Engineering Structures 222 (2020) 111129

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Influence of ground motion duration and isolation bearings on the seismic T

response of base-isolated bridges
⁎
Afraa Labiba Hassana, AHM Muntasir Billahb,
a
Department of Civil Engineering, Lakehead University, Thunder Bay, ON P7B 5E1, Canada
b
Department of Civil Engineering, Lakehead University, Thunder Bay, ON P7B 5E1, Canada

A R T I C LE I N FO A B S T R A C T

Keywords: This study investigates the effect of ground motion duration on the seismic response of base-isolated bridges.
Long-duration motion Three different isolation bearings are considered in this study, such as lead rubber bearing (LRB), Shape memory
Seismic isolation alloy wire-based lead rubber bearing (SLRB), and Friction pendulum system (FPS). Each isolation system is
Highway bridges designed to provide a similar isolation period so that the seismic response is comparable. Using 20 long-duration
Residual displacement
motions and 20 spectrally matched short-duration records; the seismic performance of the base-isolated bridges
Seismic response
is evaluated and compared using nonlinear time history analyses. Response parameters considered for this study
are the base shear in the piers, the acceleration of the bridge deck, maximum and residual displacement of the
isolation bearings as well as the energy dissipation capacity. The results indicate that the long-duration motions
cause more damage to different bridge components as compared to the short-duration motions in-terms of higher
deck acceleration, pier base shear, and residual isolator displacement. It is also observed that ground motions
having similar magnitude with different significant durations can significantly affect the isolator as well as the
bridge response.

1. Introduction One of the main objectives of using isolation bearings is to modify

Seismic base isolation is a proven technique for reducing the seismic
demand on structures and improving seismic performance. Seismically
isolated structures have shown reduced damage and superior seismic
performance during extreme earthquake events. The advantages of
protecting structures in seismic regions using isolation bearings have
been recognized all over the world [1–5]. Especially for bridges, seismic
isolation bearings are the preferred choice over other seismic protection
systems for seismic upgrading of existing structures and new con-
structions. Different types of rubber-based seismic isolation bearings
such as high damping rubber bearings (HDRB), steel-reinforced elas-
tomeric bearings (SREB), lead rubber bearings (LRB), fiber-reinforced
elastomeric bearings (FREB), and friction pendulum systems (FPS) have
gained popularity for effective seismic control and improving the
seismic performance of infrastructures [6–9]. Among many advantages
base-isolated bridges have over non-isolated bridges, major advantages
include: (i) decoupling of superstructure and substructure, (ii) reduced
seismic force demand in substructure, (iii) reduced substructure and
foundation dimensions, (iv) avoid retrofit and traffic disruptions after
strong earthquakes, and (v) overall reduction in the life cycle cost of the
bridge.
the fundamental period of the structure away from the predominant
earthquake period thus reducing the seismic demand on the structure.
However, the occurrence of several long-duration earthquakes over the
last decade has created interest among researchers on evaluating the
seismic performance of structures under long-duration motions. Long-
duration motions typically occur in subduction zones which are plate
boundaries such as Tokyo (Japan), Taiwan (China), and Santiago
(Chile). In western North America, the Cascadia subduction zone,
stretching from northern Vancouver Island in British Columbia to
northern California is the probable location that could experience long-
duration ground motions. Recent studies [10] on the impact of sub-
duction zone earthquakes have indicated that the long-duration mo-
tions can have a detrimental effect on structures having a fundamental
period of vibration of 1.0 s or larger. However, base-isolated structures
typically have a fundamental period of vibration longer than 1.0 s.
Because of this fact, it is necessary to evaluate the performance of base-
isolated bridges under long-duration earthquakes. For any structure
whose performance continues to worsen under reversed cyclic loading,
the ground motion duration is thought to be a substantive parameter for
performance prediction [11]. However, there is still a lack of consensus
among the researchers whether the ground motion duration should be

⁎
Corresponding author.
E-mail address: muntasir.billah@lakeheadu.ca (A.M. Billah).

