Soil Dynamics and Earthquake Engineering 31 (2011) 648661

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

Example of application of response spectrum analysis for seismically isolated

curved bridges including soil-foundation effects
Sevket Ates a,n, Michael C. Constantinou b
a
b

Department of Civil Engineering, Karadeniz Technical University, 61080 Trabzon, Turkey

Department of Civil, Structural and Environmental Engineering, University at Buffalo, The State University of New York, 212 Ketter Hall, Buffalo, NY 14260, USA

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 3 October 2010
Received in revised form
26 November 2010
Accepted 1 December 2010
Available online 17 December 2010

This paper presents seismic behaviour of isolated curved bridges in the earthquake prone regions.
For the seismic isolation of bridges, double concave friction pendulum bearings are placed between
the deck and the piers, and the abutments as isolation devices. A curved bridge is selected to exhibit the
application for seismic isolation. The mentioned bridge is a three-span featuring cast-in-place concrete
box girder superstructure supported on reinforced concrete columns found on drilled shafts and on
integral abutments founded on steel pipe piles. Additionally, the bridge is located on site underlain by a
deep deposit of cohesionless material. The drilled shaft-soil system is modelled by equivalent soil springs
method and is included in the nite element model. The soil modelled as a series of springs is connected to
the drilled shaft at even intervals.
The multi mode method of analysis is typically implemented in a computer program capable of
performing response spectrum analysis. The response spectrum specied for the analysis is the 5%damped spectrum modied for the effects of the higher damping. Each isolator is represented by its
effective horizontal stiffness with a linear link element.
As seen from the results of the outlined analysis, usage of the isolation devices offers some advantages
for the internal forces on the deck for the considered curved bridge as per the non-isolated curved bridge.
The response spectrum analysis is substantially required to make a decision of the displacement
capacity of the double concave friction pendulum bearings used in the study.
& 2010 Elsevier Ltd. All rights reserved.

1. Introduction
For bridge applications, contemporary seismic isolation systems
provide horizontal isolation from the effects of earthquake shaking
to reduce forces. The main function of the seismic isolation system
is to increase the period of vibration by increasing the lateral
exibility in the bridges or other structures.
The double concave friction pendulum (DCFP) bearing is an innovative and viable isolation system that is becoming a widespread application for the earthquake protection of structures. The DCFP bearings
consist of two spherical stainless steel surfaces and an articulated slider
covered by a Teon-based high bearing capacity composite material.
The concave surfaces may have the same radii of curvature. Also, the
coefcient of friction on the two concave surfaces may be the same or
not. Hyakuda et al. [1] presented the response of a seismically isolated
building in Japan where DCFP bearings are utilized. Experimental and
analytical results on the behaviour of a system having concave surfaces
of both equal and unequal radii and both equal and unequal coefcient
of friction at the upper and lower sliding surfaces were presented by

Corresponding author.
E-mail address: sates@ktu.edu.tr (S. Ates).

0267-7261/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.soildyn.2010.12.002

Tsai et al. [2]. Constantinou [3] and Fenz and Constantinou [46]
described the principles of operation of the DCFP bearing and
presented the development of the forcedisplacement relationship
based on equilibrium. The theoretical forcedisplacement relationship
was veried through characterization testing of bearings with sliding
surfaces having the same and then different radii of curvature and
coefcients of friction. Finally, practical considerations for analysis and
design of DCFP bearings were presented.
Few researchers have dealt with the dynamic response of
straight and curved box girder bridges. Sennah and Kennedy [7]
highlighted the most important references related to development of current guide specications for the design of straight and
curved box-girder bridges. DeSantiago et al. [8] analyzed a series of
horizontally curved bridges using simple nite element models and
reported that the bending moment in girders of a curved bridge can
be about 23.5% higher than moments in girders of a straight bridge
of similar span and design conguration. Mwafy and Elnashai [9]
carried out a detailed seismic performance assessment of a multispan curved bridge including soilstructure interaction effects.
Constantinou et al. [10] manifested analysis and design procedures for
seismically isolated bridges and examples of analysis and design of
seismic isolation systems. Ates and Constantinou [11] carried out a
parametrical study associated with the effects of the earthquake

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

ground motions on the seismic response of isolated curved bridges

including soilstructure interaction.
Soilstructure interaction (SSI) effect on the seismically isolated
bridges has been also studied by many researchers. Tongaonkar and
Jangid [12] observed that the soil surrounding the pier has signicant
effects on the response of the isolated bridges and under certain
circumstances the bearing displacements at abutment locations may
be underestimated if the SSI effects are not considered in the response
analysis of the system. Cases in which SSI needs to be incorporated in
seismically isolated bridge design are identied and ways to take
advantage of SSI in order to enhance safety level and reduce design
costs are recommended by Spyrakos and Vlassis [13]. Ucak and
Tsopelas [14] found that the results from comprehensive numerical
analyses show that soilstructure interaction causes higher isolation
system drifts as well as, in many cases, higher pier shears when
compared to the bridges without SSI. In the light of the studies, SSI can
have both benecial and detrimental effects on the response of the
isolated bridges depending on the characteristics of the ground motion.
The main goal of this study is to set forth the dynamic response
of isolated curved bridges subjected to response spectrum. The
soilstructure interaction is also taken into account by springs
representing the soil beneath footing and drilled shaft. Displacement capacities of the DCFP bearings are also evaluated.

2. Double concave friction pendulum bearings (DCFP)

The double concave friction pendulum bearings are made of two
concave surfaces, which are called upper and lower, and is shown
in Fig. 1.
The concave surfaces may have the same radii of curvature. Also,
the coefcient of friction on the two concave surfaces may be the
same or not. The maximum displacement capacity of the bearing is
2d, where d is the maximum displacement capacity of a single
concave surface. Note that due to rigid body and relative rotation of
the slider, the displacement capacity is actually slightly different
from 2d. The forcedisplacement relationship for the DCFP bearing
is given by the following equation:




Ff 1 R1 h1 Ff 2 R2 h2
W
F
Ub
1
R1 h1 R2 h2
R1 h1 R2 h2
where W is the vertical load, R1 and R2 are radii of the two concave
surfaces, h1 and h2 are the part heights of the articulated slider and
Ub is the total displacement (bearing displacement), and the sum of
the displacements on the upper and lower surfaces are given by
Ub 2d Ub1 Ub2

herein Ub1 and Ub2 are the displacements of the slider on the upper
and lower concave surface, respectively, and the individual displacements on each sliding surfaces are


FFf 1
R1 h1
Ub1
3
W

R2
The upper concave surface

h1
h2
2

The lower concave surface

d

ds

R1

Fig. 1. Double concave friction pendulum (DCFP) bearings.

