is the basic principle for carrying out a dynamical analysis using the mode superposition method assuming that the system behaves linearly. These are effects very difficult to take into account with the mode superposition method. relative coordinates.65): T && + 2 ⋅ ω ⋅ ξ ⋅ Y + ω 2 ⋅ Y = φn ⋅ {p(t )} & Yn (2. For instance. in which the normalizing procedure has been done with respect to the mass matrix. Therefore the following equation may be inferred from (2. As mentioned before. describing the equation of motion for SDOF systems. are used due to the convenience regarding the effects of base motion on the system. is evident.22).n The similarity between the previous expression and equation (2. ω n and ξ n . Again.72) using (2. 23 . for common structural systems subjected to extreme dynamical loading. This has a similar effect as damping and has a significant contribution to the structural response after yielding. Equation (2.60) leads to the global response of the system in terms of single degree of freedom equations in geometric coordinates.74) ii. As already discussed in the chapters referring to SDOF systems it is possible to solve equations (2. ω and ξ having the same values as the corresponding ones written in modal coordinates MG.72) is written in terms of modal coordinates. It is known that the formation of plastic hinges in a structure designed in a redundant way leads to the dissipation of energy transmitted by dynamic loading.72) or (2. not only due to the fact that certain elements are near yielding but also due to cracking.75) for each of the N modes and therefore achieve the modal coordinates Y Y vector { }.31).75) n n n n n MG.Standard Methods for Seismic Analyses & {Y&} = [Φ] −1 && ⋅ {q (t )} (2. since this method assumes that the structural properties remain constant in time. 2. as in a strong earthquake.
i and ki. Introducing the above vectors the equation of motion for MDOF systems submitted to base motion will have the following form: & & [M ]⋅ {q&(t )} + [C ]⋅ {q (t )} + [K ]⋅ {q (t )} = = − [M ] ⋅ ({ X }⋅ qsX (t ) + { Y }⋅ qsY (t ) + { Z }⋅ qsZ (t )) 1 && 1 && 1 && (2.1 + fk. in the form of (2.77) a set of vectors {1X } .... fi..1 = 0 fi. These are written so that nth line corresponds to the nth degree of freedom and the corresponding value will be unity. regarding the equation of motion in modal coordinates.. it is evident that if a degree of freedom..N = 0 (2.. given in paragraph 2... one has: In the present document whenever support motion is discussed for MDOF systems. composing the matrices [M ] . Therefore the symbol * will be omitted in the following expressions.1 + fc. is under direction J.. it is reasonable to neglect this fact due to simplification regarding the common structural dimensions..2 + fc. if the degree of freedom is in the same direction as that of the vector.i.40).2 = 0 . [C ] and [K ] . will be written as follows: fi.. {q&sY (t )} and {q&sZ (t )} . mi. only the inertia forces. It will be assumed here that the first two directions are in the surface plane and Z corresponds to vertical direction.6 about the influence coefficients..N + fc.78) It is obvious that this equation is of the same form as (2. i..i. the relative coordinates are used. fi .. are in terms of absolute coordinates..Standard Methods for Seismic Analyses Because no dynamic load is applied on any degree of freedom. Therefore it is practical to introduce into equation of motion (2. { Y } and 1 {1Z }. Considering again the nth modal coordinate.. and Z.N + fk. Reducing (2.76) As in SDOF systems...36).2 + fk.i .. may then be applied. 7 24 . corresponding to direction X. The procedures described in the previous paragraph. It should also be noted that the support acceleration vector has three components & & & {q&sX (t )} .77) && The vector {qs (t )} is the support acceleration vector and depends on the particular support conditions.. .76) to relative coordinates and expressing the equation in matrix form leads to: & & & [M ]⋅ {q& * (t )} + [C ]⋅ {q * (t )} + [K ]⋅ {q * (t )} = −[M ]⋅ {q&s (t )}7 (2. ci. otherwise it is zero. the set of equations of motion. Y. Referring to the explanation. However. only the motion of the degrees of freedom under J direction will affect the motion of the actual degree of freedom i..
it is possible to solve the equation of motion in the form (2. For instance.60).80) without using Pnj (first line in (2. the mass distribution and the direction of each degree of freedom. As it may be inferred it only depends on the vibration mode shape. From now on it will be omitted due to simplification. in a given MDOF system.80) It was mentioned before that the support motion from an earthquake is of the form of an excitation. By superposition analysis. 2 & & && Y&n + 2 ⋅ ω n ⋅ ξ n ⋅ Yn + ω n ⋅ Yn = −PnJ ⋅ qsJ (t ) (2. regarding the linear behaviour of the system. which means that it may be solved analogously as for a SDOF system.80) is of minor interest.81)). that the modal participation factor serves also as a measure of each mode contribution for the response in geometric coordinates. 2 & & && Y&n′ + 2 ⋅ ω n ⋅ ξ n ⋅ Yn′ + ω n ⋅ Yn′ = qsJ (t ) ⇒ Yn′ Final Modal Coordinate.82) It appears from this expression. 25 . to the following differential equation. This parameter may be used again to compute the actual modal coordinate by simply multiplying it by the solution determined as mentioned above (second line in (2. for mode n and direction J. PnJ. as the modal participation factor is a dimensionless parameter and the behaviour of the system is linear. is possible to solve the equation separately for each direction.81) As before.Standard Methods for Seismic Analyses 2 & & Y&n + 2 ⋅ ω n ⋅ ξ n ⋅ Yn + ω n ⋅ Yn = T T T && = −φn ⋅ [M ] ⋅ { X }⋅ qsX (t ) − φn ⋅ [M ] ⋅ { Y }⋅ qsY (t ) − φn ⋅ [M ] ⋅ { Z }⋅ qsZ (t ) 1 1 && 1 && (2. than other modes having their displacements mainly in other directions. q n.72).80) is of the same form as (2. Moreover. Therefore the minus sign in (2. is denominated the 1 modal participation factor of the nth mode for direction J.i ⋅ PiJ ⋅ Yn′ (t ) (2. Equation (2. the equation of motion under direction J for the nth degree of freedom may be computed applying the transformation (2. affecting each acceleration value qsJ . consider the response of a degree of freedom under X direction.J (t ) = ∑φ i =1 N n. Generally the sign of the response does not have any important significance in an earthquake analysis. which will lead.79) T && The term φn ⋅ [M ] ⋅ { j }.81)). Yn ⇒ Yn = PnJ ⋅ Yn′ (2. It is expected that modes with displacements mainly under X direction will contribute more to this response.
85) According to the principle of conservation of energy. and the potential energy of the spring. q(t)=0. The motion of this system may be assumed harmonic i.: q (t ) = Z0 ⋅ sin (ω ⋅ t ) (2. the force of the spring is 0 and the velocity is maximum. act to absorb the energy when the system is freely vibrating.Standard Methods for Seismic Analyses 2. The entire energy of the system is then given by the kinetic energy of the mass: Tmax = 1 2 ⋅ m ⋅ (Z0 ⋅ ω ) 2 (2. Thus the same result is established as in (2. Under these conditions is evident that: when the systems is in its neutral position.10): 1 1 2 ⋅ m ⋅ (Z0 ⋅ ω ) = ⋅ k ⋅ Z0 2 2 ⇔ ⇔ ω= k m (2.e. Consider a simply supported beam as shown in figure 7. 26 . Selection of the The main advantage of this method is that it provides a simple procedure to determine a good approximation of the natural frequency of MDOF systems.86) Approximate analysis vibration shape of a general system. for the present conditions.12 Vibration Analysis by the Rayleigh Method The Rayleigh method is widely used as it provides a simple method of evaluating the natural frequency both for SDOF and MDOF systems. This implies that the energy of a SDOF system. The total energy in this case consists of the sum of the kinetic energy of the mass. as shown in figure 1. T.84) when the system is at maximum displacement the velocity of the mass equals 0 which means that the entire energy of the system is the potential energy of the spring: Vmax = 1 2 ⋅ k ⋅ Z0 2 (2. Basic concepts The basic concept in this method is the principle of conservation of energy. Z0 ⋅ ω . V. the previous expressions must be equal. must remain constant if no damping forces. fc.83) where Z0 is the amplitude and ω the frequency.
90) where m(x) is the mass per unit length. varying harmonically in time (see figure 7): q (x.88) Introducing equation (2. Z(t).87) is: T = 1 ⋅ 2 ∫ dq (x. is given by the following expression. t ) = ψ (x ) ⋅ Z (t ) (2.89) The kinetic energy of the beam vibrating as assumed in (2. representing the ratio of the displacement at any point x to a reference displacement. d 2q (x. The previous assumption of the shape function. as shown in figure 7. Z(t). This may be achieved writing the deformed shape in terms of a shape function.87) with Z (t ) = Z0 ⋅ sin(ω ⋅ t ) . V.Standard Methods for Seismic Analyses Figure 7 – Simply supported beam with a selected deformed shape possible This beam may be considered as a MDOF system as it has an infinite number of degrees of freedom. t ) m (x ) ⋅ dx 0 dt L 2 (2. ψ (x ) . The flexural strain energy. of a prismatic beam. Proceeding as above to find the maximum strain energy. effectively reduces the beam to a SDOF system as the knowledge of a single function allows the evaluation of the displacement of the entire system.87) into this expression and letting the reference displacement. t ) 1 L V = ⋅ EI (x ) ⋅ dx 2 2 0 dx 2 ∫ (2. one may write the maximum kinetic energy as follows: 27 . To apply the Rayleigh method one has to assume a deformed shape for the fundamental mode of vibration so that it may be possible to compute the maximum potential and kinetic energy. EI being the bending stiffness. take its maximum value one finds the following expression for the maximum strain energy. ψ (x ) . Vmax: Vmax d 2ψ (x ) 1 2 L = ⋅ Z0 ⋅ EI (x ) ⋅ dx 2 0 2 dx 2 ∫ (2.
A good approximation to the natural frequency / vibration shape may be obtained considering the static performance of the system.93) Here Zn is the amplitude.87) the displacements for the degree of freedom n is given by the expression: qn (t ) = Zn ⋅ sin (ω ⋅ t ) (2. In the following the application of this procedure in determining the natural frequency of a MDOF system with N degrees of freedom is explained. the true vibration shape yields the lowest frequency. any shape other than the natural vibration shape requires the action of additional external constraints that contribute to stiffen the system and therefore to increase the corresponding frequency. ψ (x ) . is the deflected shape resulting from the application the weight in the direction where the principal vibratory motion is expected to take place. Wn. According to (2. Any shape function satisfying the geometrical boundary conditions may be selected as it represents a possible vibration shape. ψ (x ) . Therefore considering the system in figure 5a) one would assume the weight load being vertical as this is the direction where the vibration motions are expected to take place. The maximum potential energy is given by: 28 .91) The application of the principle of conservation of energy leads to the following natural vibration frequency: ω= ∫ d 2ψ (x ) EI (x ) ⋅ dx 2 0 dx L 2 ∫ m(x ) ⋅ψ (x ) dx 0 L 2 (2. However. Consequently from the infinity of vibration shapes possible in a general system. The frequency is then evaluated assuming that the vibration shape. which depends on the position of the mass and may be taken as the displacement at the degree of freedom when the system is acted upon by the weight load.92) The accuracy of the vibration frequency obtained by the Rayleigh method depends entirely on the shape function assumed. One common assumption is to identity the inertia forces with the weight of the masses in the system. In a multistorey building the vibration shape is mainly due to horizontal displacements of each storey and so the inertia forces should be put in the horizontal direction.Standard Methods for Seismic Analyses Tmax = 2 1 2 2 L ⋅ Z0 ⋅ ω m(x ) ⋅ ψ (x ) dx 0 2 ∫ (2. The potential energy is given by the sum of the work of each weightload.