https://doi.org/10.1016/j.engstruct.2020.111129
Received 24 January 2020; Received in revised form 12 June 2020; Accepted 14 July 2020
Available online 23 July 2020
0141-0296/ © 2020 Elsevier Ltd. All rights reserved.
A.L. Hassan and A.M. Billah
considered as a significant parameter for seismic performance assess-
ment and design of structures [12].
In the case of isolated bridges, earthquake duration effects have not
been studied despite most bridges located in areas susceptible to long-
duration motions being equipped with different types of isolation
bearings. Han et al. [13] investigated 320 damaged bridges following
the Wenchuan (China) earthquake in 2008. They found that 14% of the
bridges were severely damaged while 39% of bridges suffered moderate
damage. Among the damaged bridges, many of them were isolated with
laminated rubber bearings [14]. For performance-based seismic design
(PBSD) of new bridges or seismic retrofitting of existing bridges, it is
required that the isolated bridge must satisfy the life safety criteria and
meet collapse prevention objectives under large infrequent earth-
quakes. However, current design codes and guidelines do not consider
the effect of long-duration earthquakes on the design of seismically
isolated bridges neither for new construction nor for the retrofitting of
existing bridges. A good number of studies have been conducted over
the past several years that explored the ground motion duration effect
on different types of structures. Chandramohan et al. [15] investigated
the collapse response of a steel moment frame and bridge pier under
long-duration and spectrally matched short-duration motions. This
study shows that long-duration motions tend to decrease the collapse
capacity of the bridge pier as compared to the short-duration motions.
Lopez et al. [16] experimentally investigated the performance of sub-
standard bridge columns under long-duration motions. They found that
bridge piers experienced severe damage under the long-duration mo-
tions as compared to the spectrally equivalent crustal ones. Pan et al.
[17] investigated the seismic performance of conventional low-rise
light-frame wood houses under long-duration ground motions. They
concluded that the median collapse capacities of the studied buildings
reduced by 26%-61% under the long-duration motions as compared to
the short-duration motions. Hammad and Moustafa [18] numerically
investigated the performance of a special concentrically braced frame
(SCBF) under short and long-duration ground motions. They found that,
for a given intensity measure, the long-duration motions result in a
higher probability of collapse. Kabir et al. [19] evaluated the seismic
fragility of a highway bridge under long-duration, near-fault, and short-
duration far-field motions. They concluded that long-duration ground
motions dominate both component level and bridge system failure
probabilities.
All past studies that investigated the performance of seismically
isolated bridges, compared the performance under near-fault motions
with pulse-like velocity and far-field motions. Shen et al. [20] evaluated
the performance of a lead rubber bearing isolated bridge under pulse-
like near-fault motions. They found that when the pulse period of the
ground motion is close to the effective period of the system, the bridge
response is amplified. Dicleli [21] studied the impact of substructure
and isolator properties along with the frequency characteristics of the
ground motions on the performance of seismically isolated bridges.
They identified the selection of ground motions according to the site
characteristics as a major factor for effective isolation system design.
Ozbulut and Hurlebaus [22] investigated the comparative performance
of different isolation bearings under near-fault motions. Eröz and Des-
Roches [23] compared the performance of LRB and FPS isolation
bearings in a typical Multi-Span Continuous Concrete Girder bridge
considering the effect of vertical ground motions. They concluded that
both FPS and LRB improved the seismic performance of the bridge
comparably. Although the vast literature supports the advantages of
seismic isolation, no literature can be found that investigated the
seismic performance of isolated bridges under long-duration motions.
This study aims to evaluate the seismic response of isolated bridges
under long-duration motions. Three different seismic isolation systems
and spectrally matched long and short-duration record sets are em-
ployed to demonstrate the effect of ground motion duration on the
seismic response of isolated bridges. The isolation bearings considered
include Lead Rubber Bearing (LRB), Shape memory alloy wire-based
2
Engineering Structures 222 (2020) 111129
Lead Rubber Bearing (SLRB), and Friction Pendulum System (FPS).
Using 20 long-duration (LD) motions and 20 spectrally matched short-
duration (SD) records, the performance of a three-span bridge isolated
with three different isolation systems are evaluated. Response para-
meters considered for this study are the base shear in the piers, the
acceleration of the bridge deck, maximum and residual displacement of
the isolation bearings as well as the energy dissipation capacity.
2. Ground motion Duration:
The seismic performance of bridges is significantly influenced by the
ground motion intensity, frequency content, and duration of ground
motion. The seismic response of bridges during a long-duration earth-
quake can vary significantly from short-duration ground motions [15].
During a long-duration earthquake, a bridge structure is expected to
experience a large number of inelastic displacement reversals, which
tend to cause significant damage and degradation of strength and
stiffness [24]. To evaluate the seismic response under long-duration
motion, it is necessary to quantify the duration metric to define a
ground motion as a long-duration motion. Different definitions of
ground motion duration exist in the literature. Bommer and Martinez-
Pereira [25] reviewed approximately 30 different definitions of strong-
motion duration. Among various duration matrices, arias intensity (AI)
[26], significant duration (tD5-95) [27], bracketed duration (DB) [28],
Cumulative Absolute Velocity (CAV) [29], and dimensionless duration
metric (ID) [30] are commonly used to define the ground motion
duration metric.
All these duration matrices have been used by different researchers
for specific purposes. Among these matrices, 5%-75% significant
duration (Ds5-75) has been used and recommended by most of the re-
searchers [11,12,15,31,32] for evaluating structural response under
long-duration motions. Based on the recommendations from previous
studies, this study employed a 5%-75% significant duration (Ds5-75) to
quantify the ground motion duration. The calculation of the significant
duration of a specific ground motion is shown in Fig. 1, which re-
presents the time interval over which a specific percentage of the in-
tegral (shown in equation (1)) is accumulated:
Tr
π
AI =
2g
∫ a2 (t ) dt
0 (1)
where a(t) represents ground acceleration history, Tr denotes the total
recorded time of the ground motion, and g is the acceleration due to
gravity.
Previous researchers [15] suggested that spectral shapes need to be
taken into account when assessing structural response to long-duration
motions. While investigating the ground motion duration effect on the
seismic response of structures, it is desirable to compare the response
with ‘similar’ short-duration ground motion. This “similarity” can be
achieved by matching the spectral acceleration shape [33]. In this
study, the duration effect on the seismic response is evaluated by
comparing the responses under a long-duration record set and a spec-
trally matched short-duration record set both containing 20 ground
motions.
In this study, a set of 50 long-duration ground motions are first
selected from six different historical earthquake events: 2011 Tohoku,
Japan; 2010 Maule, Chile; 2007 Sumatra, Indonesia; 2003 Hokkaido;
Japan, 1999 Chichi, Taiwan; and 1985 Valparaiso, Chile. These ground
motion records are obtained from the PEER NGA West2 ground motion
database [34] and the Center for Engineering Strong Motion Data [35].
The list of selected long-duration motions is provided in Table 1.
Among these ground motions, records having tD5-75 > 25 sec are se-
lected as the long-duration motions. All the selected ground motions
have a PGA > 0.1 g and PGV > 10 cm/sec. This ground motion
selection process provided a suite of long-duration motions containing
32 records, with a geometric mean Ds5-75 of 34 sec. To compare the
A.L. Hassan and A.M. Billah Engineering Structures 222 (2020) 111129

Fig. 1. Significant duration calculation of a ground motion.

seismic performance of the three isolated bridges under the long and
short-duration motions, another 40 ground motions are selected from
the PEER NGA-West2 database [34]. Among these ground motions,
records having Ds5-75 < 20 sec and epicentral distance of more than
15 km are selected. The short-duration accelerograms are selected using
the spectral equivalency approach of Chandramohan et al. [15] which
resulted in similar spectral shapes like the long-duration counterpart in
the period range of [0.05, 4.00] s. To minimize the sum of squared error
differences of the 5%-damped linear response spectra associated with
each pair of short and long-duration records, a matching process is
performed. To minimize the error, the short-duration motions are am-
plitude scaled, without affecting the frequency content of the ground
motions, when necessary using a simple amplitude scaling factor of up
to 5. This resulted in the creation of a spectrally matched long-duration
and short-duration set each containing 20 records. Since the short-
duration motions are spectrally matched to long-duration motions, the
differences in the seismic response of the isolated bridges could be at-
tributed to the difference in their duration characteristics.
Fig. 2A compares the response spectrum of a long-duration ground
motion (from the 1992 Landers earthquake) and the spectrally
equivalent short-duration motion (from the 1986 Taiwan earthquake).
Fig. 2B shows the acceleration time histories of the two records. The list
of selected 40 ground motions (20 long-duration and 20 spectrally
equivalent short-duration motions) are provided in Table 1. Fig. 3
shows the selected scaled short and long-duration individual records
along with their means as compared to the target spectrum.

Table 1
Database of long-duration and spectrally equivalent short-duration motion.
Long Duration Motions Spectrally matched short duration motion

GM Pair Earthquake Year Station Magnitude Ds 5–75 Earthquake Year Station Magnitude Ds 5–75
No. (s) (s)

1 Landers 1992 Thousands Palm post 7.2 26 Taiwan SMART1(45) 1986 SMART1 I11 6.3 11
office
2 Landers 1992 Indio-Jackson Road 7.2 26 Northridge-01 1994 Camarillo 6.7 14
3 Valparaiso, Chile 1985 Vina del Mar 7.8 32 Iwate 2008 Miyagino-ku 6.9 6
4 Valparaiso, Chile 1986 Zaplar 7.8 30 Morgan Hill 1984 Hollister Diff Array #1 6.2 11
5 Valparaiso, Chile 1985 Llolleo 7.8 28 Northridge-01 1994 Sun Valley - Roscoe Blvd 6.7 6
6 Valparaiso, Chile 1986 San Isidro 7.8 30 Christchurch, New 2011 SWNC 6.1 2
Zealand
7 Maule, Chile 2010 Santiago Maipu 8.8 33 Northridge-01 1994 LA - Pico & Sentous 6.7 9
8 Maule, Chile 2010 Talca 8.8 53 Taiwan SMART1(5) 1986 SMART1 M10 6.3 4
9 Maule, Chile 20,101 Angol 8.8 31 Irpinia, Italy-01 1980 Brienza 6.9 4
10 Tohuku Japan 2011 Sendai 9.0 55 Kobe, Japan 1995 Kobe University 6.9 3
11 Tohuku Japan 2012 Sakura 9.0 29 Chi-Chi, Taiwan-03 1999 CHY028 6.2 6
12 Tohuku Japan 2011 Shiogama 9.0 56 Northridge-01 1994 Tarzana - Cedar Hill A 6.7 4
13 Tohuku Japan 2012 Tsukidate 9.0 58 Iwate 2008 Maekawa Miyagi 6.9 7
Kawasaki City
14 Michoacan, 1985 Villita Corona Centro 8.0 32 Darfield, New Zealand 2010 Heathcote Valley 7.3 8
Mexico Primary School
15 Kocaeli, Turkey 1999 Fatih 7.5 28 Chi-Chi, Taiwan-06 1999 HWA025 6.2 3
16 Kocaeli, Turkey 2000 Bursa Tofas 7.5 26 Chi-Chi, Taiwan-04 1999 KAU085 6.2 20
17 El Mayor-Cucapah 2010 Ejido Saltillo 7.2 33 Chi-Chi, Taiwan 1999 TCU075 6.2 18
18 El Mayor-Cucapah 2010 Tamaulipas 7.2 27 Landers 1992 Amboy 7.2 17
19 Hokkaido, Japan 2003 Hayakita 8.3 25 Chi-Chi, Taiwan 1999 TTN026 6.2 20
20 ChiChi, Taiwan 1999 CHY008 7.6 32 Christchurch, New 2011 LINC 6.2 7
Zealand