Ub2


FFf 2
R2 h2
W

649

In Eqs. (3) and (4), Ff1 and Ff2 are the friction forces on the
concave surfaces 1 and 2, respectively. The forces are given by
Ff 1 m1 W sgnU_ b1

Ff 2 m2 W sgnU_ b2

where m1 and m2 are the coefcient of friction on the concave surfaces

1 and 2, respectively; U_ b1 and U_ b2 are sliding velocities at the upper
and lower surfaces, respectively; and sgn(U) denotes the signum
function. Most applications of the DCFP bearings will likely utilize
concave surfaces of equal radii, namely, R1 R2. In this study, each
radius is calculated as 88 in. Parts heights of the articulated slider h1
and h2 are nearly equal in most cases. Thus, the effective coefcient of
friction is equal to the average of m1 and m2, and is given by

me

m1 R1 h1 m2 R2 h2
R1 R2 h1 h2

In Eq. (1), the rst term is the stiffness of the pendulum

component (spring forces) and the second term is the stiffness of
the friction component. The natural period of vibration is given by
the following equation:
s
s
R1 R2 h1 h2
Re
8
T 2p
2p
g
g
where g is the acceleration of gravity; and Re is the effective radius
of curvatures. Eq. (8) shows that the natural period of vibration is
independent of mass, but it is controlled by the selection of the
radius of the spherical concave surfaces. The important parameter
is employed as Re R1 + R2 h1  h2 4.27 m. It is also shown in
Eq. (8) that the stiffness of the pendulum depends on the weight
carried by bearing. The coefcient of the friction of the two concave
surfaces depends on the bearing pressure and is given by

m1,2 fmax fmax fmin ea9v_ b 9

where fmax and fmin are the maximum and minimum mobilized
coefcients of friction, respectively; and a is a parameter that
controls the variation of the coefcient with the velocity of sliding.
Analysis of seismically isolated bridges will be performed for
each seismic loading case considered (design basis or maximum
considered earthquake) for two distinct sets of mechanical properties of the isolation system.
Lower bound properties are dened to be the lower bound values
of characteristic strength and post-elastic stiffness that can occur
during the lifetime of the isolators. Typically, the lower bound values
describe the behaviour of fresh bearings, at normal temperature and
following the initial cycle of high speed motion. The lower bound
values of properties usually result in the largest displacement demand
on the isolators. Upper bound properties are dened to be the upper
bound values of characteristic strength and post-elastic stiffness that
can occur during the lifetime of the isolators and considering the
effects of aging, contamination, temperature and history of loading
and movement. Typically, the upper bound values describe the
behaviour of aged and contaminated bearings, following the movement that is characteristic of substantial trafc loading, when
temperature is low and during the rst high speed cycle of seismic
motion. The upper bound values of properties usually result in the
largest force demand on the substructure elements.
The lower and upper bound values of mechanical properties are
determined from nominal values of properties and the use of
system property modication factors. The nominal properties are
obtained either from testing of prototype bearings identical to the
actual bearings or from test data of similar bearings from previous
projects and the use of appropriate assumptions to account for

650

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

uncertainty. Typically, the analysis and design of the isolated

bridge is based on the available data from past tests of similar
bearings. The assumptions made for the range of mechanical
properties of the isolators are then conrmed in the prototype
testing that follows. If the selection of the range of mechanical
properties is properly made, the prototype bearing testing will
conrm the validity of the assumptions and therefore the validity of
the analysis and design. Accordingly, modications of the design
would not be necessary. Such modications often lead to delays
and additional costs.

3. Response spectrum method of analysis

The method is based on representing the behaviour of isolators
by linear elastic elements with stiffness equal to the effective or
secant stiffness of the element at the actual displacement. The
effect of energy dissipation of the isolation system is accounted for
by representing the isolators with equivalent linear viscous elements on the basis of the energy dissipated per cycle at the actual
displacement. The response is then calculated by use of response
spectra that are modied for the effect of damping larger than 5% of
the critical ones. Given that the actual displacement is unknown
until the analysis is performed, the method requires some iteration
until the assumed and calculated values of isolator displacement
are equal.
The 5%-damped elastic response spectrum represents the usual
seismic loading specication. Spectra for higher damping need to
be constructed for the application of multi mode method. Elastic
spectra constructed for higher viscous damping are useful in the
analysis of linear elastic structures with linear viscous damping
systems. Moreover, they are used in the simplied analysis of
yielding structures or structures exhibiting hysteretic behaviour
since simplied methods of analysis are based on the premise that
these structures may be analyzed using equivalent linear and
viscous representations. The typical approach of constructing an
elastic spectrum for damping greater than 5% is to divide the 5%damped spectral acceleration by a damping coefcient or damping
reduction factor B:
Sa T, b

Sa T,5%
B

10

where Sa(T,b) is the spectral acceleration at period T for damping

ratio b. Note that the spectral acceleration is the acceleration at
maximum displacement and is not necessarily the maximum.
Therefore, it is related directly to the spectral displacement Sd
through
 2
T
Sa
11
Sd
2p
The damping reduction factor B is a function of the damping
ratio and may be a function of the period. Eq. (10) is typically used
to obtain values of coefcient B for a range of values of period T and
for selected earthquake motions. The results for the selected
earthquake motions are statistically processed to obtain average
or median values, which upon division of the value for 5% damping
to the value for damping b results in the corresponding value of B.
The results are affected by the selection of the earthquake motions
and the procedures used to scale the motions in order to represent a
particular smooth response spectrum. Furthermore, the values of
factor B used in codes and specications are typically on the
conservative side, rounded and based on simplied expressions.
The values of factor B recommended by FEMA 440 [15] and
Eurocode 8 [16] are given as follows, respectively:
B