93). respectively. is: ω = g⋅ ∑W ∑ i =1 i =1 N N n ⋅ Zn (2. & One gets qn.Standard Methods for Seismic Analyses Vmax = 1 ⋅ 2 ∑W i =1 N n ⋅ Zn (2.96) 2 Wn ⋅ Z n 29 .max = ω ⋅ Zn . may be easily found using equation (2.95) Thus the frequency in a MDOF with N degrees of freedom determined by equating the maximum values for the strain and kinetic energies.94) The maximum velocity of mass number n. Therefore the maximum kinetic energy may be written in the form: Tmax = 1 ⋅ 2 ∑ Wn 2 2 ⋅ ω n ⋅ Zn i =1 g N (2.
Standard Methods for Seismic Analyses 30 .
it is necessary to evaluate the value of the maximum response. 3. which may be easily determined once its equation of motion. is the same as for a SDOF system having ω = ω n and ξ = ξ n (see equation (2. and therefore to achieve the time dependent response of the referred SDOF system. 31 . is fully known. This is a linear approximate method based on modal analysis and on a response spectrum definition. submitted to the same external action.5. ω . Ynmax . the maximum modal response of the nth mode.75)) It should be emphasized that this procedure only leads to the maximum response.1 Response Spectrum Concept To explain the response spectrum concept.Standard Methods for Seismic Analyses 3. with a specified critical damping ratio. were already discussed in paragraphs 2. Figure 8 – Typical representation of response spectrum If the procedure of determining the maximum response is repeated for a sufficient range of SDOF systems. Seismic Analysis by Response Spectra Response spectrum analysis is perhaps the most common method used in design to evaluate the maximum structural response due to the seismic action.1 to 2. Therefore the response spectrum analysis is often considered to be the most attractive method for the seismic design of a given structural system. This saves up a lot of calculation effort with evident consequences in the time consumed and CPU requirements. it is possible to define a function and represent it in a diagram similar to the one shown in figure 8. and for different natural vibration frequencies. ξ . The procedures used to formulate and solve the equation of motion. For the response spectrum definition. one considers a SDOF system submitted to an external action that may be either an applied force or a support displacement. The maximum response is established for each mode by means of the adequate response spectrum. instead of fully describing the response. q (t ) . According to the analogy between SDOF and MDOF systems. q (t ) .
f i Figure 9 2) After a certain value of frequency.ξ ) . is the inverse of the cyclic frequency in Hz (cycles per second) 32 . ξ . the relative displacement tends to zero. Figure 8 represents a typical relative displacement response spectrum. F F i F Figure 10 It should be noted that the maximum responses. Sa (ω .5. in the same graph different response spectra. It is worth to analyse the evolution of the response spectrum function: 1) For low values of frequency. was already discussed in paragraph 2. This is easily explained if one remembers the concept of the natural vibration frequency.ξ ) may be presented in every desired form.ξ ) . 8 The period T. for displacements. in seconds. velocities.ξ ) . in a SDOF system.10). or even in the form of internal forces or bending moments in a given point of the SDOF system. one may see that the maximum value for the relative displacement tends to a certain value. 5% and 10%) are shown as in figure 8. S (ω . which is the support displacement. usually found in common structural systems.Standard Methods for Seismic Analyses This diagram is generally known as a response spectrum. S (ω . tend to zero.ξ ) . and accelerations. Usually it is represented with the xaxis being the natural vibration frequencies or periods of vibration8 of the SDOF and the yaxis being the corresponding maximum response values. qs. close to zero. described by expression (2. In fact high values of frequency correspond to a very stiff system. corresponding to the same action and to different damping ratios usually found in common structures (2%. In fact a SDOF with a low value of ω is very flexible and behaves as shown in figure 9 when submitted to a support displacement. for values of critical damping ratio. Sd (ω .ξ ) . q*. ω . i. The response motion will then be as shown in figure 10 – the relative displacements. The meaning of the relative displacement. Sd (ω . Sv (ω .e. Generally. q*.
may be found using to the analogy between equations (2. ω n and critical damping ratio.Standard Methods for Seismic Analyses The available response spectra used for design purpose. Sd (ω .80) and (2. is available.22) for MDOF and SDOF systems. for the SDOF system with both the same natural vibration frequency. Instead of solving mathematically an expression in the form of (2.2) 33 . ξ n . For direction J. and so the modal coordinate. in most of the Seismic Design codes. Yn′. the term PiJ may be omitted.J (t ) = ∑φ i =1 N n.B.max = PiJ ⋅ Sd.80). may be easily achieved if the displacement response spectrum. Sd (ω n . 3. N.J (ω n . the modal & participation factor is recovered as: Yn.ξ ) .ξ n ) .1): qn. Yn′.max is established from the response spectrum.i ⋅ PiJ ⋅ Yn′ (t ) (3. the maximum value for the modal coordinate in terms of displacements.2 Response Spectrum Analysis Applied to MDOF Systems It was concluded in chapter 2. respectively.11 that the equation of motion for the nth degree of freedom under a support excitation in direction J for a given MDOF system may be written as in (3.max . Yn′ (t ) .1) As mentioned. The procedure is illustrated in figure 11.ξ n ) & (3.max = Sd .: An accelerogram is a record of the ground accelerations either measured in a certain place or generated artificially. Figure 11 After establishing the maximum value for the modal coordinate. Yn′. are defined by means of an accelerogram representing a typical earthquake in the region of the structure.
or velocities. i.1).max . Therefore one has Sv (ω ) = ω ⋅ Sd (ω ) and Sa (ω ) = ω 2 ⋅ Sd (ω ) .5) This method usually gives good results if the modal frequencies of the modes contributing for the global response are sufficiently separated to each other. Thus this method can’t provide information on the failure mode of the structure. According to this rule the maximum response in terms of a given parameter. To minimize these disadvantages it is necessary to combine the modal responses. In fact this is the main disadvantage of the response spectra analyses: The result provided is a set of extreme values that don’t take place at the same time and therefore do not correspond to an equilibrium state. velocities. The assumption behind the reasoning expressed in (3. Y&n. Therefore only two methods are presented. which is an important information from the engineering point of view. ξ n ) (3. Yn.Standard Methods for Seismic Analyses In the same way one may calculate the maximum response in terms of & & accelerations. G≈ ∑ (G ) n n =1 m 2 (3. & Y&n.max . (displacements. These are determined remembering that each vibration mode will have an expression in the form of (2.J (ω n .3) & & Yn. 1) SRSS (Square Root of Sum of Squares) This is one of the most frequently used modal combination methods. It should be mentioned that there is some controversy about which method leads to better results. Yn. certainly will correspond to an upper limit of the global response with a low probability of occurrence. Alternatively this may be done by means of the socalled pseudoresponsespectra. since is very unlikely for the maximum modal responses to happen simultaneously. will be more adequate.e. In the design codes. Otherwise another method. & & 9 34 . max = PiJ ⋅ S a. to sum the maximum values of each modal coordinate.ξ n ) & (3.max = PiJ ⋅ Sv. Sa (ω ) or Sv (ω ) are accessible9. (Gn )2 .e.42) for the corresponding equation of motion. usually the first method to be discussed below is suggested. There are several ways of carrying out this and it is out of the purpose of the present text to discuss them. accelerations or even internal forces) may be estimated through the square root of the sum of the m modal response squares.max . G. i. such as the one following. J (ω n . if the corresponding spectra. However is up to the designer to choose more accurate procedures of combining the modal response if the SRSS method can’t be applied.4) We now discuss the problem of establishing a reasonable value for the global maximum response of the system. contributing to the global response.
shows that this material will roughly behave linearly until yielding and thereafter nonlinearly until failure. The symbols ε y and ε u stand for yielding and ultimate strains. This means that even for the maximum response situation the internal forces on the different structural elements of the system are assumed to be proportional to the displacements achieved. ρ in . whereas the SRSS method considers these to be independent.Standard Methods for Seismic Analyses 2) CQC (Complete Quadratic Combination) The reason why this method is more effective in evaluating the maximum response when the modal frequencies are close to each other is due to the fact that it considers the correlation between modal responses. In fact if two vibration modes have close frequencies their contribution to the global response is not independent. given by the following expression: ρ in = (1 − β ) 8 ⋅ ξ 2 ⋅ (1 + β in ) ⋅ β in 2 3 2 2 in + 4 ⋅ ξ 2 ⋅ β in ⋅ (1 + β in ) 2 (3. The correlation between modes i and n is estimated using the parameter.7) 3. in figure 12. G≈ ∑∑ ρ n =1 i =1 m m in ⋅ Gi ⋅ Gn (3.3 Ductile Behaviour Consideration As may be understood by the discussion so far. Usually this method is used if ω n +1 ω n ≤ 1.6) The parameter β in is β in = ωi . earthquake analysis by response spectra is based on the assumption that the system behaves linearly.5 . Figure 12 – Typical stressstrain curve for steel in uniaxial tension or compression 35 . respectively. a sketch of the stressstrain curve for steel. However this hypothesis is far from reality for structural materials as reinforced concrete or steel. For instance. ωn The global response is achieved applying the following expression.
The larger this value the more ability of the material to dissipate energy after yielding. η . σ. and therefore the more ductile. Figure 13 Therefore if the response spectra method is used to design a structural system. one can utilize the nonlinear behaviour and design structures for less values of stresses. The horizontal motion of the mass will induce bending moments on the column whereas the vertical motion of the mass will lead to a compression / 36 . the MDOF system in figure 5 b). with a sufficient number of plastic hinges allowed before collapse. The determination of this coefficient is also a matter of controversy. The seismic design criteria consider that a structure submitted to an extreme earthquake should be prevented from collapse but significant damage is expected. Consider. F. or internal forces. It should be stated that ductility does not depend only on the material characteristics but also on the system and the direction of loading. yielding is expected which will lead to inelastic response of the structure. This idea is illustrated in figure 13. Assuming that the deflections.Standard Methods for Seismic Analyses The capacity of the material to absorb deformations in a stabilized way is called ductility. However. This is done by means of the coefficient. it is accepted that in order to maximize the nonlinear behaviour of the system and thus its behaviour coefficient. called the reduction factor or behaviour coefficient the physical meaning of which is shown in figure 13. Therefore this type of action must be included among the design load conditions for the Ultimate Limit State design. for instance. δ . the stresses / internal forces corresponding to the maximum deformations previously achieved may be reduced to take into account the yielding of the material. Usually.e. produced by a given earthquake are essentially the same whether the structure behaves linearly or yields significantly. Under these conditions. the value given for the behaviour coefficient is much less than the real one as the elastic response is reduced using further reduction coefficients (see chapter 6). it is desirable to design it in a redundant way i. One way of measuring ductility is the ratio of ultimate deformation to the yielding deformation.
the reduction factor is taken as unity. This is the reason why in most of the analyses. For the first situation the momentrotation curve will show that the element has capacity to absorb deformations after yielding and so ductile behaviour may be assumed. for vertical seismic action. On the other hand.Standard Methods for Seismic Analyses tension. 37 . the axial force – axial deformation diagram often show brittle behaviour and so η = 1 is usually adopted.
Standard Methods for Seismic Analyses 38 .