3
A.L. Hassan and A.M. Billah
Fig. 2. Comparison of the (a) response spectra and (b) time histories of spectrally equivalent long-duration and short-duration record. The long-duration record is
from the 1992 Landers earthquake, recorded at the Thousands palm post office station. The short-duration record is from the 1986 Taiwan earthquake, recorded at
the SMART1 I11 station, scaled by a factor of 0.75.
3. Description of the bridge
A three-span continuous concrete girder bridge located in Victoria,
British Columbia, Canada is considered in this study. The bridge su-
perstructure is supported on two-column bents with unequal bent
heights of 15.5 m and 11.5 m. The concrete superstructure is composed
of 250 mm thick cast in place concrete slab supported on three simply
supported precast concrete University of Nebraska (NU) I-girders (NU
2000) girders at 4.5 m spacing as shown in Fig. 4. The existing bridge is
not equipped with seismic isolation bearings. The concrete super-
structure sits on elastomeric pads at each abutment and pier locations.
Each column bent has two circular piers connected by a rectangular
concrete pier cap. The column spacing within a bent is 6.6 m. The
1500 mm circular reinforced concrete piers are reinforced with
28–30 M (diameter 29.9 mm) longitudinal rebars and 15 M (diameter
16 mm) bars as spiral reinforcement. The considered bridge is classified
as a major route bridge according to the current Canadian Highway
Bridge Design Code (CHBDC) [36]. However, with the current struc-
tural configuration, the bridge does not satisfy the performance re-
quirements of a major route bridge outlined in CHBDC. The bridge is
founded on deep foundations both at the pier and abutment locations
which consist of steel HP - Bearing Piles (H-shaped cross-section having
parallel flange surfaces and equal web and flange thicknesses). The pier
columns sit on a rectangular concrete pile cap which is connected to
two rows of steel HP Piles. Conventional seat type abutments supported
on steel HP piles support the bridge ends. The details of the geometric
properties of different bridge components are provided in Table 2.
4. Seismic isolation bearings
To investigate the effect of ground motion duration on seismically
isolated bridges, three different types of isolation bearings are con-
sidered in this study, namely lead rubber bearing (LRB), Friction
Fig. 3. Mean response spectra for long-duration (a) and short-duration (b) ground motions compared to the target response spectrum of bridge location.
4
Engineering Structures 222 (2020) 111129
pendulum system (FPS), and Shape memory alloy wire-based lead
rubber bearing (SLRB). Among these three types of bearings, LRB and
FPS have been widely implemented around the world for seismic iso-
lation of bridges and buildings. Over the last few years, researchers
have developed several configurations of shape memory alloy (SMA)
wire-based isolation bearings [22,37,38] and reported improved per-
formance under seismic actions. In this study, SMA-based LRB equipped
with the double-cross wire configuration [39] has been considered to
evaluate its effectiveness under the long-duration motions.
To achieve comparable structural performance with three different
types of isolation bearings, the main design objective of the three iso-
lation systems is to achieve a similar fundamental isolation period of
the bridge while keeping the column elastic under the design level
earthquake. The target isolation period is set as 1.5 s. The design and
analysis of the LRB and FPS are performed to satisfy the requirements of
CHBDC [36] and the provisions of the AASHTO Guide Specifications for
Seismic Isolation Design [40] following the procedure outlined in
NCHRP Report [41]. The design of SLRB is performed following the
method described by [38,42] while meeting the requirements of the
CHBDC [36]. The properties of the three isolation bearings used in this
study are summarized in Table 3.
5. Finite element modeling
A detailed three dimensional (3D) finite element model of the
bridge is developed using a nonlinear fiber-based finite element pro-
gram Seismostruct (Seismosoft 2020) [43]. Fig. 5 shows the different
components of the finite element model of the bridge which includes
some idealizations and realistic assumptions. For seismically isolated
bridges, it is a common assumption that the superstructure (girder and
deck) will remain elastic under seismic excitation. As a result, the
concrete girders are modeled using an elastic beam element having
equivalent stiffness and mass. Using the NU2000 girder geometry and
A.L. Hassan and A.M. Billah Engineering Structures 222 (2020) 111129

Fig. 4. (a) Plan and Elevation of the bridge, (b) Typical cross section at pier location, (c) Cross section of the pier cap beam, and (d) Cross section of the column.

Table 2 segments based on the tributary area of each girder segment. Forced-
Geometric properties of the bridge. based fiber elements in Seismostruct are used to model the circular
Properties Dimension Unit
concrete columns and the bent cap (Fig. 5c). For defining the confined
and unconfined concrete, the constitutive relationship proposed by
Dimension of the pier cap (L × d × w) 11,392 x1,600 × 1,800 mm Mander et al. [44] is used. The steel reinforcement is defined using the
Dimension of pier (diameter) 1500 mm Menegotto-Pinto stress-strain relationship [45] that includes the iso-
Dimension of the pier pile cap (L × d × w) 12,500 × 1,800 × 4,000 mm
Deck Thickness 250 mm
tropic strain hardening property. This steel model considers low-cycle
Pile Section HP 360 × 132 fatigue for steel bars [46]. However, to predict the actual effect of
Girder Section NU2000 ground motion type, capturing the cyclic deterioration of the structure
Length of pile (varies) 20–25 m is crucial. Analytical validation of experimental cyclic deterioration by
No. of piles at pier 12
Billah et al. [47] demonstrates the accuracy of the program Seismos-
No. of piles at abutment 12
truct in predicting the cyclic strength degradation of structures under
long-duration motions.
concrete properties, the elastic beam elements are defined by calcu- Since the bridge is supported on seat type abutments, the abutment
lating the values of EA, EI2, EI3 and GJ, where A is the cross-section response in both longitudinal and transverse directions is represented
area, E is the elastic modulus, I2 and I3 are the moments of inertia using bilinear springs as recommended by Caltrans [48] and CHBDC
around local axes (2) and (3), G is the shear modulus, and J is the [36]. The abutment longitudinal response is modeled combining the
torsional constant (Fig. 5b). Lumped masses are defined on girder soil response and pile response as shown in Fig. 5d. Ignoring the con-
tribution of the wing wall, only the pile response in the transverse