4
5:6ln100b

and
B

r
0:05 b
0:10

13

The values of factor B in various codes and specications are

nearly identical for values of damping ratio less than or equal to
30%. This is the limit of damping ratio for which simplied methods
of analysis can be used.
Consider a seismically isolated structure represented as a single
degree of freedom system with mass m, weight W and lateral force
displacement relation having bilinear hysteretic characteristics as
shown in Fig. 2. The system is characterized by characteristic
strength Qd and post-elastic stiffness Kd. For the friction pendulum
system, the characteristic properties are dened as below, correspondingly:
Qd mW
Kd

14

W
Re

15

where m is the coefcient of friction at large velocity of sliding. The

displacement of the system for an earthquake, which is described
by a particular smooth response spectrum, can be identied as D.
The effective period of the system is given by [17,18]
s
W
Teff 2p
16
Keff g
Keff Kd

Qd
W mW

Re
D
D

17

Substituting Eqs. (14) and (15)) into Eq. (16), the effective period
of the system is rewritten as follows:
s
1
Teff 2p
18
g=Re mg=D
In addition, the effective damping of the same system is given by
[17,18]


1
Ed
beff
19
2p Keff D2
where Ed is the energy dissipated per cycle at displacement D and
period Teff. For the behaviour depicted in Fig. 2, the energy dissipated
per cycle is given by
Ed 4Qd DY

20

where Y is the yield displacement of the system. Assuming Y is equal

to zero, Eqs. (14), (16) and (17) are substituted into Eq. (19) and the
effective damping of the system is easily obtained and then given in

Lateral force

Qd

Kd
Lateral
displacement

12
Fig. 2. Idealized forcedisplacement relation of typical seismic isolation system.

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

more useful form as



2
m
beff
p m D=Re

21

It should be noted that Eq. (21) is valid when Y is zero as mentioned

above. The peak dynamic response of this system may be obtained
from the response spectrum assuming that the system is linear elastic
with effective period Teff. Based on the value of effective damping beff,
the damping reduction factor B is calculated. The response of the
system depending on spectral displacement and spectral acceleration
is calculated as the response obtained for 5% damping divided by
factor B. However, since the calculation is based on an assumed value
of displacement D, the process is repeated until the assumed and
calculated values of displacement are equal. This procedure represents a simplied method of analysis that is typically used for seismically isolated structures.






4. Description of example curved bridge

A curved bridge is selected to exhibit the application for seismic
isolation. The bridge was used as an example of bridge design
without an isolation system in the Federal Highway Administration
(FHWA) Seismic Design Course, Design Example No. 6, prepared by
Berger/Abam Engineers Inc. [19]. The bridge is to be modelled and
analyzed in a seismic zone with an acceleration coefcient of 0.7g
dened by Caltrans [20].
The conguration of the bridge is a three-span featuring cast-inplace concrete box girder superstructure supported on reinforced
concrete columns found on drilled shafts and on integral abutments founded on steel pipe piles. The bridge is located on site
underlain by a deep deposit of cohesionless material.
The alignment of roadway over the bridge is sharply curved,
horizontally (1041), but there is no vertical curve. The two intermediate bents consist of rectangular columns with a cross beam on
top. The geometry of the bridge, section properties and foundation
properties are assumed to be the same as in the original bridge in
the FHWA example. It is presumed that the original bridge design is
sufcient to sustain the loads and displacement demands when
seismically isolated as described herein. The bridge is only used for
comparing purpose. The following assumptions are also made for
earthquake analyses of the bridges under consideration:

assumption as the base isolation attempts to reduce the earthquake response in such a way that the structure remains within
the elastic range.
The deck of the bridge is curved and is supported at discrete
locations along its longitudinal axis by cross diaphragms.
Both superstructure and substructure are modelled as lumped
mass systems divided into the number of small discrete
segments. Each adjacent segment is connected by a node and
at each node three degrees of freedom are considered. The mass
of each segment is assumed to be distributed between the two
adjacent nodes in the form of point masses.
Stiffness contributions of non-structural elements such as
sidewalk and parapet are neglected.
The force-deformation behaviour of the bearing is considered to
be linear.
The bearings provided at the piers and abutments have the same
dynamic characteristics.
The drilled shaft is represented for all motions using a spring
model with frequency-independent coefcients. The modelling
of the drilled shaft on deformable soil is performed in the same
way as that of the structure and is coupled to perform a dynamic
SSI analysis.

Figs. 37 show, respectively, the plan and its dimensions,

developed elevation, framing plan, horizontal sections of the
substructure, section of the superstructure and section at the
center line of the pier as an intermediate bent. The bridge is
isolated with two isolators at each abutment and pier location for a
total of 8 isolators. The isolators are directly located above the cap
of the rectangular columns and the abutments. Two isolators are
intentionally used instead of more isolators due to the fact that the
distribution of load on each isolator is accurately calculated.
Additionally, the use of more than two isolators per a location
would have caused difculties in the calculation of the axial and
increased the cost.
Diaphragms in the box girder at the abutment and pier locations
above the isolators are also taken into account, in view of rigidity
and self weight in the nite element model of the curved bridge.

 Bridge superstructure and piers are assumed to remain in the

elastic state during the earthquake excitation. This is a reasonable

Fig. 4. Developed elevation of the curved bridge.

88.00m

.25

Pier 1

11.80m

33.50m

27

651

27

.25

Pier 2

Rc =48.77m
The center line
of Abutment A

The center line

of Abutment B

All arch lengths are as per along the center line of the bridge.
Fig. 3. The curved bridge plan and its dimensions.

The center line

of the bridge

652

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

88.00m
33.50m
Y
m
.25
27

27

.25

90cm

23cm

Radial
Direction

76cm

Chord
Direction
Fig. 5. Framing plan of the curved bridge.

The center line of the

bridge

100cm

50cm 50cm

Column section:

Drilled shaft section:

The center line of the

pier1 or 2
240cm
85cm

85cm

170cm

Deck section:
1180cm

100cm

30cm

290cm

30cm 280cm

30cm 290cm

30cm

100cm

170cm

25cm

18cm
Fig. 6. Horizontal sections of the substructure and the deck of the curved bridge.