: p(t ) && & q (t ) + 2 ⋅ ξ ⋅ ω ⋅ q (t ) + ω 2 ⋅ q (t ) = m (2. t.e. may be expressed in the same form as (2.20). Therefore this method is one of the most effective for the solution of nonlinear response. q(t).g. and the dynamical loading. in the present text. the structural properties are assumed to remain constant during the entire loading history and further it is assumed that the structure behaves linearly. k) from one step to the next.22) It should be noted that both the response. i.20) If the starting time is different from 0.22) for a damped SDOF system. may be applied. During each step the response is evaluated from the initial conditions existing at the beginning of the step (displacements and velocities) and the loading history in the interval. p(t). The response to general dynamic loading of a SDOF system subjected to initial & conditions q0 and q 0 is deduced considering first the corresponding free vibration response as in equation (2. Nevertheless. depend on time. 4. a linear time history analysis is adopted i. Seismic Response by TimeHistory Analysis TimeHistory analysis is a stepbystep procedure where the loading and the response history are evaluated at successive time increments. & q + q0 ⋅ ξ ⋅ ω q (t ) = e −ξ ⋅ω ⋅t q0 ⋅ cos (ω d ⋅ t ) + 0 ⋅ sin (ω d ⋅ t ) ωd (2. p(t). With this method the nonlinear behaviour may be easily considered by changing the structural properties (e. stiffness.τ )) + ωd (4.τ ) = e −ξ ⋅ω ⋅ (t τ ) q (τ ) ⋅ cos (ω d ⋅ (t .τ )) q (t .e. among the many methods available. the above expression may be written in a general form introducing τ as the time corresponding to the initial conditions: & q (τ ) + q (τ ) ⋅ ξ ⋅ ω ⋅ sin(ω d ⋅ (t . ∆t – steps. Duhamel’s Integral to General Dynamic The equilibrium equation for a given general dynamic loading. due to load applied at another time τ.1 Response of a SDOF System Loading. already discussed in chapter 2. As a consequence the mode superposition method.Standard Methods for Seismic Analyses 4.1) 39 . The purpose of Duhamel’s integral is to achieve the response at any time.
dq (τ ) .4) The entire loading history may be considered to consist of a succession of such short impulses.2) ∫ The second term in this equation represents the area of the plot p(τ) in the time interval ∆τ. for t>τ: p(τ ) ⋅ dτ dq (t ) = e −ξ ⋅ω ⋅(t τ ) ⋅ sin (ω d ⋅ (t − τ )) m ⋅ ωd (4.2) as follows: & m ⋅ dq (τ ) = p(τ ) dτ (4. one may write the differential response. the total response may be established by summing all the differential responses developed during the loading history. each producing its own differential response according to the expression above. and reference 2. This is the same as saying that the response at time t is given by the integral of the differential displacements since time t=0 until time t.Standard Methods for Seismic Analyses Now we consider the same SDOF system acted upon by a load p(τ). There are several procedures to evaluate this integral and it is out of the purpose of this text to discuss them here. is a free vibration motion subjected to an initial & velocity. which leads to the result: & q + q0 ⋅ ξ ⋅ ω ⋅ sin (ω d ⋅ t ) + q (t ) = e −ξ ⋅ω ⋅t q0 ⋅ cos (ω d ⋅ t ) + 0 ωd (4.5) This result is known as Duhamel’s Integral and is one of the most important results in Structural Dynamics as it may be used to express the response of any damped SDOF system subjected to any form of dynamical loading.e. For a differential time interval. provide useful information about the evaluation of the Duhamel Integral for SDOF systems. the free damped vibration response must be added to the solution.3) Using the previous relation and noticing that the response after the termination of the short duration impulse. chapter 7. as follows. in the interval ∆τ given by the impulsemomentum relationship: & m ⋅ ∆q = p(τ ) dτ (4.6) t 1 −ξ ⋅ω ⋅ (t −τ ) + ⋅ p(τ ) ⋅ e ⋅ sin (ω d (t − τ )) dτ m ⋅ ω d ∫0 10 References 1. p(τ). This load & induces into the system a velocity variation.. Because the system is assumed to be linear. section 4.: q (t ) = 1 ⋅ m ⋅ ωd ∫ p(τ ) ⋅ e 0 t −ξ ⋅ω ⋅ (t −τ ) ⋅ sin (ω d (t − τ )) dτ (4. dτ. dq(t).10 To take into account initial conditions. this area is simply p(τ)dτ. 40 . i. which allows to rewrite equation (4. ∆q . p(τ ) dτ .
8) Once this procedure is done for all normal coordinates.n ⋅ (t − τ )) dτ (4.2 Linear Time History Analysis for MDOF Systems It may be inferred from the discussion held in paragraphs 2.n as expressed in & (2.8).: Yn (t ) = e −ξ n ⋅ω n ⋅t ω d.11.8) must be solved. Yn(t).6).6) for damped SDOF systems is composed by two terms with the same nature as discussed in paragraph 2.72) in which the vector {p(t )} represents the general dynamic loading applied in the corresponding degrees of freedom. The determination of the modal coordinates of a given MDOF systems.e. The first term reflects only the influence of the initial conditions and the second term corresponds to the loading effect on the structural response. one applies the expression (2. that the solution given by the Duhamel Integral may be used to determine the modal coordinates of a given MDOF system submitted to general dynamic loading.n ⋅ ∑ ∫ eξ i =1 0 N t n ⋅ω n ⋅τ ⋅ φin ⋅ pi (τ ) ⋅ sin (ω d . 2 T & & Y&n (t ) + 2 ⋅ ω n ⋅ ξ n ⋅ Yn (t ) + ω n ⋅ Yn (t ) = φn ⋅ {p(t )} (4.Standard Methods for Seismic Analyses As one may notice the general response in the form (4. The mode superposition method is then used to determine the global response of the system. i.8) would have to be written in the form of (4.4.7) The modal coordinate Yn(t). then a set of time intervals must be established taking into account the desired accuracy of the time history representation.n and q 0. If the time history has m time intervals then it is obvious that mxNxN equations in the form of (4. To obtain the global response of the system it is necessary to compute the equation of motion for the N degrees of freedom. This is done by means of expression (2.60) to obtain the time dependent equation of motion for each degree of freedom in geometric coordinates. If the system is submitted to initial conditions different from zero. is accomplished from equation (2. NxN equations in the form of (4. it is required to solve the set of N equations as (4. This will lead to the global response of the system at any desired time t.61) considering the vectors {q 0 } and {q 0 }.6 to 2. in order to obtain the equation of motion for a given degree of freedom at a time t in a MDOF system with N degrees of freedom. Therefore one may conclude that to establish the deformed shape of a structure at a certain time t. has the same form as (4.60). then it is obvious that equation (4. It should be noticed that. with ξ = ξ n and ω = ω n . 4. In most 41 .8) must be solved. For this case one & would have to compute the modal initial conditions q0. assuming that the system starts from rest.5). If one wants to represent the time history of the displacements.
However.10) The analogy between expression (4. q sJ (t ) . Y and Z. using the mode superposition approach and the assumed linear behaviour of the system. If expression (2. && Separating the support acceleration vector. As stated in paragraph 2. q s (t ) .11. it may be concluded.9) and (4.8) the term & [M ] ⋅ {1J } ⋅ q&sJ (t ) which leads to: Yn (t ) = e −ξ n ⋅ω n ⋅t {p(t )} by ω d. Yn(t). for some structures or certain types of analysis the number of degrees of freedom may be very high.3 Time History Analysis for Earthquakes As mentioned before. Thus. the modes corresponding to high frequencies have a small contribution for the response of the structure.n ⋅ ∫e 0 t ξ n ⋅ω n ⋅τ && ⋅ PnJ ⋅ qsJ (τ ) ⋅ sin (ω d. the definition of the modal coordinates Yn(t) may be done for each direction separately. are 42 .n ⋅ (t − τ )) dτ (4.11) & A solution is achieved by substituting into equation (4.9) may be written as: T && PnJ ⋅ q sJ (t ) = φ n ⋅ [M ] ⋅ { J } ⋅ q sJ (t ) 1 && (4. One of the most common techniques is to assume the load subdivided into a sequence of time intervals. steps.9) Remembering the expression of the modal participation factor. PnJ. we have for each degree of freedom a dynamic load given by the && product of the mass. Thus all the results in paragraph 2. that the decision about the number of degrees of freedom and the desired accuracy for the time history representation affect directly the number of calculations to accomplish and therefore must be chosen carefully taking into account the time consuming and the CPU requirements to proceed a time history analysis.12) The problem now consists in solving this expression above for each modal coordinate. in its three components. it is obvious that the second term in (4. 4. an earthquake action is considered as a base motion computed on the basis of the support acceleration. the equation of motion for the nth mode under direction J is the following: 2 & & && Y&n (t ) + 2 ⋅ ω n ⋅ ξ n ⋅ Yn (t ) + ω n ⋅ Yn (t ) = PnJ ⋅ q sJ (t ) (4. This is actually one of the main disadvantages of the method.80) is used. which makes the application of this method impracticable.5 and 2.7) is evident expressing the load vector {p(t )} as: & {p(t )} = [M ] ⋅ {1J } ⋅ q&sJ (t ) (4. and the corresponding acceleration value. m.11 may be applied.Standard Methods for Seismic Analyses cases less than N modes are considered since. along the axes X. in which the modal coordinates.
in which each acceleration component should be divided. For the next time intervals. i. Yn. equation (4.0 (τ ) ∆τ ⋅τ (4. i. it is possible to compute the corresponding global response in terms of geometric coordinates using superposition.n ∫ Once all the modal coordinates have been determined for the time i ⋅ ∆t . Therefore at the time i ⋅ ∆t the response for the modal coordinate n is of the same form as (4. X. ∆t.n ⋅ (t − τ )) dτ (4. Yn(∆t). qsJ . q sY (t ) and q sZ (t ) . becomes: Yn (∆t ) = e −ξ n ⋅ω n ⋅∆t ωd.i −1 and Yn.6).i 1 ω d.0 (τ ) + sJ ⋅ τ ⋅ sin (ωd .: && && q sJ (τ ) = q sJ . ∆ti.: & + Yn.n ⋅ PnJ ⋅ ∫ ∆t 0 && && q (τ + ∆t ) − qsJ . ∆ti1. Stepbystep integration method with linear variation of the load In order to perform a time history analysis of a given structure. The acceleration is assumed to vary linearly within the referred interval && && between the initial value. normally.14) ∆τ It should be noted that this expression is exact for the first time interval assuming that the system is at rest until the load is applied. regarding & the continuity of the response. The repetition of this procedure for each time interval leads to the time history response of the structure. and therefore && && && making automatically available to the designer the values of q sX (t ) . 43 .0 (τ ) e && + ⋅ PnJ ⋅ eξ n ⋅ω n ⋅τ ⋅ qsJ . According to the desired accuracy of the time history analysis.15) −ξ n ⋅ω n ⋅ ∆t && && ∆t qsJ (τ + ∆t ) − qsJ .n ⋅ t ) + Yn (i ⋅ ∆t ) = e −ξ ⋅ω n ⋅ ∆t Yn. in terms of displacements and velocities.e. and the final value.n (4. These parameters are achieved computing the response at the end of the previous time interval.n ⋅ (t − τ )) dτ 0 ∆τ ω d. Y and Z.i −1 must be determined. qsJ (τ + ∆t ) .0 (τ ) .i 1 ⋅ cos (ω d. the designer uses an accelerogram of a certain earthquake considered to be a typical seismic action. A complete accelerogram contains the record of the acceleration for the three directions corresponding to the three cartesian axes.0 (τ ) && eξ n ⋅ω n ⋅τ ⋅ qsJ .Standard Methods for Seismic Analyses calculated.n ⋅ t ) + n. As previously stated an accelerogram may be a record of the ground accelerations measured in a certain place during the period of an earthquake.13) Thus.12) for the modal coordinate.0 (τ ) + ⋅ τ ⋅ sin (ω d . the designer decides the number of time intervals. This procedure is called the stepbystep integration method and next we shall briefly describe one of the many different ways to solve it.i 1 ⋅ ξ ⋅ ω n Y ⋅ sin (ω d.e. the initial conditions.0 (τ ) + && && q sJ (τ + ∆t ) − q sJ .