Table 3
Properties of the three isolation bearings.
LRB FPS SLRB

Parameter Value Unit Parameter Value Unit Parameter Value Unit

Characteristic Strength 187 kN Shear Stiffness, Ks 9 kN/mm Shear Stiffness, Ks 9.3 kN/mm
Post Elastic Stiffness 1.75 kN/mm Friction coefficient at fast velocities 0.06 SMA wire diameter 2.5 mm
Effective Stiffness 3.6 kN/mm Friction coefficient at slow velocities 0.03 Superelastic strain limit 13.5 %
Yield Force 231 kN Curvature radii of friction pendulum 2.5 m Characteristic Strength 61 kN
Post-yield hardening ratio, r 0.19 Bearing hardening ratio 0.117
Initial Stiffness, Ki 9.23 kN/mm

5
A.L. Hassan and A.M. Billah Engineering Structures 222 (2020) 111129

Fig. 5. Details of the finite element model of the bridge (a) Modeling details at pier location, (b) elastic beam element for superstructure, (c) nonlinear beam-column
element for column and bent cap, (d) abutment model in longitudinal direction, (e) abutment model in transverse direction, (f) gap element model, (g) bilinear model
of LRB, (h) bearing 2 model of FPS, and (i) bearing 1 model of SLRB.

direction is considered as shown in Fig. 5e. Other researchers have used hardening model could not accurately capture the response of SLRB
a similar approach for modeling the abutment response under seismic under reversed cyclic loading. This bearing 1 element requires defining
excitation [49]. The gap between the deck and abutment is modeled some dimensionless parameters A, β, and γ. The parameters β and γ
using a zero-length nonlinear spring (bilinear gap element) following represent the hysteretic shape variable that controls the shape of the
the suggestion of Muthukumar and DesRoches [50]. This impact ele- hysteretic loop. As suggested by Constantinou and Adnane [55], the
ment requires the definition of several parameters (Kt1, Kt2, Δy, and Δm) rule A/(β + γ) = 1 is followed for defining the hysteretic shape vari-
as shown in Fig. 5f. Here, the maximum gap opening (Δm) is considered ables. Following this, the values for the parameters A, β, and γ are
to be 25.4 mm with the yield displacement (Δy) is assumed to be 10% of defined as 1, 0.5, and 0.5, respectively.
Δm .
In this study, the isolation bearings are modeled using a zero-length
5.1. Validation of finite element model
link element. The hysteretic behavior of LRB is represented using a
bilinear kinematic hardening model available in Seismostruct. Three
Although the ability of the FE program in predicting structural re-
parameters are needed to define the hysteresis behavior of LRB such as,
sponse under static and dynamic loading has been verified by different
the initial stiffness (Ki), post-yield hardening ratio (r), and yield force
researchers [56,57], the accuracy of the adopted bearing modeling
(Fy), which are obtained from the properties of LRB as shown in Table 3.
techniques is verified against available experimental results to confirm
The FPS bearing is defined using the zero-length bearing 2 element
the adequacy of the isolation bearing models used in this study. Fig. 6
(friction pendulum bearing) available in Seismostruct. The model re-
shows the comparison of the lateral load-displacement response of the
quires the definition of friction coefficient, which is calculated ac-
three isolation bearings obtained from the experiment and the finite
cording to the Constantinou et al. [51] friction model, the radius of
element models. The first two experiments of LRB and FPS bearings are
curvature of the friction pendulum, and vertical force on the bearing.
reported in Constantinou et al. [58]. The LRB was tested under vertical
Since the bridge considered does not have any shear key, the cutting off
stress of 6.7 MPa under lateral sinusoidal motion with a frequency of
process of shear keys [52] has not been considered in this study. For
0.35 Hz and a peak shear strain of 58%. Fig. 6a compares the hysteretic
representing the SLRB, the zero-length bearing 1 element is used which
behavior of the LRB specimen calculated from the FE model and the
follows the hysteretic rules proposed by Wen [53] and Park et al. [54].
experiment. The comparison shows that the bilinear kinematic model
Previously, Dezfuli and Alam [39] showed that the bilinear kinematic
can mimic the experimental force-displacement behavior with