5. Modelling of the drilled shaft for piers

The drilled shaft can be modelled by equivalent soil springs
method that is illustrated in Fig. 8. With the use of this technique,
the drilled shaft is included in the nite element model, and the
foundation soil is modelled as a series of springs connected to the
drilled shaft at even intervals. It should be noted that spring
stiffness must be accurately selected to represent the best behaviour. The soil springs at each depth are calculated using a
coefcient of horizontal subgrade reaction that increases linearly
with depth and is inversely proportional to the cross sectional
dimension of the drilled shaft [21].
A sufcient number of springs should be used along the length
of the drilled shaft. The springs near the surface are usually the
most important to characterize the response of the drilled shaft
surrounded by the soil; thus a closer spacing may be used in that
region. However, in general springs evenly spaced at about half the
diameter of the drilled shaft are recommended [19].
In this study, the diameter of the drilled shaft is 2.40 m and the
drilled shaft is 18 m in length. Consequently, 15 springs are used on
the center along the length of the drilled shaft. The springs are
arranged at 1.20 m intervals but at the ends the segments are taken
as 0.60 m in such a way that all the spring constants are calculated

based on 1.20 m tributary length of the drilled shaft. The mentioned arrangement is implemented such as in Fig. 8. The soil
properties beneath the foundation of the bridge are given in Table 1
[19]. In this table, g is the total unit weight; f is the internal angle of
friction; c is cohesion and nh is the constant of horizontal subgrade
reaction. New ll will be required at the abutments. The ll has
similar properties to the native soil.
In order to calculate the horizontal stiffness of the equivalent soil
springs, the coefcient of horizontal subgrade is given below [19]:
kh

nh z
D

22

in which z is the depth in reference to the ground surface; D is the

diameter of the drilled shaft. The horizontal stiffness of the equivalent
soil springs based on the coefcient of horizontal subgrade is
ki kh DHtrib

23

where Htrib represents the height of tributary soil spring. Substituting

Eq. (22) into Eq. (23), the horizontal stiffness of the equivalent soil
springs can be rewritten as follows:
ki nh zHtrib

24

The horizontal stiffness of the equivalent soil springs is tabulated in Table 2 where Htrib is 1.20 m as per Eq. (24).

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

653

10% Slope
Rigid beam
element

640cm

30cm

Ground surface
60cm

180cm

25cm

ki

ki
14@120cm=1680cm

1800cm

Medium dense, silty

sand

1800cm

Ground Surface

240cm
60cm
Bedrock

kv =

Z
Y

kr = 0
Fig. 7. Section at the center line of the pier of the curved bridge.

As shown in Fig. 8, vertical movement of drilled shaft is

restrained by an innitely stiff spring at the base of the shaft.
Actual vertical resistance occurs via skin friction and end bearing.
However, for this analysis, the simplication of restraining only the
base of the drilled shaft is felt to be reasonable. Similarly, torsional
movement of the drilled shaft would be resisted by skin friction.
However, no torsional restraint was used in the mathematical
model. The response is not sensitive to the lack of torsional restraint
in the drilled shaft, and this can be demonstrated by simple
bounding analyses [19].

6. Modelling of the foundation stiffness for abutments

The abutments are modelled with three dimensional frame
elements. The abutment is depicted in Fig. 9a. The springs represent
the piles as shown in Fig. 9b. The stiffness of the springs is given in
Table 3 for the vertical, longitudinal, transverse and rotational
directions as per each connection point between the abutment and
the piles, respectively, taking advantage of the design example [19]
for calculating the stiffness.

7. Selection and usage of the spectrum

The response spectrum specied for the analysis is the 5%damped spectrum modied for the effects of the higher damping.
The ordinates of the 5%-damped response spectrum for values of
period larger than 0.8Teff are divided by the damping reduction

Fig. 8. Equivalent soil spring model of the drilled shaft and its geometry and
element layout.

Table 1
Soil properties for the subsurface materials.
Stratum

Depth (m)

Soil description

g (kN/m3) f (deg.) nh (kN/m3)

Alluvium

0 to 4100

19

34

4000

New Fill

Above grade

Medium dense,
silty sand
Medium dense
sand and gravel

19

34

6250

Table 2
The horizontal stiffness of the equivalent soil springs of the drilled shaft.
Depth z (m)

Spring stiffness
ki (kN/m)

0.60
1.80
3.00
4.20
5.40
6.60
7.80
9.00
10.20
11.40
12.60
13.80
15.00
16.20
17.40
18.00

3021
9077
15,133
21,189
27,231
33,287
39,343
45,385
51,441
57,497
63,553
69,595
75,651
81,707
87,749
Rigid

654

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

CL
End diaphragm
100cm
100cm
75cm

50cm

180cm

180cm

180cm

180cm

180cm

540cm

50cm

180cm

Pipe piles filled

with
concrete

540cm
1180cm

klong

ktrans

kvert
krl

krt=0
krv

Fig. 9. (a) The real abutment and (b) the analytical model of the abutment having the equivalent springs representing the soil.

Table 3
Stiffness of the equivalent soil springs standing for the pipe piles.

1 and 7
klong (kN/m)
Longitudinal translation
ktrans (kN/m)
Transverse translation
kver (kN/m)
Vertical translation
krv (kNm/rad)
Rotation about vertical axis
krl (kNm/rad)
Rotation about longitudinal axis
krt (kNm/rad)
Rotation about transverse axis

Spring location name

2 and 6

2.00
3 and 5

5356

5356

5356

5356

4320

4320

4320

4320

608,083

608,083

608,083

608,083

152,360

67,720

16,933

17,298,935

7,688,420

1,922,320

Note: krv kver d2i krl klong d2i krt ktran d2i where di is the distance between the
center lines of the abutment and the ith pipe pile. In this study, the distance is 180,
2  180 and 3  180 cm ranging from the innermost of the pipe pile to the outermost
one, respectively.

factor B for the effective damping of the isolated bridge. Whereas all
modes are assumed to be damped at 5%, in this approach, only the
isolated modes of the structure are allowed the reduction of
response due to increased damping. Note that the modication
of the spectrum for higher damping requires that the effective
period and effective damping in each principal direction should be
calculated. This is done using the single degrees of freedom system
of analysis. Fig. 10 presents the response spectrum used in multi
mode analysis of a seismically isolated bridge supported on the soil

Spectral Acceleration (g)