Standard Methods for Seismic Analyses 44 .
the chapters 23 and 24 at reference 1 and the section 4. Usually an expression is defined to prescribe the minimum lateral seismic force.3. This is due to the fact that often the response is mainly controlled by the first mode of vibration. an inertia loading provides a good approximation to the natural vibration shape of the structure. in the Rayleigh method. also designated the base shear force. 12 11 45 .1. Regarding the determination and distribution of the static equivalent forces in a given structure. As discussed before. The structure to be analysed by the equivalent static method should respect certain criteria regarding its geometrical regularity and stiffness distribution such as12: All lateral load resisting elements (such as columns or walls) should run from the base to the top without any interruption: Mass and lateral stiffness should not change abruptly from the base to the top. in Reference 3 at section 4. It is prescribed in any relevant code for earthquake analysis and is widely used especially for buildings and other common structures meeting certain regularity conditions. This is the basic concept of the Equivalent Static Method. Thus. A complete set of requirements of this type is presented. If the structural response is not significantly affected by contributions from higher modes of vibration it is reasonable to assume that with an appropriate set of inertia forces one may achieve a good approximation for the response. imposing a minimum value of frequency the higher modes contribution may be neglected. Equivalent Static Method This method is perhaps the simplest procedure at disposal for a structural engineer to perform an earthquake analysis and achieve reasonable results. for example.3. The method is also called The Lateral Forces Method as the effects of an earthquake are assumed to be the same as the ones resulting from the statical transverse loadings.2. Each code presents its own procedure to compute and to distribute the static equivalent forces in order to achieve the earthquake effects on the structure11.3. which leads to a certain minimum value of frequency/stiffness.2 in reference 3 are recommended.3. Geometrical asymmetries in height or in plan due to setbacks should not exceed certain values.Standard Methods for Seismic Analyses 5. One usual requirement for the structure regarding the application of this method is that the natural vibration period of the structure should be limited by a maximum value.
Standard Methods for Seismic Analyses 46 .
The crosssection is a box girder. End Girder 8m Mid Girder 24 m End Girder 16 m Mid Girder 24 m End Girder 16 m Mid Girder 24 m End Girder 8m 1 Pier Tap 2 . The shear tap element is a concrete box with 2m height and of length 5.1 Structural Model of the Bridge A sketch of the bridge is shown in figure 14. 12. using the methods discussed before.Standard Methods for Seismic Analyses 6.35m tall and both are rigidly connected to a shear tap element at the top.8 m Pier 2 .2m 3 Pier Tap . 15.35 m 2 y Detail x Girder Tracks Rigid crosssection Figure 15 – Structural model (See table 1 for detailed information about cross section properties) 47 . The alignment of the main span axis is straight. 6. The piers are 15.80m and 12. Each span has a length 40m and a 13m width. Figure 14 The bridge is a threespan bridge with two rail tracks. similar to one designed for the High Speed Transportation System in Taiwan.4m for each side. Case Study The present chapter presents seismic analyses of a bridge.2m 1 Pier 1 .
The three spans are independent and simply supported at the abutments and shear taps. Three types of loads are considered: Self Weight – the weight of the entire structure which is carrying the loads. The abutments (see figure 15) allow translation in the same direction as the main span axis. iv.Standard Methods for Seismic Analyses The following assumptions are made for the structural model: i. (EI)e.: The support system assumed for the piers and abutments is too much on the conservative side. The global axes X and Y are shown in figure 15. equal to ½ EI is used. The pier supports are assumed to be fixed. Superimposed dead load on the span – 200 kN/m in vertical direction which includes the weight of the components other than the main structure stated above. vi.B.25 kN/m 60 kN/m Span Length Figure 16 – One train live load The geometrical parameters of each crosssection shown in Figure 15 are summarized in Table 1. Cracked column section with effective flexural rigidity. v. To take into account the torsional effects due to train loads. It is computed as shown in figure 16. 48 . For vertical members the local coordinate system is achieved applying a positive rotation of 90º on the global coordinate system.4 m 96. In fact it would be more realistic to admit spring systems to simulate it. The local coordinate system coincides with the global coordinate system for horizontal members. ii. Both abutments allow rotations perpendicular to the bridge plane and restrain all the others. N. However the procedure adopted is considered adequate for the present purpose. 6. iii. The axis Z is defined applying the righthand rule. the tracks are assumed to be connected to the girder centroid through weightless rigid members (see Detail in figure 15). One train live load – the weight of a train occupying one track.
9 9. for all elements (piers.53 0. local coordinate y of shear centre with respect to centroid. the material assigned for all the sections is concrete of class C25/30. Ixy. The drawings A. Therefore no mass density is considered for the concrete assigned for the Rigid crosssection and the Young’s Modulus.6 20.6 31.4 Izz (m4) 47.34 0 0 0 0 A. Moment of inertia about local axis z.. the use of uniaxial finiteelements. spans or pier tap’s) less than 4m long is considered adequate. respectively. is reduced to half the standard value for the Pier crosssection.2 11.83 ey (m) 0. E.82 1000 7.1 in the Appendix represent the identification of the members and joints adopted in this study The establishment of the degrees of freedom is done according to the mass distribution and the static loads applied: 49 . and vii. The model definition must represent the real behaviour of the system and plays a fundamental role in the accuracy of the results.1 70.1 120 16. Finally.7x103 29. ey.Standard Methods for Seismic Analyses Table 1 – Geometrical properties of the crosssections A (m2) End Girder Mid Girder Rigid crosssection Tracks Pier Tap Pier Here: 18.5 Ixy (m4) 68. called members. Cross section area. Iyy. Exceptions for the Rigid crosssection and Pier crosssections are made considering assumptions vi. Izz. Torsional moment of inertia. 6.9 12.7 1000 0.7 8. This means to define the degrees of freedom of the structure. Moment of inertia about local axis y.1 70.6 Iyy (m4) 109 80.2 1000 0.9 1000 0. Regarding the geometry of the bridge.2 Frequencies and Vibration Mode Shape Determination for the Bridge The first step to accomplish a dynamical analysis is to model the structure as a MDOF system.
Of course. each frequency and the corresponding vibration mode shape of the system may be determined using equations (2.30 50 .8 99. Once the mass and stiffness matrix are computed. One of the most common ways to overcome this situation is the mass participating criterion. regarding the automatic calculation systems available nowadays. which leads to more degrees of freedom and therefore larger calculation requirements to solve the eigenvalue problem. Under this criterion. Usually this value should be bigger than 70% For this study it is decided to use the first 125 modes. This procedure leads to a lumpedmass matrix with null offdiagonal terms and it represents the simplest form of defining the mass properties of a given structure. the response determined by considering only a few modes is a good approximation as long as the mass participating in it exceeds a certain value. the computation of the 639 frequencies and vibration mode shapes considered does not represent a significant computational effort. This means that the above procedure will be repeated successively as many times as the number of the degrees of freedom to achieve all the mode frequencies and vibration shapes. [K].89 99. In this study 640 degrees of freedom were computed. The sum of the loads applied on each member is concentrated at the middle and “transformed” to a mass dividing by the acceleration of gravity.44) and (2.43). Thus it is possible to define the mass properties of the structure assuming that the entire mass is concentrated at the nodes at which the translational displacements are specified. For the present simplified model. Table 2 shows. As one may remember there will be as many modes as degrees of freedom. ordered by ascending frequency values.81 m/s2.Standard Methods for Seismic Analyses Half of the mass of each member is considered to be concentrated in the nearest joint. may be found in any publication about finiteelements and it is out of the purpose to expose them here. 9. More accuracy in the results means bigger refinement of the model. for each direction. Generally the modes with lower frequencies contribute more for the global response. the larger this value the more accurate the results. Table 2 – Total mass participation factors for the first 125 modes Transverse direction Longitudinal direction Vertical direction Mass Participation % 97. the mass participation in terms of percentage of all the mass of the system. The procedures leading to the definition of the stiffness matrix.
Table 3 – The EL Centro’s NS component acceleration response spectrum scaled up by a factor of 2 for critical damping ratio.2. As expected. a response spectrum analysis may be done.11 0.01 0. it is worth to compare the results for the first mode shape and frequency given by the eigenvalue solution with the solution provided by the application of the Rayleigh method. However this is considered to be a good approximation if one compares the figures in Appendix A.2.92 0 0.Standard Methods for Seismic Analyses As it may be seen.7 9.58 14.61 0.11 Hz for the first mode.1 14.8 15. For the present analysis a response spectrum in terms of accelerations vs period. In fact. Therefore the weight load is applied on this direction in order to compute the natural frequency as in (2. with an evident decrease of the time consumed to perform the calculations.91 It should be noticed that the response spectrum used is only considered for periods up to 0.26 9. Table A. The result is 1. The following figures refer to the deformed shape of the first three modes. As one may see in table A. practically all the mass participates in the response.1.91 sec. ξ = 5% a (m/s2) T (sec) 6.2 of the first vibration mode shape given by the eigenvalue problem solution with the one given by the Rayleigh method. each one with the corresponding displacements in one predominant direction.1 summarizes the frequencies / period and each mass participation factor for each mode in the three directions.5 13. the response results determined considering only 20% of the modes are practically the same as considering all the modes. The application of the Rayleigh method assumes that the first mode shape will have displacements mainly in the longitudinal direction. 6.3 10.1 the solution of the eigenvalue problem leads to a frequency of 1.2. Finally.90 seconds making it pointless to compute the response analysis with spectrum values for periods greater than this as all the other modes will have lower periods. In the Appendix one may find the results for the eigenvalue problem solution.96).71 10. This spectrum is computed from the NorthSouth component of El Centro earthquake scaled by a factor of 2 and is shown in table 3.3 Response Spectrum Analysis of the Bridge Once the modal frequencies and the vibration mode shapes are computed.53 Hz.51 0.4 15. with only 20% (125/640) of the modes.81 0. According to the mass participation criterion. is an upper value of the real frequency.41 0. 51 .0 0. as shown in table A.3 18. T is assumed.21 0.31 0. the period of the first mode is 0.