6
A.L. Hassan and A.M. Billah
Fig. 6. Experimental and simulated force-displacement histories of three isolation bearings (a) LRB [54], (b) FPS [54], and (c) SLRB [35].
reasonable accuracy. Fig. 6b shows the hysteretic response of the FPS
bearing obtained from the experiment and the numerical model. The
FPS was subjected to a lateral displacement cycle having a frequency of
0.6 Hz and constant vertical pressure of 30.8 MPa. Fig. 6b compares the
normalized force-displacement histories recorded during the experi-
ment and simulation for the first three cycles. The simulated force-
displacement relationship obtained using the bearing 2 elements could
predict the experimental results with reasonable accuracy. The nu-
merical model of the SLRB is validated against the experimental results
of Dezfuli and Alam [39]. The bearing was equipped with 2.5 mm SMA
wire in double-cross configuration and subjected to sinusoidal loading
up to 200% shear strain. Fig. 6c compares the hysteretic response ob-
tained from the test and numerical model confirming the accuracy of
bearing 1 element in replicating the response of SMA-LRB under re-
versed cyclic loading.
5.2. Modal analysis of the isolated bridges
One of the main design objectives of the three isolation systems was
to achieve a similar vibration period for the three bridges. Before per-
forming nonlinear time history analysis, modal analysis of the three
isolated bridges was conducted. Table 4 summarizes the first three
modal periods of the three isolated bridges along with the direction of
vibration. From Table 4 it can be seen that all three isolated bridges
have very similar modal behavior, which ensures that their seismic
Table 4
Modal analysis results of the three isolated bridges.
Mode-1 Mode-2
Isolation System Period, T (s) Direction Period, T (s)
LRB 1.527 Transverse 1.305
FPS 1.534 Transverse 1.316
SLRB 1.521 Transverse 1.299
7
Engineering Structures 222 (2020) 111129
response can be compared. All three isolated bridges achieved a vi-
bration period close to the target period of 1.5 s.
6. Results and discussions
The performance of the three isolation systems under the long-
duration and spectrally equivalent short-duration motions are com-
pared in terms of different structural response parameters such as
maximum isolator displacement (MID), residual isolator displacement
(RID), peak normalized base shear (PNBS) in the column, maximum
pier deformation (MPD), maximum deck acceleration (MDA), and iso-
lator energy dissipation (IED). The seismic response of the isolated
bridges subjected to the ground motions described above is compared
through nonlinear time history analysis.
6.1. Pier base shear
One of the main objectives of seismically isolated bridges is to de-
couple the superstructure from substructure to reduce the seismic de-
mand on the bridge piers. This reduced base shear demand can reduce
the overall dimension of the substructure including the foundation re-
sulting in a cost reduction. In this study, the pier base shear in different
isolated bridges under the two sets of ground motions is compared in
terms of the peak normalized base shear (PNBS). The base shear ob-
tained from the time history analysis of the different isolated bridges is
Mode-3
Direction Period, T (s) Direction
Longitudinal 1.032 Torsional
Longitudinal 1.037 Torsional
Longitudinal 1.029 Torsional
A.L. Hassan and A.M. Billah
Fig. 7. Peak normalized base shear (PNBS) for bridge piers isolated with different isolation bearings under LD and SD ground motions (a) LRB, (b) SLRB, and (c) FPS.
normalized by the weight of the deck. Fig. 7 shows the box plots of the
PNBS obtained from the dynamic analysis of the bridges isolated with
three different systems under the LD and SD ground motions. In each
figure, the PNBS of each isolated bridge is compared under the LD and
spectrally equivalent SD ground motions. In each diagram, the height
represents the numerical range of the data (maximum and minimum
values) for the PNBS. The “boxes” represent the 25th through the 75th
percentile. The horizontal line inside the box is the median value (or
50th percentile) and the cross mark (x) with value represents the mean
value of the response parameter. From Fig. 7, it can be observed that
irrespective of the isolation system used, the LD ground motions tend to
increase the base shear demand in the pier. This can be attributed to the
higher seismic energy imparted by LD motions which resulted in higher
base shear demand. This outcome is in agreement with previous re-
search, which has shown that LD motions increase the base shear de-
mand. The largest individual PNBS is observed in the SLRB system
under the LD ground motions. However, among the three isolation
systems, SLRB resulted in the largest average PNBS of 0.43 under the LD
ground motion suite. On the other hand, under the SD motions, the
average PNBS obtained from the SLRB isolated system is only 0.27.
From the average values of PNBS presented in Fig. 7, it can be observed
that under the SD motions, the average PNBS of the three systems is
almost identical. On the other hand, under the LD motions the values
are significantly different where the bridge piers with SLRB experience
7.5% and 22% higher PNBS as compared to the bridges with FPS and
LRB bearings, respectively. Additionally, the PNBS experienced under
the LD motions is significantly larger for all three isolation systems as
compared to the SD counterpart. This indicates that the isolated bridges
designed following current design codes can significantly under-
estimate the substructure design forces under the LD ground motions.
6.2. Maximum pier displacement
Fig. 8 compares the maximum displacement experienced by the
bridge piers isolated with three different bearings under the LD and SD
ground motions. Since isolation bearings increase the flexibility of the
structure, it is expected that the pier displacement will increase for an
Fig. 8. Maximum pier displacement (MPD) for bridge piers isolated with different isolation bearings under LD and SD ground motions (a) LRB, (b) SLRB, and (c) FPS.
8
Engineering Structures 222 (2020) 111129
isolated bridge. Analogous to the PNBS, the maximum pier displace-
ment (MPD) is also influenced by the ground motion duration which is
reflected by the increased pier displacement under the LD motions.
Because of the accumulated damage under the LD motions, the stiffness
of the pier was reduced, resulting in an elongated period of the bridge
thereby increasing the maximum pier displacement. Among the three
isolated bridges, the LRB isolated bridge pier experienced maximum
average displacement of 53 mm under the LD motions which is 4% and
20% higher than the FPS and SLRB isolated bridge pier average max-
imum displacement, respectively. Conversely, under the SD motions,
SLRB isolated pier experienced the smallest deformation which is 25%
and 31% less than the LRB and FPS system, respectively. However, ir-
respective of the isolation system used LD motions tend to increase the
displacement demand in the bridge pier as compared to the SD motions.
Evaluation of the pier top displacement for all three isolation systems
reveals that there is an average 30% increase in MPD under the LD
motions as compared to the SD motions.
6.3. Maximum deck acceleration
Another major objective of using seismic isolation bearing is to re-
duce the acceleration of the bridge deck induced by the ground motion.
Typically, the deck acceleration is proportional to the seismic force
applied to the structure. Fig. 9 compares the deck acceleration in the
three isolated bridges under the LD and SD ground motions. All three
systems experienced higher deck accelerations when subjected to the
LD ground motions. In this regard, it should be noted that similar
performances of the bridges under LD motions which increased the base
shear lead to analogous performances in increasing the peak deck ac-
celeration. Of the three isolation systems, SLRB effectively reduced the
bridge deck acceleration as compared to the FPS and LRB systems ir-
respective of the ground motion duration. As seen from Fig. 7, the SLRB
resulted in higher shear force in the pier but resulted in smaller deck
acceleration (Fig. 9). Dezfuli and Alam [39] reported a similar ob-
servation where they compared the performance of SLRB and LRB
under near-fault ground motions. For the LRB system, the LD motion
increased the deck acceleration by 30% as compared to the SD motions
A.L. Hassan and A.M. Billah Engineering Structures 222 (2020) 111129

Fig. 9. Maximum deck acceleration (MDA) for bridges isolated with different isolation bearings under LD and SD ground motions (a) LRB, (b) SLRB, and (c) FPS.

whereas this increase is 25% and 32% for SLRB and FPS system, re- |F +| + |F −|
keff =
spectively. |Δ+| + |Δ−|