Stiffness

Sa (T,5%)

Sa (T,5%)

1.50
0.8Teff
1.00
0.50
0.00
0.00

2.11sec
0.50

1.00

1.50

2.00
2.50
Period (sec)

3.00

3.50

4.00

Fig. 10. Acceleration response spectrum curve for soil prole type C.

prole type C [20]. For the soil type, the data acceleration spectrum
is dedicated in Table 4. The effective period is Teff 2.636 s, the
effective damping is beff 0.30 and the damping reduction factor
B1.82. The ordinates of the 5%-damped spectrum for period
larger than 2.11 s are divided by factor 1.82.
Analysis by the multi mode method should be independently
performed in two orthogonal directions and the results be combined
using the 10030% combination rule. The two orthogonal directions
may be any two arbitrary perpendicular directions that facilitate the

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

analysis. The most convenient is the use of the longitudinal and

transverse bridge directions. For curved bridges, the longitudinal axis
may be taken as the chord connecting the two abutments. The vertical
ground acceleration effect can be included, using rational methods of
analysis and combined using the 1003030% rule. The interested
reader is referred to Wilson et al. [22]. The procedure is employed in
this study.

655

elements are used representing columns, cap beam and drilled

shafts, and the springs representing soil stiffness. Additionally, the
DCFP bearings on the abutments and the cap beams of the piers are
dened using equivalent beam element as well. The super-elevation having a slope of 10% is also taken into account in the nite
element model in order that torsional stresses of the girders are not
overlooked. The cross sectional properties of the bridge are given in
Figs. 6 and 7. Additionally, the modulus of elasticity of concrete is
taken as 3.2  107 kN/m2.

8. Finite element model of the curved bridge

The nite elemet model of the curved bridge consists of solid
elemets having 3 degrees of freedom at each nodal point. They are
in horizontal and vertical translational directions. In this model, the
drilled shaft is also included by equivalent soil springs mentioned
above in such a way that the foundation soil is considered as a series
of springs connected to the drilled shaft at even intervals. The three
dimensional nite element models are represented with the
diagram in Fig. 11. The model generated in SAP2000 [23] has
9031 nodal points and 5060 solid area elements. In addition frame
Table 4
Spectral acceleration used in response spectrum analysis.
Period (s)

0.05
0.05
0.10
0.24
0.30
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.10
2.11
2.25
2.50
2.60
2.68
2.72
2.75
3.00
3.25
3.50
3.75
4.00

Spectral acceleration (g)

Original

Vertical

Modied for lower

bound properties

Modied for upper

bound properties

0.70
0.70
1.29
1.77
1.80
1.72
1.44
1.19
0.95
0.78
0.64
0.55
0.51
0.51
0.48
0.41
0.40
0.39
0.38
0.37
0.32
0.29
0.25
0.23
0.20

0.49
0.49
0.90
1.24
1.26
1.20
1.01
0.83
0.67
0.55
0.45
0.39
0.36
0.36
0.34
0.29
0.28
0.27
0.27
0.26
0.22
0.20
0.18
0.16
0.14

0.70
0.70
1.29
1.77
1.80
1.72
1.44
1.19
0.95
0.78
0.64
0.55
0.51
0.51
0.48
0.41
0.40
0.39
0.24
0.23
0.19
0.18
0.15
0.14
0.12

0.70
0.70
1.29
1.77
1.80
1.72
1.44
1.19
0.95
0.78
0.64
0.55
0.51
0.28
0.27
0.23
0.22
0.22
0.22
0.21
0.18
0.16
0.14
0.13
0.11

9. Modelling of double concave friction pendulum for response

spectrum analysis
Each isolator may be modelled as a vertical three dimensional
beam element rigidly connected at its two ends of length h, area A,
moment of inertia about both bending axes I and torsional constant
J. The element length is the height of the bearing, h0.30 m and the
area is the contact area, which is a circle of 0.25 m diameters. Note
that the element is intentionally used with rigid connections at its
two ends so that P  D effects can be properly accounted for in the
case of the double concave bearing.
To properly represent the axial stiffness of the bearing, the
modulus of elasticity is specied to be related but less than the
modulus of steel, so it can be taken as 1.05  107 kN/m2. The
bearing is not exactly a solid piece of metal so that the modulus is
reduced to half to approximate the actual situation. Torsional
constant is set J0 or a number near zero since the bearing has
insignicant torsional resistance. Moreover, shear deformations in
the element are de-activated by specifying very large areas in shear.
The moment of inertia of each element is calculated by the
following equation:
I

Keff h3
12E

25

where Keff is the effective stiffness of the bearing. The required

values outlined above for the linear link element for the DCFP
bearings are demonstrated in Table 5.
Response spectrum analysis is performed using the response
spectrum of Fig. 10 for 0.7g after division by parameter B for periods
larger or equal to 0.8TM, where TM is the effective period and B is the
parameter that relates the 5%-damped spectrum to the spectrum at the
effective damping. Values of 0.8TM are 2.72 s for lower bound analysis
and 2.11 s for upper bound analysis. Values of spectral acceleration
used in the analysis are presented in Table 4. The modied values of the
acceleration per the above rule are in grey colour in Table 4.
Eigen modal and multi mode response spectrum analyses were
performed in SAP2000 [23]. Figs. 12 and 13 offer the mode shapes of
the rst three modes of vibration of the isolated bridge in the lower

Table 5
The required values of friction pendulum for the response spectrum analysis.

Fig. 11. Three-dimensional nite element model of the curved bridge.

Bearing
location

Parameter Lower bound

properties

Upper bound
properties

Abutment

Keff (kN/m)
h (m)
E (kN/m2)
A (m)
I (m4)

388.76
0.30
1.05  108
0.051
9.177  10-9

840.57
0.30
1.05  108
0.051
4.374  10-9

Pier

Keff (kN/m)
h (m)
E (kN/m2)
A (m2)
I (m4)

781.03
0.30
1.05  108
0.051
18.436  10-9

1190.80
0.30
1.05  108
0.051
28.109  10-9

656

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

and upper bound analyses, respectively. Analyses are performed by

separately applying the earthquake excitation in the chord, radial
and vertical bridge directions as dened in Fig. 5. The vertical
response spectrum is taken as a 70% portion of the horizontal 5%damped spectrum without any modication for increased damping
and then is given in Table 4.