Usually the soils are classified for earthquake analysis according to their consistence as soft or hard and/or according to the soil being sandy or argillaceous. Z. there are regions with a seismic intensity higher than others.Standard Methods for Seismic Analyses Because the present analysis is merely an example. A correct analysis would require the consideration of the response spectrum corresponding to the soil conditions of the area where the bridge is built. In most of the Seismic Codes.34. The structure analysed will be in a region for which the regional coefficient.2. equal to 2/3. αv. The behaviour coefficient assumed. Another factor the designer needs to take into account is the geographical localization of the structure. Before discussing the results for the present bridge it is worth to make the following consideration with respect to the modal participation factors of the modes shown in the figures of Appendix A. For instance the response spectrum above is used regardless the soil nature. The maximum value is 0.22. longitudinal and vertical. i. i. it is decided to adopt the CQC modal combination. It is known that the soil characteristics influence a great deal the way the seismic waves reach a structure and affect its dynamical behaviour. simplifications are assumed. For each direction. Thus the designer is requested a critical attitude when analysing the results attained. It is not usual to use the same response spectrum to compute the vertical loading as done in this example.4 and the minimum is 0. this fact is taken into account by scaling up or down the given response spectra by means of regional coefficients. is 0. the analysis using response spectra provides an envelope of the response. transverse. For the present case. It should be noted that since the earthquake action is in the form of an excitation. given the exemplificative nature of this text. it is decided to use the set of four regional coefficients. internal forces evaluated by means of linear analysis are reduced to 50%. In fact. In fact the vertical motions are generally of a lower intensity than horizontal. a correct analysis would require the use of a response spectrum typical for Taiwan instead of El Centro’s NS component.e. this is taken into account by reducing the vertical action using a coefficient. Since the mode frequencies were very close.e. Still. the expected earthquake intensity is scaled down to 85% of the one expected in the most sensitive region. Therefore the results are presented regardless of the sign. η. in the Code of Taiwan. For the present analysis. 52 . Z. is 2. response spectrum loads are created from the response spectrum shown in table 3. depending on many factors.
nonlinear behaviour is allowed and therefore the reduction factor. equal to 1.136 α y ⋅ η 1. The coefficient αy may be defined as the ratio of the seismic design action used to the seismic design action leading to formation of a sufficient number of plastic hinges for overall structural instability. allow the designer to reduce the member forces considering to structural type of the system using a coefficient.11 the modal participation factor.34 = = 0. In fact it is observed that for each mode the largest modal participation factor is achieved precisely for the predominant direction of the displacements. αy.87 166. The results to be presented correspond to the members and joints identified in the figures in Appendix A.01 Longitudinal direction Vertical direction 0. As is defined in some modern seismic codes for earthquake analysis.4 Results of the Response Spectrum Analysis The results obtained are processed in a different manner according to the direction of the loading and its type (displacements or forces in the members). In the following these results are presented separately. 53 . This may easily be confirmed by comparing the figures in the Appendix with the results in table 4. Therefore the reduction coefficient adopted for multiplying the internal member forces due to earthquake loading in the horizontal direction is given by: Z 0.Standard Methods for Seismic Analyses Table 4 – Modal Participation Factors for modes 1.22 5.1) Tables A.3. PnJ.02 2. Internal forces due to earthquake loading in horizontal direction For this type of results.84 124.73 0. 6.25.94 5.46 201. η.1.2 As referred in the end of paragraph 2. the Design Specifications elaborated by the Taiwan High Speed Rail Corporation.3.4 in the Appendix A.1 to A. is a good measure of the contribution of the nth mode for the global response in J direction.25 × 2 (6.3.3 present the reduced internal member forces for both earthquake loadings in transverse and longitudinal direction. is used as discussed in paragraph 4. Transverse direction Mode 1 Mode 2 Mode 3 0. 2 and 3.
34 = = 0. Displacements The displacements were reduced by simply using the same coefficients for Z and αv.2) Tables A.5 TimeHistory Response Analysis of the Bridge Once the modal frequencies and the vibration mode shapes are computed.4) Tables A. Therefore the following reduction coefficients are considered: Displacements due to earthquake loading in horizontal direction: Z = 0.227 (6.181 αy 1.34 = 0.6 in Appendix A.3 ⋅ Sy + 0. the results achieved are usually combined.3 present the reduced internal member forces for earthquake loading in vertical direction.3) (6.3 ⋅ Sx + 0.7 to A.3 ⋅ Sz = max Sy + 0.3. So the reduction coefficient will be of the same form as in (6.34 Displacements due to earthquake loading in the vertical direction: α v ⋅ Z = 2 3 ⋅ 0.5). Combination of Orthogonal Seismic Effects To account for the directional uncertainty of earthquake motions and the simultaneous occurrences of the corresponding internal forces in three perpendicular directions.3 ⋅ Sx + 0. 54 .5) 6.5 and A. The maximum displacement.25 (6.3 present the reduced displacements.9 in Appendix A. Smax is given by: Smax Sx + 0. α v ⋅ Z 2 3 × 0.3.3 ⋅ Sz Sz + 0.3. This is due to the fact that the response is mainly influenced by the vertical vibration modes.Standard Methods for Seismic Analyses Internal forces due to earthquake loading in vertical direction Usually ductility is not taken into account to compute the internal forces obtained when the earthquake acts under vertical direction.3 ⋅ Sy (6. which are normally associated with brittle behaviour.3. internal force or moment.1) assuming η equals to the unity. a time history analysis may be performed. For the present case the Taiwan High Speed Rail Corporation suggests the combination according to (6.
and the simplification used previously for the response spectra analysis.3 the considerations about the coefficients related to the soil nature. to the regional coefficients and to the vertical direction apply here. Each load corresponds to the application of El Centro’s NS component in one direction of the bridge. In this case two situations would have to be considered corresponding to the application of each horizontal acceleration component for both transverse and longitudinal direction of the bridge. Figure 17 represents the acceleration plot of the NS component of the El Centro earthquake.Standard Methods for Seismic Analyses Considering the exemplificative nature of this analysis and. As in chapter 6. in this analysis the same values for the reduction coefficients as adopted for the response spectrum analysis are used. a (m/s 2 ) 8 6 4 2 0 0 2 4 6 8 5 10 15 20 25 30 t (sec) Figure 17 – Accelerogram of the NS component of the El Centro earthquake scaled up by a factor of 2 A complete time history analysis of this bridge would require the use of three components of the acceleration vector. 6.6 Results of the TimeHistory Response Analysis The results are computed in the same way as in the response spectrum analysis. 55 . only the North – South component of the El Centro’s accelerogram scaled up by a factor of 2 is used to compute the three earthquake loadings. Therefore. This is due to the fact that both analyses rest on the mode superposition method based on the assumption that the system behaves linearly.
4.3.9 present the results similar to the ones presented in tables A. X.227 Because the same accelerogram is used to define the support acceleration in the three directions.9. M y (kNm) 24000 18000 12000 6000 0 6000 12000 18000 24000 Figure 18 – Time variation of the moment at the base of pier 1 due to earthquake loading in the transverse direction t(sec) 0 5 10 15 20 25 30 35 56 .4. the rule presented in (6. Each result refers to the maximum value during the whole time history reduced by applying the reduction coefficients summarized here: Member forces due to horizontal earthquake loading: 0.5) may be applied again.1 to A.340 Displacements due to vertical earthquake loading: 0. It should be noted that the value S is to be inserted regardless of the sign. The graphs shown in the following provide some examples of time variation of certain results.3. Y and Z.181 Displacements due to horizontal earthquake loading: 0.1 to A. As mentioned the timehistory method allows a much more complete analysis because it provides the time evolution of any kind of result. to account for the directional uncertainty of the earthquake motions and the low probability of simultaneous occurrence of the maximum response for each direction.136 Member forces due to vertical earthquake loading: 0.Standard Methods for Seismic Analyses Tables A.
Standard Methods for Seismic Analyses V (kN) 8000 6000 4000 2000 0 0 2000 4000 6000 8000 5 10 15 20 25 30 35 t (sec) Figure 19 – Time variation of the shear force in longitudinal direction at base of pier 2 due to earthquake loading in the longitudinal direction δ (mm) 32 24 16 8 0 0 8 16 24 32 Figure 20 – Time variation of the transverse displacement at midspan section of the middle span due to earthquake loading in the transverse direction t(sec) 5 10 15 20 25 30 35 57 .
Wtot = 72 894kN and Z. The base shear force for the earthquake acting in the transverse direction.6 × 72 894 = 12 734 kN 9.6 m/s2. the deformed shape from the inertia loading is an approximation to the natural vibration shape and therefore introduces additional stiffness/frequency and consequently higher internal forces in the structure.6) ⋅ Sa (T ) ⋅ tot V = α y ⋅η g where T is the fundamental period in the direction under consideration.3. we use the NS component of the El Centro earthquake scaled up by a factor of 2. Since we use a simplified method. as discussed in paragraph 2. Vz.2).4).136 × 12. The Design Specifications of the Taiwan High Speed Rail Corporation prescribe that the bridge shall be designed and constructed to resist a minimum lateral seismic force. Therefore equivalent static analysis is normally used only in the predesign phase for this type of structures.7) As expected. given by the expression: W Z (6.65 sec for the longitudinal direction (cf. is: Vz = 0. The sum of these internal forces equals 10 019 kN.12. In this case. this method provides good results when applied to structures meeting certain “regularity” conditions with respect to geometry. this parameter may be determined by the Rayleigh method which gives T=0. which implies a more careful analysis used in the design. Bridges are not usually part of this group of structures as they are normally rather complex. stiffness and mass distribution.81 (6. V. In fact.Standard Methods for Seismic Analyses 6. Wtot is the total weight of the structure accounting for the train loads. 58 .7 Equivalent Static Analysis of the Bridge As discussed in chapter 5. this value is higher than the ones obtained using the previous analyses. For instance. The value for the acceleration is computed from table 3 by linear interpolation and is equal to 12. paragraph 6. consider the results from the response spectrum analysis for the shear force in the longitudinal direction in the support joints (2 and 5) when the earthquake acts in the same direction (see table A. Moreover. αy and η have the same meaning as in the previous analyses. bridges are often important infrastructures in social and economic terms. as in the previous analyses. Sa(T) is the acceleration corresponding to the fundamental period determined by means of a typical response spectrum. However in this paragraph we illustrate the application of this method by computing the base shear force when the earthquake is in the transverse direction.
7) is to be applied for each direction of the bridge so that a complete set of internal forces and displacements may be obtained. This implies the determination of the fundamental period for the three directions.Standard Methods for Seismic Analyses It should be noted that expression (6.5) should be applied to obtain the maximum design values in terms of displacements and internal forces. 59 . As in the previous analyses a combination rule such as in (6.
Reprografia DECivil – Instituto Superior Técnico. Structural Dynamics: Theory and Computation – third edition.Standard Methods for Seismic Analyses References [1] [2] [3] [4] [5] [6] R. Luís Guerreiro. Design Specifications. Reprografia DECivil – Instituto Superior Técnico. Eurocode 8: Design of structures for earthquake resistance . 1975 Mario Paz. 1999. 1991.European Committee for Standardization.W. Dynamics of Structures.Part 1. CEN . Revisões de análise modal e análise sísmica por espectros de resposta. Clough and Joseph Penzien. Taiwan High Speed Rail Corporation. McGrawHill. 1991. Dinâmica de estruturas. João Azevedo and Jorge Proença. Van Nostrand Reinhold. 60 . Draft No 6. 2003. 2000.
Appendix .
A.1 Model Identification Members Identification Joints Identification .
.
ii.2. Mode 1 Mode 2 Mode 3 Mode 1 applying the Rayleigh Method . iv. iii.1 Deformed Shapes i.A.2 Eigen – Value Solution Table A.