where Δ+ and Δ- are the maximum and minimum horizontal dis-

6.4. Isolation bearing response placements obtained from each time history analysis, and F+ and F- are
the corresponding forces. From Fig. 14, it can be observed that, under
To compare the performance of the three isolation systems, different LD motions, the maximum values of isolator displacement are generally
operational characteristics of the isolation bearings such as the max- smaller as compared to the spectrally equivalent SD motions while the
imum shear strain (γmax), residual displacement (RD), effective hor- corresponding forces are higher. As a result, under the LD motions, the
izontal stiffness (KH,eff), and total dissipated energy (E) are obtained for bearings exhibited a higher effective horizontal stiffness. The mean
the two sets of ground motions. Table 5 summarizes the average values effective horizontal stiffness of SLRB is much higher than that of LRB
of these parameters obtained from the dynamic analysis of the three and FPS. For example, under SD motions, the mean value of KH,eff of
isolated bridges under the LD and SD ground motions suite. SLRB is 80% and 49% higher than that of LRB and FPS, respectively
The maximum isolator displacement (MID) and residual isolator which is 52% and 15% higher under LD motions as compared to the
displacement (RID) obtained from the nonlinear time history analyses LRB and FPS, respectively. This supports the results obtained for PNBS
using 40 ground motions are summarized in Figs. 10 and 11, respec- which showed that the bridge pier isolated with SLRB experienced
tively. These figures present the 10th, 25th, 75th, and 90th percentile larger base shear as compared to LRB and FPS isolated bridges.
cumulative probabilities of MID and RID along with the mean value of The post-earthquake functionality of bridges is one of the major
each response parameter Fig. 12 shows the time variation of lateral requirements of the current performance-based seismic design.
displacement of the three isolation bearings under the LD motion #7 Isolation bearings need to have an adequate re-centering capacity to
and spectrally equivalent SD ground #7. ensure the post-earthquake functionality of bridges. A large amount of
A general comparative assessment from the average values in residual deformation, accumulated after the end of an earthquake, in-
Figs. 10 and 11 show that the LRB experienced larger MID and RID as dicates the inadequate restoring capacity of the isolation system [59].
compared to SLRB and FPS irrespective of the ground motion type. As Different researchers proposed several advanced isolation systems for
seen from Fig. 10, the SD motions tend to impose higher deformation on controlling the residual displacement in isolation bearings
the isolation bearings as compared to the LD motions irrespective of the [37,38,60,61]. In an attempt to reduce the residual displacement of
bearing types. For example, the SD motion increases the MID for LRB conventional isolation bearing, Wei et al. [60] proposed a vertical
and FPS on an average by 20% whereas it is increased by 35% for SLRB. spring-viscous damper-concave Coulomb friction isolation system.
Among the three bearing types, SLRB sustained the smallest MID for Through an extensive performance-based assessment, they found that it
both SD and LD ground motions. As reported in Table 5, LRB reached an is possible to achieve zero residual displacements on the proposed
average maximum shear strain of 100% whereas the maximum shear isolation system by using super lubrication in the middle of the contact
strain in FPS and SLRB is 88% and 82%, respectively under the SD surface and a variable increment ratio of concave friction distribution.
ground motions. Similarly, under the LD motions, the mean peak shear Fig. 11 shows the summary of residual isolator displacement (RID)
strain is 78%, 52%, and 77%, for LRB, SLRB, and FPS, respectively. This experienced by different isolations systems under the LD and SD mo-
phenomenon can be attributed to the differences in the bearings’ ef- tions. From Fig. 11, it is evident that irrespective of the bearing used,
fective horizontal stiffness (Table 5). In this study, the bearings’ effec- the LD motions tend to increase the RID significantly. The results show
tive stiffness is calculated as the secant stiffness of the isolation system that the SLRB system is successful in recovering the deformation at the
at the amplitude of interest. Here, the effective stiffness of the isolation end of ground motions for both LD and SD motions. Under the SD
system under the LD and SD ground motions is calculated using the motions, SLRB experienced a mean RID of 10 mm, which is increased
following expression [58]: up to 14 mm under the LD motions. The recentering ability of the SMA
wires used in the SLRB system resulted in the near-perfect restoring
Table 5 characteristics. On the other hand, the LRB system experienced 28 mm
Mean operational characteristics of different isolation bearings. RID under the SD motions, which is increased by more than two times
Ground Motions Isolation KH,eff (kN/ E (kN.m) γmax (%) RID (mm)
under the LD motions. Compared to LRB, the FPS system experienced
System mm) 32% and 56% less mean RID under the SD and LD motions, respec-
tively.
Short-Duration LRB 1 430 100 28 The behavior of the three isolation systems under SD and LD ground
SLRB 1.8 337 82 10
FPS 1.21 401 88 19
motions can be compared by looking at Fig. 12 which compares the
variation of bearing displacement with time for ground motion pair #7
Long-Duration LRB 1.56 554 78 65
(SD and LD) along with their MID and RID. Irrespective of the isolation
SLRB 2.38 422 52 14
FPS 2.17 502 77 28 system used, the SD motions resulted in higher MID as compared to the
LD motions. This phenomenon can be attributed to the differences in

9
A.L. Hassan and A.M. Billah Engineering Structures 222 (2020) 111129

Fig. 10. Maximum isolator displacement (MID) for bridges isolated with different isolation bearings under LD and SD ground motions (a) LRB, (b) SLRB, and (c) FPS.

the bearings’ effective horizontal stiffness (Table 5) under the SD and SLRB system under the ground motion pair #18 where both LD and SD
LD motions. Under the LD motions, the bearings exhibited a higher motions have the same magnitude (Mw = 7.2) but different significant
effective horizontal stiffness which in turn resulted in lower MIDs as durations of 27 s and 17 s, respectively. It is found that, under the LD
compared to the SD motions. An opposite trend is observed for RIDs motion, the RID for the SLRB is 14.4 mm which is only 2.6 mm under
which are larger under the LD motions as compared to the SD motions. the SD motion. Similar observations are found for the FPS system where
For the same ground motion, the LRB sustained a RID of 41 mm under the ground motion duration is found to have a significant impact on the
the LD motion which is only 3 mm and 11 mm for the SLRB and FPS, isolation bearing restoring capacity.
respectively. Similarly, under the SD motion, the LRB experienced the
largest RID (22 mm) which is 95% and 82% higher than that of SLRB
6.5. Isolator energy dissipation
and FPS, respectively. The results indicate that LRB may not be suitable
isolation bearing option when subjected to long-duration motions. On
The energy dissipation capacity of the isolation bearing indicates
the other hand, both FPS and SLRB can be an effective option in seismic
the isolator’s ability to dissipate a large amount of energy imparted by
zones susceptible to long-duration motions. These systems can effec-
the ground motion. Increased duration of ground motion indicates that
tively prevent the maximum displacement of the bearing, reduce the
the structure will be subjected to a higher number of displacement
residual deformation and thus the collapse probability of the bridge
cycles thus experience large energy demand, which can introduce sig-
during a long-duration motion can be avoided.
nificant structural damage. Fig. 14 compares the hysteretic response of
Fig. 13 shows the relationship between the ground motion duration,
the three isolation bearings under the LD and SD ground motion pair
magnitude, and the accumulated residual displacement in three dif-
#10. The total energy dissipation is obtained by calculating the area
ferent isolation systems under long-duration motions. From Figs. 11 and
enclosed by the hysteresis loop as shown in Fig. 14. Table 5 also sum-
12, it has been observed that the long-duration motions tend to increase
marizes the mean isolator energy dissipation of the three systems under
the residual displacement demand in the isolation bearings as opposed
the SD and LD ground motion suite. From these results, it can be ob-
to the MID which tends to decrease under long-duration motions
served that, under LD motions, the maximum values of isolator dis-
(Fig. 10). This can be attributed to the fact that the LD motions impose a
placement are generally about the same or smaller as compared to
large number of inelastic displacement cycles as compared to SD mo-
spectrally equivalent SD motions. On the other hand, cumulative values
tions, which reduce the re-centering capability of the isolation bearings.
such as the cumulative energy dissipation of the isolation system are
From Fig. 13 it can be seen that, as the ground motion duration and
larger under the LD motions as compared to the SD counterpart. This
magnitude increase, the RID also increases. However, a large magni-
outcome is due to the increased number of inelastic cycles imposed by
tude does not affect the RID if the duration is smaller. For example,
long-duration motion records that significantly increase the cumulative
when the LRB is subjected to the Mw 8.8 motion with a significant
energy dissipated. From Table 5 it can be observed that, under both LD
duration of 31 s (LD#9), the RID is 45 mm. On the other hand, when the
and SD motions, the LRB system has the highest energy dissipation
significant duration is increased to 53 s under Mw 8.8 motion (LD#8),
capacity as compared to the FPS and SLRB systems. A similar ob-
the RID is increased by 200% to 135 mm. It is to be noted that both
servation can be made from Fig. 14, which shows a fatter hysteretic
LD#8 and LD#9 are the recorded ground motions from the 2010 Maule,
curve for the LRB system indicating larger energy dissipation. The LRB
Chile earthquake recorded at different stations. This indicates the
system dissipates energy via the lead core deformation whereas the FPS
ground motion duration significantly affects the restoring capacity of
system dissipates hysteretic energy via friction. In the SLRB system, the
the isolation bearing which can significantly affect the post-earthquake
SMA wires act as supplemental dampers which could result in increased
functionality of the bridge. An additional comparison is made for the
energy dissipation. However, due to the re-centering ability of the SMA

Fig. 11. Residual isolator displacement (RID) for bridges isolated with different isolation bearings under LD and SD ground motions (a) LRB, (b) SLRB, and (c) FPS.