10. Numerical results for the DCFP bearing

The results of these analyses are presented in Tables 6 and 7 in
terms of the bearing displacements, isolation shear force and bearing
axial forces due to earthquake. Observations to be made in the results
of multi mode response spectrum analyses are that the displacements

Mode 1: T1 = 3.400 sec

Mode 2: T2 = 3.305 sec

Mode 3: T3 = 2.951 sec

Fig. 12. The rst three modes of vibration of the isolated curved bridge with lower bound properties of the DCFP.

Mode 1: T1 = 2.636 sec

Mode 2: T2 = 2.530 sec

Mode 3: T3 = 2.087 sec

Fig. 13. The rst three modes of vibration of the isolated curved bridge with upper bound properties of the DCFP.

Table 6
Signicant response quantities for the abutments obtained by multi mode response spectrum analysis of the isolated bridge with the DCFP.
Response name

Earthquake direction
Chord (100%)

DM (cm)
DTM (cm)

Bearing shear force (kN)

Bearing axial force (kN)
Bearing total axial force (kN)

Properties type of
the DCFPs
Radial (100%)

Vertical (100%)

37.30
65.50
41.12
47.85
q
65:502 0:30  37:302 0:30  0:432 66
q
47:852 0:30  41:122 0:30  0:432 50

0.46
0.43

159.86
186.02
313.85
545.73
153.41
865.18
180,59
1081.31
q
1,229:782 0:30  865:182 0:30  153:412 1260
q
1,229:162 0:30  1,081:312 0:30  180:592 1270

1.78
3.87
1229.78
1229.16

Lower bound
Upper bound
Lower bound
Upper bound
Lower
Upper
Lower
Upper
Lower

bound
bound
bound
bound
bound

Upper bound

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

of the abutment and pier bearings are different due to the pier
exibility effect and the inertia effects in the substructure. Accordingly, the abutment bearings experience lesser displacement. Maximum axial bearing forces develop when the earthquake excitation is
in the vertical direction only. Maximum bearing displacement occurs
when the bridge is under the radial directional earthquake. This result
for earthquake is nearly the same as the expected value as 70 cm. This
is in good agreement in view of selection of the displacement capacity

657

of the DCFP bearings recommended by Constantinou et al. [24]. On the

other hand, maximum shear forces at the isolated level are formed in
case of the bridge isolated by the DCFP bearings is having the upper
bound properties.
Tables 6 and 7 show a comparison of signicant response
quantities obtained by the multi mode response spectrum methods
of analysis. In each type of the analyses, quantity DTM is calculated as
the vector sum of bearing displacements due to chord, radial and

Table 7
Signicant response quantities for the piers obtained by multi mode response spectrum analysis of the isolated bridge with the DCFP.
Response name

Earthquake direction
Chord (100%)

DM (cm)
DTM (cm)

Bearing shear force (kN)

Bearing axial force (kN)
Bearing total axial force (kN)

Properties type of
the DCFPs
Radial (100%)

Vertical (100%)

47.00
64.50
42.37
47.55
q
64:502 0:30  47:002 0:30  2:032 66
q
47:552 0:30  42:372 0:30  0:202 49

2.03
0.20

Lower bound
Upper bound
Lower bound

327.60
345.03
408.82
699.71
278.98
612.53
297.22
849.92
q
2227:112 0:30  612:532 0:30  278:982 2236
q
2226:622 0:30  849:922 0:30  297:222 2243

3.02
6.85
2227.11
2226.62

Lower
Upper
Lower
Upper
Lower

Upper bound

8000
12000
16000
30
60
Bridge Length (m)

2000
4000
Lower Bound Properties
Upper Bound Properties

8000
0

30
60
Bridge Length (m)

4000
8000
Lower Bound Properties
Upper Bound Properties

16000
0

30
60
Bridge Length (m)

8000
12000
16000
0

30

60

90

0
2000
4000
6000
8000

90

12000

4000

Bridge Length (m)

6000

90

Bending Moment (kNm)

Bending Moment (kNm)

Bending Moment (kNm)

Bending Moment (kNm)

Lower Bound Properties

Upper Bound Properties

4000

Upper bound

Non-isolated Bridge

Bending Moment (kNm)

Bending Moment (kNm)

Isolated Bridge
0

bound
bound
bound
bound
bound

90

60
30
Bridge Length (m)

90

30
60
Bridge Length (m)

90

0
4000
8000
12000
16000

Fig. 14. (a) Bending moment about the center line axis of the isolated and non-isolated curved bridges subjected to vertical directional earthquake. (b) Bending moment about
the center line axis of the isolated and non-isolated curved bridges subjected to chord directional earthquake. (c) Bending moment about the center line axis of the isolated and
non-isolated curved bridges subjected to radial directional earthquake.

658

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

Torsional Moment (kNm)

1000
2000
3000
4000

Lower Bound Properties

Upper Bound Properties

5000
0

Torsional Moment (kNm)

Non-isolated Bridge

30
60
Bridge Length (m)

0
1000
2000
3000
4000

Lower Bound Properties

Upper Bound Properties

5000
0

30

0
1000
2000
3000
4000
5000

90

Torsional Moment (kNm)

Torsional Moment (kNm)

Isolated Bridge
0

60

2000
3000
4000
5000

90

8000
12000
Lower Bound Properties
Upper Bound Properties

30
60
Bridge Length (m)

Torsional Moment (kNm)

Torsional Moment (kNm)

30

60

90

Bridge Length (m)

4000

20000

90

1000

Bridge Length (m)

16000

30
60
Bridge Length (m)

0
4000
8000
12000
16000
20000

90

30
60
BridgeLength (m)

90

Fig. 15. (a) Torsional moment about the center line axis of the isolated and non-isolated curved bridges subjected to vertical directional earthquake. (b) Torsional moment
about the center line axis of the isolated and non-isolated curved bridges subjected to chord directional earthquake. (c) Torsional moment about the center line axis of the
isolated and non-isolated curved bridges subjected to radial directional earthquake.