019 0.001 0.011 0.82 14.244 1.019 0.000 0.56 22.029 0.11 51.65E+02 5.001 0.667 0.005 0.264 0.003 0.044 0.516 0.94E+04 4.045 0.023 0.30 19.93E+04 1.11 5.018 0.01 47.25E+05 1.001 0.016 0.055 0.956 0.69E+04 6.061 0.078 0.061 0.48E+04 6.89 31.72E+04 9.52 40.492 1.036 0.022 0.92E+04 3.080 0.01E+04 9.014 0.10E+05 1.73 13.073 0.244 0.190 0.023 0.099 Z 0.34 57.107 0.037 0.259 0.064 0.56E+04 1.19 21.002 0.053 0.30 58.647 0.163 21.013 3.55 13.77 56.38 27.75E+04 2.97 40.32 11.59E+04 6.070 0.000 0.421 0.072 0.04E+04 1.95 Period (sec) 0.027 0.01E+04 5.86 20.119 0.26 9.130 0.05 28.083 0.08E+04 2.000 0.110 0.000 0.856 3.024 0.84E+04 1.85 41.082 0.002 0.001 0.095 0.45E+03 7.041 0.038 0.000 0.007 0.011 0.62 37.026 0.02E+03 1.632 0.001 0.003 0.072 37.054 0.019 0.94 23.000 0.087 Y 0.017 0.039 0.189 0.017 0.25E+03 7.263 0.000 0.055 0.033 0.224 5.22 47.000 0.06 18.020 0.059 0.18 58.876 0.027 0.23E+02 5.89E+04 7.001 0.002 0.183 0.036 0.294 0.34 21.027 0.087 9.89 26.451 0.44E+03 7.017 Mass participation (%) X 94.29E+05 1.389 0.106 0.11 2.074 0.904 0.386 0.99E+04 2.79 44.406 0.355 0.68E+02 1.42E+03 4.343 0.81E+04 5.029 0.406 0.021 0.016 0.032 0.77E+04 1.001 0.016 0.000 0.000 66.045 0.011 0.097 0.72 52.047 0.890 0.632 0.000 0.72E+04 8.408 0.001 0.068 0.100 0.006 0.044 0.021 0.020 0.005 4.120 0.31 37.053 0.044 0.000 0.122 0.001 0.077 0.384 0.19 28.243 0.519 2.070 0.06E+05 1.57 34.26 17.369 0.025 0.028 0.000 0.37 3.647 1.023 0.000 0.004 0.024 0.83E+01 2.25E+04 2.116 0.81E+04 1.54E+03 8.78 3.31 10.34E+05 1.04 44.01E+04 2.15E+04 1.061 0.050 0.042 0.30E+04 3.37E+05 Frequency (Hz) 1.42 21.136 0.49E+04 5.589 0.217 10.11E+04 3.032 0.002 0.236 0.48 22.106 0.003 0.80 4.003 0.018 0.961 0.007 0.747 0.000 1.022 0.514 0.085 0.09E+05 1.024 0.132 0.000 0.77 50.016 0.74 13.008 0.012 0.91E+04 1.29E+04 1.21E+03 5.059 21.067 0.049 0.005 0.616 0.035 0.66E+04 7.046 0.039 0.082 0.59 22.10 22.70E+02 6.061 Mode Eigenvalue ( d/ 2) Frequency (H ) Period ( ) Mass participation (%) .015 0.91 35.004 0.047 0.067 0.Mode 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 Eigenvalue (rad/sec2) 4.09E+03 3.000 0.265 0.012 0.003 0.297 0.080 0.001 0.32E+04 1.036 0.490 0.070 0.61 52.25 16.637 0.05 18.63E+04 1.000 0.019 0.
039 0.026 0.003 0.059 0.032 0.017 0.90 93.027 0.010 0.001 0.79 109.007 0.74E+05 5.034 0.05 81.81 76.011 0.013 0.47 95.03 90.002 0.66E+05 3.010 0.28E+05 (Hz) 62.037 0.26E+05 3.32E+05 2.008 0.Mode 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 (rad/sec2) 1.014 0.007 0.010 0.019 0.011 0.004 0.20E+05 3.000 0.02 90.80E+05 3.136 0.016 0.000 0.13 135.000 0.169 0.018 0.001 0.19E+05 2.97E+05 6.007 0.009 0.003 0.83 114.013 0.035 0.61E+05 2.011 0.000 0.09 111.20E+05 5.26 98.000 0.001 0.014 0.43 66.030 0.28E+05 2.132 0.02 124.99E+05 3.011 0.55 121.013 0.24 131.008 0.018 0.003 0.07E+05 7.014 0.068 0.008 0.002 0.105 0.008 0.011 0.064 0.34E+05 4.019 0.15E+05 6.022 0.119 0.072 0.002 0.039 0.68 133.94 120.94E+05 5.405 0.002 0.011 0.000 0.43 123.001 0.92E+05 5.041 Mode Eigenvalue (rad/sec2) Frequency (Hz) Period (sec) Mass participation (%) X Y Z .012 0.62 96.72E+05 2.016 0.179 0.70 77.34 70.000 0.143 0.009 0.008 0.36 82.008 0.012 0.000 0.024 0.013 0.12E+05 2.097 0.016 0.37E+05 2.86E+05 1.005 0.001 0.000 0.010 0.536 0.003 0.083 0.009 0.000 0.001 0.004 0.68E+05 5.56 75.000 0.183 0.99 85.015 0.015 0.747 0.018 0.000 Z 0.39E+05 5.030 0.015 0.20E+05 4.006 0.000 0.71 87.78 73.053 0.90E+05 1.013 0.044 0.379 0.06 129.009 0.45E+05 3.000 0.000 0.028 0.74E+05 4.004 0.000 0.002 0.043 0.000 0.032 0.27E+05 6.79 126.99 68.001 0.060 0.010 0.78 (sec) 0.007 0.000 0.027 0.012 0.000 0.90E+05 2.55E+05 1.087 0.004 0.03 103.536 0.024 0.007 0.007 0.014 0.017 0.10E+05 7.001 0.59 116.000 0.006 0.66 89.000 0.002 0.69 63.59E+05 6.002 0.000 0.000 0.87E+05 4.402 0.025 0.41 79.77E+05 1.13 104.008 0.010 0.017 0.011 0.57 69.13E+05 3.013 0.351 0.47E+05 2.004 0.87E+05 4.002 0.015 0.000 Y 0.072 0.026 0.000 0.60E+05 1.60E+05 5.06E+05 3.109 0.79E+05 5.000 0.049 0.009 0.000 0.178 0.008 0.16 122.85 119.008 0.15 62.015 0.46 70.98E+05 2.81 134.193 0.009 0.001 0.31 74.216 0.008 0.10 99.000 0.015 0.051 0.73 66.036 0.008 0.007 0.145 0.10E+05 3.000 0.029 0.015 0.011 0.000 0.018 0.95E+05 1.011 0.14 119.174 0.007 X 0.009 0.012 0.80 116.008 0.00 76.047 0.008 0.109 0.53 111.152 0.000 0.003 0.000 0.06 88.002 0.27E+05 2.009 0.000 0.004 0.004 0.85E+05 7.74E+05 1.008 0.61E+05 3.03 88.53E+05 1.37E+05 5.000 0.
032 0.41 144.000 0.78 191.006 0.65E+05 9.006 0.54 170.003 0.000 0.007 0.42E+06 1.29E+06 1.005 0.002 0.66 156.000 0.018 0.007 0.006 0.021 0.000 0.010 0.000 0.006 0.07E+06 1.000 0.33 157.036 0.45E+06 138.29E+05 9.96 163.022 0.000 0.007 0.000 0.12E+06 1.000 0.103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 7.000 0.039 0.10 183.306 0.002 0.006 0.00 164.007 0.011 0.41 144.006 0.47 190.031 0.079 0.002 0.000 0.000 0.000 0.15E+06 1.038 0.30 168.005 0.88 151.09E+06 1.30 165.33E+06 1.000 0.22E+06 1.000 0.021 0.17E+06 1.000 0.58E+05 7.000 0.01 181.000 0.000 0.79 158.64 189.018 0.002 0.12E+05 8.05E+06 1.060 0.000 0.005 0.83E+05 9.000 0.003 0.007 0.000 0.005 0.020 0.006 0.000 0.08E+06 1.000 0.000 0.000 0.065 0.000 0.006 0.08E+05 9.006 0.000 0.025 0.006 0.000 0.74 166.438 .017 0.73 170.000 0.76 0.074 0.017 0.57 139.44E+06 1.007 0.000 0.031 0.97E+05 1.005 0.001 0.68E+05 8.046 0.15E+06 1.000 0.89 176.006 0.52 143.000 0.002 0.006 0.000 0.000 0.000 0.23E+05 8.006 0.61 171.
Deformed Shape – Mode 1 i. XY Plane view . Isometric view ii.
.
.
Deformed Shape – Mode 2 i. XZ Plane view . Isometric view ii.
.
.
Deformed Shape – Mode 3 i. Isometric view ii. XY Plane view .
.
.
Isometric view ii. XY Plane view .Deformed Shape – Mode 1 – using the Rayleigh method i.
.
.
3 Tables of Results for Response Spectra Analysis i. vii. iii. iv.A. ix. v. ii. vi. viii. Internal Forces in the Girder – Transverse Loading Internal Forces in the Piers – Transverse Loading Internal Forces in the Girder – Longitudinal Loading Internal Forces in the Piers – Longitudinal Loading Internal Forces in the Girder – Vertical Loading Internal Forces in the Piers – Vertical Loading Displacements – Transverse Loading Displacements – Longitudinal Loading Displacements – Vertical Loading .
Earthquake loading in transverse direction (kNm) Forces Span ID Element 1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30 Moments Torsion Bending Y Bending Z Joint Axial Shear Y Shear Z 1 14 14 15 15 16 16 17 17 0 10 10 17 17 20 20 22 22 57 57 57 47 47 40 40 32 32 3230 3230 3230 3203 3203 3081 3081 2994 2994 8454 8454 8454 8397 8397 7978 7978 7878 7878 74744 61962 61962 49224 49224 36822 36822 24838 24838 0 227 227 442 442 583 583 672 672 18 18 19 19 20 20 21 21 22 22 27 27 32 32 35 35 39 39 42 42 21 21 21 21 32 32 42 42 56 56 2840 2840 2632 2632 2404 2404 2132 2132 1787 1787 7657 7657 7369 7369 7139 7139 6867 6867 6742 6742 13773 13773 5594 5594 9307 9307 17166 17166 24033 24033 691 691 634 634 527 527 366 366 142 142 23 24 25 25 26 26 27 27 28 28 44 47 47 47 51 51 55 55 56 56 63 96 96 96 87 87 68 68 47 47 1513 2052 2052 2052 1529 1529 914 914 419 419 6469 3413 3413 3413 3115 3115 2520 2520 2027 2027 26992 31121 35119 35119 40991 40991 44393 44393 45763 45763 0 0 194 194 535 535 790 790 963 963 29 29 30 30 31 31 32 32 33 33 58 58 57 57 55 55 52 52 49 49 25 25 20 20 51 51 70 70 91 91 284 284 848 848 1453 1453 1965 1965 2552 2552 1524 1524 850 850 762 762 1061 1061 1470 1470 44940 44940 41618 41618 35850 35850 28047 28047 17987 17987 1039 1039 982 982 801 801 536 536 191 191 34 35 36 36 37 37 38 38 39 39 46 44 44 44 41 41 38 38 35 35 98 65 65 65 57 57 43 43 31 31 2872 715 715 715 1023 1023 1292 1292 1521 1521 1713 7657 7657 7657 7945 7945 8044 8044 8313 8313 12311 9704 8884 8884 6284 6284 5421 5421 9275 9275 0 0 132 132 374 374 542 542 655 655 40 40 41 41 42 42 43 43 44 44 31 31 27 27 23 23 20 20 15 15 17 17 16 16 30 30 41 41 51 51 1719 1719 1908 1908 2053 2053 2137 2137 2243 2243 8539 8539 8809 8809 9012 9012 9103 9103 9448 9448 15527 15527 22930 22930 31072 31072 39659 39659 48680 48680 705 705 680 680 582 582 431 431 236 236 8 0 59 2265 9476 57721 0 .3.1 .Member Forces in the Girder .Table A.