10
A.L. Hassan and A.M. Billah
Fig. 12. Time variation of lateral displacement of (a)LRB, (b) SLRB and (c) FPS under long duration and spectrally equivalent short duration motion pair #7.
wires in the SLRB system, the resulting hysteresis curve is thinner than
the LRB and FPS system, which resulted in the reduction of total energy
dissipated by the SLRB system. Moreover, due to increased effective
horizontal stiffness of the SLRB system (see Table 5) resulting from the
inclusion of SMA wire, the force demand on the isolation bearings in-
creased thus increasing the pier base shear as observed in the previous
section.
7. Conclusions
Observation from recent earthquakes and the damage of structures
have revealed the susceptibility of the existing structures to long-
duration ground motions. However, current design codes, as well as
different performance-based design methodologies, are typically based
on response and damage measures calibrated to short-duration non-
pulse type earthquakes. The current seismic isolation bearing design
guidelines are not adequately equipped to incorporate the effects of
longer duration in the design process. Hence, this study aimed to pro-
vide new insight into the response of seismically isolated bridges
equipped with three different isolation bearings subjected to long-
duration motions and to contrast the demands under spectrally
equivalent short-duration motions. In general, it has been observed that
the ground motion duration affects the seismic response of the isolated
bridges. This can be attributed to a large number of cycles that result in
11
Engineering Structures 222 (2020) 111129
the cumulative damage effects from long-duration seismic events and
imparts more energy to the structure. Based on the study the following
conclusions are drawn:
• It is possible to achieve a similar isolation period for bridges with
three different isolation systems. Although the performances of the
isolated bridges are comparable under the SD ground motions, their
response significantly varied under the LD motions.
• Long-duration motions cause more damage to different bridge
components as compared to the short-duration motions in-terms of
higher deck acceleration, pier base shear, and residual isolator dis-
placement. Although differences that should be captured exist be-
tween the response under two sets of ground motions, the overall
benefit of seismic performance improvement of the isolated bridges
is comparably and successfully achieved over non-isolated bridges.
• Under LD motions, LRB systems are susceptible to significantly large
residual displacement which might eventually disrupt the post-
earthquake functionality of the bridge. However, the FPS system
showed promising performance in reducing the RID and im-
plementing the SLRB system could effectively eliminate the RID is-
sues associated with conventional LRB bearings.
• Under SD motions, the bridge piers experienced similar base shear
irrespective of the isolation systems. When subjected to LD motions,
SLRB experienced 7.5% and 22% higher PNBS as compared to the
A.L. Hassan and A.M. Billah
Fig. 13. Relationship of residual isolator displacement (RID) with significant duration and magnitude under LD ground motion.
FPS and LRB bearings, respectively. On the other hand, the LRB
system experienced 30% higher deck acceleration under LD motions
whereas this increase is 25% and 32% for SLRB and FPS system,
respectively compared to SD motions.
• The maximum displacement demand on the isolation bearings
varied from 10% (LRB and FPS) to 18% (LRB and SLRB) under the
SD motions, which are increased up to 35% under the LD motions.
Compared to the LRB, the FPS system experienced 32% and 56% less
mean RID under SD and LD motions, respectively. This is a clear
indication that considering the duration effect is more important in
the design of the isolators and less on assessing the effectiveness of
seismic isolation on bridge substructure protection.
• As the ground motion duration and magnitude increases, the RID
also increases. The residual displacement demand in the isolation
bearing increases with the elongation of ground motion duration.
However, a large magnitude does not affect the RID if the duration is
smaller. Ground motions having similar magnitude with different
significant durations can significantly affect the isolator as well as
the bridge response.
• Based on the analytical results of this study, the following con-
siderations should be taken into account for practical design of
isolated bridges under long-duration ground motions:
a) Design targets should be set at critical values of residual deforma-
tion as LD motions tend to increase the residual deformation sig-
nificantly. Based on this study, it is recommended to consider lower
residual displacement capacity for the isolation bearings when
subjected to long-duration motion or consider higher spectral ac-
celeration demands.
b) It is recommended to increase the energy dissipation capacity of the
bridge system either by increasing the number of isolation bearings
or adding supplementary energy dissipation devices.
c) Since the long-duration motions increase the pier base shear de-
mand, a site and structural system-specific adjustment should be
implemented for the base shear demand for designing bridge piers in
locations where long-duration ground motions are expected.
12
Engineering Structures 222 (2020) 111129
d) Testing requirements should be imposed to evaluate the damping
mechanism and energy dissipation capacity under a large number of
displacement reversals before implementing in sites susceptible to
long-duration motions.
The current study did not consider the effect of heating and sub-
sequent strength degradation on the analytical model of the isolation
bearings. However, that requires testing of isolation bearings under
long-duration motions to include the effects of the large number of
displacement cycles along with longer durations. Further studies need
to be conducted considering different bridge models with uncertainty in
geometry/material properties as well as isolation bearing character-
istics at material fiber levels. Future studies should consider near-fault
ground motions with velocity pulse to evaluate the comparative re-
sponse of isolated bridges under long-duration motions and pulse-like
near-fault motions.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Acknowledgment
The Natural Sciences and Engineering Research Council (NSERC) of
Canada through the Discovery Grant and Lakehead University SRC
Research Development Fund (RDF) supported this study. The financial
supports are greatly appreciated.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.engstruct.2020.111129.
A.L. Hassan and A.M. Billah
Fig. 14. Hysteretic response of (a) LRB, (b) SLRB and (c) FPS under long duration and spectrally equivalent short duration motion pair #10.
References
[1] Jangid RS. Seismic response of isolated bridges. J Bridge Eng 2004;9(2).
[2] Tsopelas P, Constantinou MC, Okamoto S, Fujii S, Ozaki D. Experimental study of
bridge seismic sliding isolation systems. Eng Struct 1996;18(4).
[3] Tongaonkar NP, Jangid RS. Seismic response of isolated bridges with soil-structure
interaction. Soil Dyn Earthq Eng 2003;23:287–302.