Isolated Bridge

Non-isolated Bridge
0

1000
2000
3000

Lower Bound Properties

Upper Bound Properties

Shear Force (kN)

Shear Force (kN)

30
60
Bridge Length (m)

3000

90

30
60
Bridge Length (m)

30
60
Bridge Length (m)

90

30
60
Bridge Length (m)

90

90

160

40
60
80

Lower Bound Properties

Upper Bound Properties

Shear Force (kN)

20
Shear Force (kN)

2000

4000

4000

100

200
240
280
320
360

30
60
Bridge Length (m)

90

400
800
Lower Bound Properties
Upper Bound Properties

1200
0

30
60
Bridge Length (m)

90

Shear Force (kN)

0
Shear Force (kN)

1000

400
800
1200

Fig. 16. (a) Shear forces of the isolated and non-isolated curved bridges subjected to vertical directional earthquake. (b) Shear forces of the isolated and non-isolated curved
bridges subjected to chord directional earthquake. (c) Shear forces of the isolated and non-isolated curved bridges subjected to radial directional earthquake.

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

Isolated Bridge

Non-isolated Bridge
0

400
800
Lower Bound Properties
Upper Bound Properties

1200
0

30
60
Bridge Length (m)

Axial Force (kN)

Axial Force (kN)

400
800
1200

90

Axial Force (kN)

Axial Force (kN)

30
60
Bridge Length (m)

90

30
60
Bridge Length (m)

90

30
60
Bridge Length (m)

90

0
2000
4000
6000

2000
4000
6000
8000

8000
0

30
60
Bridge Length (m)

90

0
Axial Force (kN)

0
Axial Force (kN)

659

2000
4000
6000

2000
4000
6000
8000

8000
0

30
60
Bridge Length (m)

90

Fig. 17. (a) Axial forces of the isolated and non-isolated curved bridges subjected to vertical directional earthquake. (b) Axial forces of the isolated and non-isolated curved
bridges subjected to chord directional earthquake. (c) Axial forces of the isolated and non-isolated curved bridges subjected to radial directional earthquake.

Table 8
Structural accelerations of the marked point of the curved bridge (PGA
0.7g 686.70 cm/s2).

Table 9
Structural displacements of the marked point of the curved bridge.
Earthquake direction

Earthquake
direction

Isolated
(Lower)
Max.
Vertical
ax
ay
az

Displacement (cm)

Acceleration (cm/s2)
Isolated
(Upper)
Max.

NonIsolated
Max.

Isolated (Lower)
Max.
Vertical
ux

0.30
93.85
1082.27

3.78
314.17
1425.02

Chord
ax
ay
az

160.22
1.98
13.36

239.22
3.43
12.32

1636.93
5.03
9.88

Radial
ax
ay
az

1.96
185.72
214.48

3.66
420.55
207.06

5.64
1736.62
789.46

vertical earthquake components combined using the 1003030%

rule. The axial bearing forces are calculated as the sum of bearing
axial forces due to chord, radial and vertical earthquake components
combined using the 1003030% rule. Note that for the case of the

Non-Isolated
Max.

0.00
0.25
2.64

0.00
0.28
2.62

0.03
2.44
13.13

Chord
ux
uy
uz

46.43
0.46
0.03

41.73
0.46
0.03

15.98
0.05
0.08

Radial
ux
uy
uz

0.46
47.55
0.36

0.53
65.86
0.33

0.03
8.28
4.95

Uy

0.18
88.90
1084.91

Isolated (Upper)
Max.

uz

bearing axial forces the combination used is the one that corresponds to 100% of vertical earthquake. This case is the worst for the
DCFP bearings since the axial load becomes the maximum and
lateral displacement is less than the maximum value. The results
demonstrate very good agreement with the pre-calculated bearing
displacement demands.

660

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

11. Numerical results of the isolated curved bridge

As per the results, internal forces for the isolated and nonisolated curved bridges subjected to acceleration response spectra
are claried with Figs. 1417. The graphics include the lower and
upper bound properties of the double concave friction pendulum,
and maximum and minimum values of the internal forces such as
bending and torsional moments, and shear and axial forces for both
bridge congurations. The response spectra are separately applied
to the considered bridge in the chord, radial and vertical directions.
Fig. 14ac compares the results of the bending moments about
radial direction obtained using the response spectrum method. It is
noted that the responses obtained at the deck are compared for
isolated and non-isolated bridge models when subjected to earthquake ground motions. When the vertical directional earthquake is
applied to the bridge with isolation bearings, the bending moments
are nearly the same for the lower and upper bound properties cases of
the DCFP bearings. But, the differences arise in case the bridge is
subject to the two horizontal directions. The deck bending moments
obtained from the response spectrum analysis of the isolated curved
bridge decrease around 6590% if these are compared with the results
obtained for the non-isolated curved bridge. The results clearly show
that consideration of double concave friction pendulum bearings for
the isolation of the curved bridges for the response spectrum analysis
decreases the deck bending moments. Additionally, effectiveness of
the seismic isolation on the bending moments of the curved bridges is
about in portion as 1015% and in the trend of a decline when the
bridge is subjected to the vertical earthquake ground motion. Hence, it
is not to ignore that the vertical direction of the earthquake is more
effective on the bridges.
The outcomes of torsional moment about chord axis of the
isolated and non-isolated curved bridges subjected to chord, radial
and vertical directional earthquakes are depicted in Fig. 15ac. In
the light of the gures, it is seen that the seismic isolation system is
important for reduction of the undesirable torsional effects.
The reduction of the deck shear and axial forces obtained for the
isolated curved bridge is obvious, if the forces are compared with those
obtained for the non-isolated curved bridge as seen in Figs. 16 and 17.
It can also be added that the responses by the lower bound
properties are smaller than the upper bound ones.
Tables 8 and 9 give an opportunity to compare the structural
accelerations and displacements on the bridge, which are marked
in Fig. 5. It can be seen that the structural accelerations considerable decrease in case seismic isolation technique is used. However,
the structural displacements as expected increase compared with
these of non-isolated bridge.