Table A.2 .Member forces in the Piers .Earthquake loading in transverse direction (kNm) Forces Pier ID Element 31 32 1 33 34 38 35 36 2 37 39 Moments Shear Z Torsion Bending Y Bending Z Joint Axial Shear Y 2 9 9 10 10 11 11 3 3 172 172 172 172 172 171 171 169 169 19 19 19 15 15 14 14 15 15 1737 1737 1737 1717 1717 1658 1658 1547 1547 2876 2876 2876 2874 2874 2871 2871 2865 2865 22935 16897 16897 11125 11125 5630 5630 1583 1583 77 20 20 38 38 76 76 118 118 4 5 12 12 13 13 6 6 167 174 174 174 174 174 173 173 15 19 19 19 16 16 16 16 1338 3242 3242 3242 3218 3218 3143 3143 2848 4065 4065 4065 4062 4062 4057 4057 3085 33124 21957 21957 11058 11058 1276 1276 142 65 11 11 53 53 103 103 7 171 16 2967 4039 6077 133 .3.
Table A.3.3  Member Forces in the Girder Earthquake loading in longitudinal direction (kNm)
Forces Span ID Element
1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30
Joint
Axial
Shear Y
Shear Z
Torsion
Moments Bending Bending Y Z
1
14 14 15 15 16 16 17 17
0
254 254 669 669 1011 1011 1283 1283
249
249 249 217 217 171 171 132 132
24
24 24 23 23 21 21 21 21
100
100 100 93 93 75 75 58 58
145
117 117 89 89 88 88 93 93
0
997 997 1780 1780 2334 2334 2695 2695
18
18 19 19 20 20 21 21 22 22
1582
1582 1881 1881 2152 2152 2422 2422 2760 2760
99
99 92 92 114 114 156 156 215 215
20
20 20 20 20 20 20 20 20 20
32
32 28 28 37 37 49 49 64 64
58
58 55 55 68 68 82 82 97 97
2751
2751 2546 2546 2157 2157 1533 1533 643 643
23 24
25 25 26 26 27 27 28 28
2994 408
408 408 754 754 1090 1090 1359 1359
254 162
162 162 92 92 73 73 112 112
20 21
21 21 21 21 22 22 22 22
70 42
42 42 38 38 34 34 42 42
105 126
111 111 99 99 84 84 77 77
0 0
329 329 678 678 663 663 498 498
29
29 30 30 31 31 32 32 33 33
1628
1628 1924 1924 2218 2218 2487 2487 2823 2823
139
139 126 126 80 80 49 49 97 97
22
22 22 22 22 22 22 22 21 21
49
49 43 43 30 30 35 35 53 53
83
83 63 63 73 73 97 97 112 112
567
567 809 809 929 929 849 849 432 432
34 35
36 36 37 37 38 38 39 39
3055 2972
2972 2972 2625 2625 2290 2290 2022 2022
143 328
328 328 263 263 192 192 135 135
21 20
20 20 20 20 20 20 20 20
60 94
94 94 86 86 63 63 42 42
117 103
103 103 109 109 101 101 96 96
0 0
665 665 1816 1816 2582 2582 3074 3074
40
40 41 41 42 42 43 43 44 44
1753
1753 1457 1457 1160 1160 889 889 551 551
96
96 105 105 153 153 205 205 261 261
21
21 21 21 21 21 21 21 23 23
29
29 40 40 68 68 84 84 100 100
98
98 63 63 49 49 57 57 87 87
3266
3266 3175 3175 2766 2766 2057 2057 1115 1115
8
0
297
23
110
133
0
Table A.3.4  Member forces in the Piers  Earthquake loading in longitudinal direction (kNm)
Pier ID
Element
31 32
Joint
Axial
Moments Bending Bending Shear Y Shear Z Torsion Y Z
Forces
2
9 9 10 10 11 11 3 3
278
278 278 276 276 273 273 269 269
3319
3319 3319 3309 3309 3274 3274 3206 3206
9
9 9 7 7 5 5 5 5
2
2 2 2 2 2 2 2 2
36
12 12 22 22 37 37 48 48
51776
40166 40166 28925 28925 17804 17804 6614 6614
1
33 34 38 35 36
4 5
12 12 13 13 6 6
263 297
297 297 296 296 293 293
3069 6700
6700 6700 6684 6684 6633 6633
6 12
12 12 11 11 9 9
2 3
3 3 3 3 3 3
54 47
16 16 40 40 69 69
856 82745
59634 59634 36911 36911 13710 13710
2 37 39
7
287
6506
8
3
83
992
Table A.3.5  Member Forces in the Girder  Earthquake loading in vertical direction (kNm)
Forces Span ID Element
1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30
Moments Shear Z Torsion Bending Y Bending Z
Joint
Axial
Shear Y
1
14 14 15 15 16 16 17 17
0
31 31 79 79 115 115 142 142
2598
2598 2598 2240 2240 1671 1671 1046 1046
159
159 159 157 157 151 151 145 145
945
945 945 877 877 732 732 591 591
3583
2964 2964 2335 2335 1723 1723 1142 1142
0
10394 10394 19428 19428 25828 25828 29903 29903
18
18 19 19 20 20 21 21 22 22
171
171 193 193 199 199 196 196 182 182
297
297 669 669 1373 1373 1960 1960 2452 2452
138
138 128 128 116 116 100 100 80 80
373
373 400 400 515 515 628 628 725 725
604
604 232 232 476 476 848 848 1145 1145
30744
30744 28168 28168 22732 22732 14952 14952 5341 5341
23 24
25 25 26 26 27 27 28 28
152 476
476 476 490 490 492 492 497 497
2592 2169
2169 2169 1973 1973 1538 1538 1020 1020
67 119
119 119 87 87 55 55 30 30
735 601
601 601 589 589 495 495 368 368
1264 1458
1667 1667 1980 1980 2167 2167 2235 2235
0 0
4392 4392 12353 12353 18336 18336 22364 22364
29
29 30 30 31 31 32 32 33 33
501
501 505 505 506 506 507 507 512 512
401
401 384 384 1052 1052 1556 1556 1985 1985
29
29 60 60 95 95 123 123 151 151
211
211 149 149 406 406 527 527 622 622
2176
2176 1974 1974 1644 1644 1253 1253 887 887
23880
23880 22472 22472 18305 18305 12140 12140 4351 4351
34 35
36 36 37 37 38 38 39 39
508 148
148 148 162 162 168 168 171 171
2111 2648
2648 2648 2446 2446 1962 1962 1377 1377
168 46
46 46 55 55 68 68 77 77
639 787
787 787 778 778 679 679 564 564
816 821
743 743 560 560 415 415 496 496
0 0
5362 5362 15222 15222 22858 22858 28298 28298
40
40 41 41 42 42 43 43 44 44
165
165 146 146 118 118 92 92 65 65
668
668 281 281 1046 1046 1676 1676 2250 2250
81
81 85 85 91 91 99 99 109 109
450
450 414 414 610 610 743 743 873 873
750
750 1060 1060 1400 1400 1767 1767 2169 2169
30866
30866 30035 30035 25898 25898 19249 19249 10489 10489
8
0
2610
114
932
2590
0
Table A.Member forces in the Piers .6 .3.Earthquake loading in vertical direction (kNm) Pier ID Element 31 32 Joint Axial Moments Bending Bending Shear Y Shear Z Torsion Y Z Forces 2 9 9 10 10 11 11 3 3 4507 4507 4507 4495 4495 4472 4472 4439 4439 424 424 424 414 414 398 398 385 385 156 156 156 153 153 147 147 138 138 133 133 133 133 133 132 132 132 132 1333 842 842 492 492 548 548 910 910 1997 554 554 935 935 2261 2261 3584 3584 1 33 34 38 35 36 4 5 12 12 13 13 6 6 4388 4446 4446 4446 4434 4434 4412 4412 382 522 522 522 511 511 493 493 124 224 224 224 221 221 215 215 131 192 192 192 192 192 192 192 1133 1676 993 993 591 591 917 917 4332 1938 380 380 1672 1672 3369 3369 2 37 39 7 4373 484 205 191 1260 4322 .
Earthquake loading in transverse direction Joint 1 8 18 4 29 7 40 Translations (mm) X 0.6 0.Table A.0 0.0 0.0 0.0 0.0 0.0 9.1 0.7 .0 0.0 0.0 0.3.0 0.0 0.2 23.0 0.8 30.0 0.2 7.0 0.Displacements .0 0.5 Y 0.0 0.0 .0 0.0 0.4 21.1 0.4 Z 0.4 0.0 0.0 Rotations (rad) X 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.0 Z 0.0 0.1 Y 0.0 0.
5 0.3 73.Earthquake loading in longitudinal direction Joint 1 8 18 4 29 7 40 Translations (mm) X 73.0 0.0 0.0 0.1 1.0 0.0 Z 0.0 Y 0.0 0.0 0.0 7.8 0.0 .0 0.8 Rotations (rad) X 0.Table A.0 0.0 0.0 0.Displacements .8 0.3.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 73.0 0.0 0.8 .0 9.5 72.0 0.0 0.0 0.0 0.8 Z 0.0 0.0 Y 0.9 72.5 73.0 72.
1 Y 0.2 0.0 0.0 0.0 0.Earthquake loading in vertical direction Joint 1 8 18 4 29 7 40 Translations (mm) X 0.0 0.0 0.2 0.3.9 .3 0.0 0.0 Z 0.0 0.0 0.4 0.2 0.0 0.0 0.0 Z 0.Table A.0 0.0 Y 0.1 0.0 0.Displacements .0 0.2 Rotations (rad) X 0.0 6.2 6.0 0.2 0.0 .0 0.2 0.2 0.3 4.2 0.5 0.0 0.0 0.0 0.0 0.2 0.0 0.3 0.
v. iii. viii. ix. Internal Forces in the Girder – Transverse Loading Internal Forces in the Piers – Transverse Loading Internal Forces in the Girder – Longitudinal Loading Internal Forces in the Piers – Longitudinal Loading Internal Forces in the Girder – Vertical Loading Internal Forces in the Piers – Vertical Loading Displacements – Transverse Loading Displacements – Longitudinal Loading Displacements – Vertical Loading . iv. vii.A.4 Tables of Results for TimeHistory Analysis i. ii. vi.