12. Conclusion
This study outlines an investigation about the responses of the
isolated and non-isolated curved bridges subjected to response
spectra as per Caltrans Seismic Design Criteria, Version 1.4 (2006)
for magnitude 7.257 0.25, 0.7g acceleration and soil prole C. In
order to seismically isolate the bridge, the double concave friction
pendulum bearings as isolation devices are placed between the
deck and the pier/the abutments as isolation devices. Response
history acceleration of the selected ground motions is considered as
the earthquake ground motion. The analyses are carried out for the
isolated and non-isolated bridges, separately. The soilstructure
interaction is also considered by springs representing the soil
beneath footing and the drilled shaft surrounding of soil. The
maximum and minimum response values of the isolated and nonisolated bridges are compared with each other for different cases.
According to the response spectrum analysis, the displacements
of the abutment and pier bearings are different due to the pier

exibility effect and inertia effects in the substructure. Accordingly,

the abutment bearings experience lesser displacement. Maximum
axial bearing forces develop when the earthquake excitation is in
the vertical direction of the response spectrum.
Maximum bearing displacement occurs when the bridge is under
the radial directional earthquake. Besides, the total maximum displacement of each bearing is calculated as the vector sum of bearing
displacements due to vertical, chord and radial earthquake components combined using the 1003030% rule and obtained as nearly
66 cm in case the lower bound properties of the DCFP bearings are
considered. The displacement capacity is overestimated as compared
to the reference regarding response history analysis [11]. Similarly, the
shear forces at the DCFP bearings level, the accelerations transmitted
to structures and displacements obtained from response history
analysis are some of the larger values than those obtained from the
response spectrum analysis.
While the structural accelerations are decreased on the bridge,
spectral displacements are increased due to the usage of the DCFP
bearings as seismic protectors.
Finally, it is pointed out that the base isolation of the considered
curved bridge subject to the response spectra signicantly decreases
the deck responses.

Acknowledgement
The authors acknowledge the Scientic and Technical Research
Council of Turkey (TUBITAK) for supporting the studies of Sevket
ATES at the University at Buffalo.

References
[1] Hyakuda T, Saito K, Matsushita T, Tanaka N, Yoneki S, Yasuda M, Miyazaki M,
Suzuki A, Sawada T. The structural design and earthquake observation of a
seismic isolation building using Friction Pendulum system. In: Proceedings of
the 7th international seminar on seismic isolation, passive energy dissipation
and active control of vibrations of structures, Assisi, Italy; 2001.
[2] Tsai CS, Chen WS, Chiang TC, Chen BJ. Component and shaking table tests for
full-scale multiple friction pendulum system. Earthquake Engineering and
Structural Dynamics 2006;35:165375.
[3] Constantinou MC. Friction pendulum double concave bearing. NEES Report.
Available at: /http://nees.buffalo.edu/dec304/FP-DC%20Report-DEMO.pdfS;
2004.
[4] Fenz DM, Constantinou MC. Behavior of double concave friction pendulum
bearing. Earthquake Engineering and Structural Dynamics 2006;35(11):
140324.
[5] Fenz DM, Constantinou MC. Spherical sliding isolation bearings with adaptive
behavior: theory. Earthquake Engineering and Structural Dynamics 2008;37:
16383.
[6] Fenz DM, Constantinou MC. Spherical sliding isolation bearings with adaptive
behavior: experimental verication. Earthquake Engineering and Structural
Dynamics 2008;37:185205.
[7] Sennah KM, Kennedy JB. State-of-art- in design of curved box-girder bridges.
Journal of Bridge Engineering 2001;6(3):15967.
[8] DeSantiago E, Mohammadi J, Albaijat HMO. Analysis of horizontally curved
bridges using simple nite-element models. The Practice Periodical on
Structural Design and Construction 2005;10(1):1821.
[9] Mwafy AM, Elnashai AS. Assessment of seismic integrity of multi-span curved
bridges in mid-America.IL, USA: Mid-America Earthquake Center Civil and
Environmental Engineering Department, University of Illinois at UrbanaChampaign; 2007.
[10] Constantinou MC, Whittaker AS, Fenz DM, Apostolakis G. Seismic isolation of
bridges. Report submitted to the State of California Department of Transportation; 2007.
[11] Ates S, Constantinou MC. Example of application of response history analysis
for seismically isolated curved bridges on drilled shaft with springs representing soil. Soil Dynamics and Earthquake Engineering 2010, doi:10.1016/
j.soildyn.2010.09.002.
[12] Tongaonkar NP, Jangid RS. Seismic response of isolated bridges with soil
structure interaction. Soil Dynamics and Earthquake Engineering 2003;23:
287302.
[13] Spyrakos CC, Vlassis AG. Effect Of soilstructure interaction on seismically
isolated bridges. Journal of Earthquake Engineering 2002;6(3):391429.
[14] Ucak A, Tsopelas P. Effect of soilstructure interaction on seismic isolated
bridge. Journal of Structural Engineering 2008;134(7):115464.

S. Ates, M.C. Constantinou / Soil Dynamics and Earthquake Engineering 31 (2011) 648661

[15] Federal Emergency Management Agency. FEMA 440, Improvement of Nonlinear

Static Seismic Analysis Procedures. Washington DC; 2005.
[16] European Committee for Standardization. Design of Structures for Earthquake
Resistance. Part 2: Bridges, Eurocode 8, EN1998-2, draft; August 2005.
[17] American Association of State Highway and Transportation OfcialsAASHTO.
Guide specications for seismic isolation design. Washington, D.C; 1999.
[18] California Buildings Standards Commission. California Building Code, Sacramento,
California; 2001.
[19] Berger/Abam Engineers, Inc. Federal Highway Administration Seismic Design
Course, Design Example No. 6, Publication no. FHWA-SA-97-011 and Barcode
no. PB97-142111; 1996.

661

[20] California Department of TransportationCaltrans. Seismic Design Criteria,

Version 1.4; 2006.
[21] Terzagi K. Evaluation of coefcient of subgrade reaction. Geotechnique, London
1955;5:4.
[22] Wilson E, Suharwardy I, Habibullah E. A clarication of the orthogonal effects
in a three-dimensional seismic analysis. Earthquake Spectra 1995;11:4.
[23] SAP2000. Integrated nite element analysis and design of structures basic analysis
reference manual. Berkeley, California: Computer and Structures Inc.; 2007.
[24] Constantinou MC, Whittaker AS, Kalpakidis Y, Fenz DM, Warn GP.. Performance
of seismic isolation hardware under service and seismic loading. Technical
report MCEER-07-0012, Buffalo, NY; 2007.