4.Table A.1 .Member Forces in the Girder .Earthquake loading in transverse direction (kNm) Forces Span ID Element 1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30 Moments Shear Z Torsion Bending Y Bending Z Joint Axial Shear Y 1 14 14 15 15 16 16 17 17 0 8 8 8 8 7 7 10 10 66 66 66 57 57 47 47 34 34 3349 3349 3349 3278 3278 3095 3095 2971 2971 8163 8163 8163 8074 8074 7660 7660 7557 7557 74498 61289 61289 48635 48635 36368 36368 24420 24420 0 264 264 541 541 718 718 819 819 18 18 19 19 20 20 21 21 22 22 16 16 22 22 26 26 29 29 31 31 18 18 17 17 33 33 49 49 66 66 2791 2791 2578 2578 2357 2357 2089 2089 1731 1731 7399 7399 7208 7208 7058 7058 6862 6862 6812 6812 14410 14410 6205 6205 9452 9452 16767 16767 23320 23320 812 812 739 739 608 608 410 410 149 149 23 24 25 25 26 26 27 27 28 28 36 40 40 40 44 44 51 51 55 55 73 121 121 121 108 108 85 85 57 57 1547 2060 2060 2060 1591 1591 985 985 449 449 6591 3388 3388 3388 3068 3068 2403 2403 1973 1973 26305 30514 34527 34527 40359 40359 43701 43701 45028 45028 0 0 244 244 672 672 1006 1006 1237 1237 29 29 30 30 31 31 32 32 33 33 59 59 58 58 53 53 48 48 42 42 24 24 20 20 61 61 87 87 111 111 311 311 812 812 1425 1425 1940 1940 2498 2498 1580 1580 922 922 700 700 1252 1252 1836 1836 44222 44222 41025 41025 35478 35478 27925 27925 18067 18067 1335 1335 1259 1259 1019 1019 673 673 242 242 34 35 36 36 37 37 38 38 39 39 40 36 36 36 34 34 33 33 29 29 120 79 79 79 70 70 53 53 36 36 2791 865 865 865 1081 1081 1252 1252 1506 1506 2135 7638 7638 7638 7900 7900 7974 7974 8223 8223 12778 10553 9764 9764 6307 6307 6207 6207 11357 11357 0 0 160 160 440 440 646 646 783 783 40 40 41 41 42 42 43 43 44 44 25 25 19 19 12 12 8 8 5 5 17 17 14 14 36 36 51 51 63 63 1766 1766 2006 2006 2188 2188 2314 2314 2484 2484 8430 8430 8693 8693 8899 8899 8991 8991 9351 9351 16383 16383 22300 22300 30906 30906 39893 39893 50096 50096 848 848 826 826 704 704 514 514 279 279 8 3 71 2551 9425 60519 0 .
Earthquake loading in transverse direction (kNm) Forces Pier ID Element 31 32 1 33 34 38 35 36 37 39 Joint Axial Shear Y Shear Z Torsion Moments Bending Bending Y Z 2 9 9 10 10 11 11 3 3 4 198 198 198 198 198 197 197 196 196 194 15 15 15 14 14 13 13 12 12 12 1814 1814 1814 1775 1775 1694 1694 1561 1561 1339 2852 2852 2852 2851 2851 2848 2848 2843 2843 2830 23193 16845 16845 10811 10811 5215 5215 1058 1058 3089 71 18 18 32 32 75 75 118 118 141 5 12 12 13 13 6 6 7 209 209 209 209 209 208 208 208 16 16 16 16 16 15 15 15 3325 3325 3325 3283 3283 3191 3191 3001 3999 3999 3999 3997 3997 3993 3993 3980 33436 21964 21964 10800 10800 1011 1011 6371 61 9 9 48 48 101 101 131 2 .4.Member forces in the Piers .2 .Table A.
4.Earthquake loading in longitudinal direction (kNm) Forces Span ID Element 1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30 Moments Torsion Bending Y Bending Z Joint Axial Shear Y Shear Z 1 14 14 15 15 16 16 17 17 0 253 253 667 667 1009 1009 1281 1281 209 209 209 175 175 141 141 122 122 22 22 22 21 21 21 21 21 21 77 77 77 72 72 58 58 44 44 115 109 109 88 88 82 82 84 84 0 835 835 1545 1545 2001 2001 2342 2342 18 18 19 19 20 20 21 21 22 22 1579 1579 1877 1877 2147 2147 2418 2418 2756 2756 89 89 80 80 117 117 159 159 218 218 20 20 20 20 20 20 19 19 19 19 27 27 17 17 28 28 47 47 62 62 45 45 50 50 53 53 68 68 85 85 2500 2500 2381 2381 1996 1996 1363 1363 608 608 23 24 25 25 26 26 27 27 28 28 2991 405 405 405 753 753 1089 1089 1359 1359 239 143 143 143 73 73 75 75 106 106 19 20 20 20 20 20 20 20 20 20 66 25 25 25 22 22 23 23 31 31 90 105 87 87 93 93 82 82 73 73 0 0 289 289 533 533 490 490 488 488 29 29 30 30 31 31 32 32 33 33 1628 1628 1924 1924 2218 2218 2487 2487 2823 2823 119 119 110 110 75 75 42 42 82 82 20 20 20 20 20 20 20 20 20 20 35 35 33 33 25 25 32 32 47 47 76 76 54 54 58 58 77 77 94 94 620 620 830 830 892 892 814 814 424 424 34 35 36 36 37 37 38 38 39 39 3056 2970 2970 2970 2623 2623 2288 2288 2019 2019 131 294 294 294 247 247 195 195 136 136 20 19 19 19 19 19 20 20 20 20 56 84 84 84 78 78 59 59 36 36 99 96 95 95 105 105 99 99 90 90 0 0 596 596 1600 1600 2308 2308 2842 2842 40 40 41 41 42 42 43 43 44 44 1751 1751 1455 1455 1158 1158 888 888 550 550 87 87 88 88 134 134 170 170 224 224 21 21 21 21 21 21 22 22 23 23 20 20 28 28 59 59 71 71 81 81 92 92 48 48 39 39 51 51 79 79 3083 3083 2988 2988 2570 2570 1905 1905 1036 1036 8 137 259 23 85 115 0 .3 .Table A.Member Forces in the Girder .
Earthquake loading in longitudinal direction (kNm) Pier ID Element 31 32 Joint Axial Moments Bending Bending Shear Y Shear Z Torsion Y Z Forces 2 9 9 10 10 11 11 3 3 238 238 238 237 237 236 236 234 234 3306 3306 3306 3301 3301 3271 3271 3205 3205 6 6 6 6 6 5 5 4 4 2 2 2 2 2 2 2 2 2 29 8 8 15 15 28 28 39 39 51731 40159 40159 28934 28934 17812 17812 6594 6594 1 33 34 38 35 36 4 5 12 12 13 13 6 6 232 232 232 232 231 231 231 231 3069 6692 6692 6692 6682 6682 6633 6633 4 7 7 7 6 6 5 5 2 3 3 3 3 3 3 3 44 25 17 17 29 29 44 44 726 82717 59629 59629 36911 36911 13696 13696 2 37 39 7 229 6507 4 3 52 992 .4 .Table A.4.Member forces in the Piers .
Member Forces in the Girder Earthquake loading in vertical direction (kNm) Forces Span ID Element 1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 2 16 17 18 19 20 21 22 23 24 25 3 26 27 28 29 30 Moments Shear Z Torsion Bending Y Bending Z Joint Axial Shear Y 1 14 14 15 15 16 16 17 17 0 15 15 48 48 47 47 55 55 2638 2638 2638 2301 2301 1730 1730 1076 1076 165 165 165 163 163 161 161 155 155 905 905 905 841 841 714 714 592 592 3951 3293 3293 2598 2598 1911 1911 1260 1260 0 10551 10551 19825 19825 26486 26486 30755 30755 18 18 19 19 20 20 21 21 22 22 76 76 89 89 94 94 93 93 104 104 255 255 651 651 1398 1398 2016 2016 2548 2548 146 146 134 134 121 121 104 104 83 83 369 369 440 440 615 615 764 764 891 891 630 630 229 229 532 532 943 943 1253 1253 31693 31693 29094 29094 23535 23535 15524 15524 5539 5539 23 24 25 25 26 26 27 27 28 28 80 397 397 397 415 415 415 415 421 421 2708 2188 2188 2188 1997 1997 1539 1539 995 995 68 162 162 162 118 118 71 71 28 28 909 618 618 618 586 586 473 473 355 355 1369 1562 1774 1774 2260 2260 2581 2581 2715 2715 0 0 4430 4430 12527 12527 18538 18538 22489 22489 29 29 30 30 31 31 32 32 33 33 424 424 424 424 418 418 414 414 419 419 380 380 397 397 1070 1070 1547 1547 1926 1926 31 31 65 65 113 113 154 154 196 196 211 211 168 168 403 403 496 496 557 557 2659 2659 2389 2389 1900 1900 1263 1263 928 928 23841 23841 22230 22230 17913 17913 11742 11742 4180 4180 34 35 36 36 37 37 38 38 39 39 409 73 73 73 91 91 71 71 67 67 2018 2735 2735 2735 2543 2543 2039 2039 1427 1427 221 44 44 44 55 55 65 65 72 72 553 961 961 961 954 954 840 840 696 696 884 855 777 777 574 574 402 402 573 573 0 0 5538 5538 15772 15772 23732 23732 29391 29391 40 40 41 41 42 42 43 43 44 44 60 60 54 54 43 43 33 33 39 39 672 672 249 249 1088 1088 1750 1750 2327 2327 78 78 83 83 92 92 98 98 100 100 514 514 412 412 614 614 735 735 853 853 824 824 1120 1120 1457 1457 1833 1833 2259 2259 32048 32048 31133 31133 26757 26757 19786 19786 10713 10713 8 0 2664 99 911 2695 0 .5 .Table A.4.
Earthquake loading in vertical direction (kNm) Pier ID Element 31 32 Joint Axial Moments Bending Bending Shear Y Shear Z Torsion Y Z Forces 2 9 9 10 10 11 11 3 3 4809 4809 4809 4811 4811 4805 4805 4784 4784 352 352 352 349 349 340 340 329 329 201 201 201 199 199 191 191 181 181 153 153 153 152 152 152 152 151 151 1696 1011 1011 504 504 564 564 998 998 1799 566 566 811 811 1909 1909 2999 2999 1 33 34 38 35 36 4 5 12 12 13 13 6 6 4747 4711 4711 4711 4623 4623 4578 4578 323 422 422 422 419 419 411 411 164 295 295 295 293 293 287 287 150 224 224 224 224 224 223 223 1314 2084 1066 1066 653 653 942 942 3610 1654 375 375 1419 1419 2667 2667 2 37 39 7 4567 405 275 221 1487 3475 .Table A.4.Member forces in the Piers .6 .
0 0.6 30.2 21.0 0.7 .Table A.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 7.0 0.Earthquake loading in transverse direction Joint 1 8 18 4 29 7 40 Translations (mm) X Y Z 0.6 Rotations (rad) X Y Z 0.0 0.0 0.0 0.0 0.5 0.4.0 0.0 0.0 9.0 0.0 .0 0.1 0.0 0.0 0.0 0.Displacements .0 0.0 0.0 0.1 23.1 0.0 0.0 0.1 0.0 0.5 0.
0 Rotations (rad) X Y Z 0.0 0.0 0.0 0.0 0.0 0.5 72.0 0.4 0.0 0.0 0.Table A.0 73.0 0.0 0.0 72.0 0.8 .3 73.0 0.0 0.4 0.5 0.Earthquake loading in longitudinal direction Joint 1 8 18 4 29 7 40 Translations (mm) X Y Z 73.0 0.9 72.0 0.4.9 0.5 73.0 0.0 0.0 0.0 0.0 0.0 .8 0.0 0.Displacements .0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.
2 0.Table A.2 Rotations (rad) X Y Z 0.6 0.5 0.0 0.0 0.2 0.Earthquake loading in vertical direction Joint 0 8 18 4 29 7 40 Translations (mm) X Y Z 0.0 0.0 0.8 0.2 0.3 0.0 9.0 0.0 0.0 0.2 0.4 9.0 0.0 0.0 0.0 0.Displacements .0 0.2 0.0 0.0 0.0 0.0 0.0 0.3 0.2 0.0 0.0 0.0 0.2 0.0 0.0 0.9 .5 0.2 0.0 .5 6.4.0 0.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.