[4] Van Engelen NC, Osgooei PM, Tait MJ, Konstantinidis D. Partially Bonded Fiber-
Reinforced Elastomeric Isolators (PB-FREIs). Struct Control Health Monitor
2015;22(3):417–32. https://doi.org/10.1002/stc.1682.
[5] Billah AHMM, Todorov B. Effects of subfreezing temperature on the seismic re-
sponse of lead rubber bearing isolated bridge. Soil Dynamics and Earthquake
Engineering 2019;126. https://doi.org/10.1016/j.soildyn.2019.105814.
[6] Warn GP, Whittaker AS. Performance estimates in seismically isolated bridge
structures. Eng Struct 2004;26:1261–78.
[7] Alam MS, Bhuiyan MAR, Billah AHMM. Seismic fragility assessment of SMA-bar
restrained multi-span continuous highway bridge isolated by different laminated
rubber bearings in medium to strong seismic risk zones. Bull Earthq Eng 2012;10.
[8] Ozdemir G, Avsar O, Bahyan B. Change in response of bridges isolated with LRBs
due to lead core heating. Soil Dyn Earthq Eng 2011;31.
[9] Van Engelen NC, Konstantinidis D, Tait MJ. Structural and nonstructural perfor-
mance of a seismically isolated building using stable unbonded fiber-reinforced
elastomeric isolators. Earthquake Engng Struct Dyn 2016;45:421–39.
[10] Marafi NA, Eberhard MO, Berman JW, Wirth EA, Frankel AD. Impacts of Simulated
M9 Cascadia Subduction Zone Motions on Idealized Systems. Earthquake Spectra
2019;35(3):1261–87.
[11] Kempton JJ, Stewart JP. Prediction equations for significant duration of earthquake
ground motions considering site and near-source effects. Earth Spectra
2006;22(4):985–1013.
[12] Hancock J, Bommer JJ. A state-of-knowledge review of the influence of strong-
13
Engineering Structures 222 (2020) 111129
motion duration on structural damage. Earthquake Spectra 2006;22(3):827–45.
[13] Han Q, Xiuli D, Jingbo L, Zhongxian L, Lyun L, Jianfeng Z. Seismic damage of
highway bridges during the 2008 Wenchuan earthquake. Earthquake Eng Eng
Vibration 2009;8:263–73.
[14] Xiang N, Li J. Seismic Performance of Highway Bridges with Different Transverse
Unseating-Prevention Devices. J Bridge Eng ASCE 2016;21(9).
[15] Chandramohan R, Baker JW, Deierlein GG. Quantifying the influence of ground
motion duration on structural collapse capacity using spectrally equivalent records.
Earthquake Spectra 2016;32(2):927–50.
[16] Lopez A, Dusicka P, Bazaez R. Performance of seismically substandard bridge re-
inforced concrete columns subjected to subduction and crustal earthquakes. Eng
Struct 2020;207:110216.
[17] Pan Y, Ventura CE, Finn L. Effects of Ground Motion Duration on the Seismic
Performance and Collapse Rate of Light-Frame Wood Houses. J Struct Eng
2018;144(8).
[18] Hammad A, Moustafa MA. Modeling sensitivity analysis of special concentrically
braced frames under short and long-duration ground motions. Soil Dyn Earthquake
Eng 2020;128:105867.
[19] Kabir MR, Billah AHMM, Alam MS. Seismic fragility assessment of a multi-span RC
bridge in Bangladesh considering near-fault, far-field and long-duration ground
motions. Structures 2019;19:333–48.
[20] Shen J, Tsai MH, Chang KC, Lee GC. Performance of a Seismically Isolated Bridge
under Near-Fault Earthquake Ground Motions. J Struct Eng 2004;130(6):861–8.
[21] Dicleli M. Performance of seismic-isolated bridges in relation to near-fault ground-
motion and isolator characteristics. Earthquake Spectra 2006;22(4):887–907.
[22] Ozbulut OE, Hurlebaus S. Optimal design of superelastic-friction base isolators for
seismic protection of highway bridges against near-field earthquakes. Earthquake
Eng Struct Dyn 2011;40:273–91.
[23] Eröz M, DesRoches R. A Comparative Assessment of Sliding and Elastomeric Seismic
Isolation in a Typical Multi-Span Bridge. J Earthquake Eng 2013;17(5):637–57.
[24] Ou YC, Song J, Wang PH, Adidharma L, Chang KC, Lee GC. Ground motion duration
A.L. Hassan and A.M. Billah
effects on hysteretic behavior of reinforced concrete bridge columns. ASCE Journal
of Structural Engineering 2013;140(3).
[25] Bommer JJ, Martinez-Pereira A. The effective duration of earthquake Strong mo-
tion. J Earthquake Eng 1999;3(2):127–72.
[26] Arias A. A measure of earthquake intensity. In: Hansen RJ, editor. Seismic design
for nuclear power plants. Cambridge, MA: MIT Press; 1970. p. 438–83.
[27] Trifunac MD, Brady AG. A study on the duration of strong earthquake ground
motion. Bull Seismol Soc Am 1975;65(3):581–626.
[28] Page RA, Boore DM, Joyner WB, Coulter HW. Ground motion values for use in
seismic design of the Trans-Alaska pipeline system. US Geological Survey Circular
1972;672.
[29] Reed JW, Kassawara RP. A criterion for determining exceedance of the operating
basis earthquake. Nucl Eng Des 1990;123(2–3):387–96.
[30] Cosenza E, Manfredi G. The improvement of the seismic-resistant design for existing
and new structures using damage criteria. Seismic Design Methodologies for the
Next Generation of Codes, H. Krawinkler, P. Fajfar eds 1997; Balkema, Rotterdam:
119–130.
[31] Bravo-Haro MA, Elghazouli AY. Influence of earthquake duration on the response of
steel moment frames. Soil Dyn and Earthquake Engineering 2018;115:634–51.
[32] Barbosa AR, Ribeiro FLA, Neves LAC. Influence of earthquake ground-motion
duration on damage estimation: application to steel moment resisting frames.
Earthquake Eng Struct Dyn 2017;46:27–49.
[33] Baker JW, Cornell CA. Spectral shape, epsilon and record selection. Earthquake Eng
Struct Dyn 2006;35(9):1077–95.
[34] Ancheta TD, Darragh RB, Stewart JP, Seyhan E, Silva WJ, Chiou BSJ, et al. PEER
NGA-West2 Database, PEER 2013/03, Pacific Earthquake Engineering Research
Center. Berkeley, California: University of California; 2011.
[35] Center for Engineering Strong Motion Data (CESMD), 2012. Center for Engineering
Strong Motion Data. [online] Available at: < http://www.strongmotioncenter.org/
> [Accessed 1 June 2019].
[36] Canadian Standard Association – CSA. CAN/CSA-S6-14: Canadian highway bridge
design code. National Research Council of Canada, Ottawa; ON.
[37] Liu H, Wang X, Liu J. The shaking table test of an SMA strands-composite bearing.
Earthq Eng Vib 2008;28.
[38] Dezfuli FH, Alam MS. Shape memory alloy wire-based smart natural rubber
bearing. Smart Mater Struct 2013;22:045013.
[39] Dezfuli FH, Alam MS. Smart Lead Rubber Bearings Equipped with Ferrous Shape
Memory Alloy Wires for Seismically Isolating Highway Bridges. J Earthquake Eng
2018;22(6):1042–67.
[40] AASHTO. Guide specifications for seismic isolation design. Fourth ed. Washington,
DC, USA: American Association of State Highway and Transportation Officials
(AASHTO); 2014.
[41] Buckle IG, Moustafa AA, Monzon EV. Seismic isolation design examples of highway
bridges. Final Rep. NCHRP 20-7/Task 262 (M2) 2011. National Cooperative for
Highway Research Program (NCHRP), Transportation Research Board,
Washington, DC.
[42] Dezfuli FH, Alam MS. Hysteresis model of shape memory alloy wire-based lami-
nated rubber bearing under compression and unidirectional shear loadings. Smart
Mater Struct 2015;24(6):065022.
[43] SeismoSoft. SeismoStruct - a computer program for static and dynamic nonlinear
analysis of framed structures available from: www.seismosoft.com; 2020.
14
Engineering Structures 222 (2020) 111129
[44] Mander JB, Priestley MJN, Park R. Theoretical stress-strain model for confined
concrete. J Struct Eng ASCE 1988;114(8).
[45] Menegotto M, Pinto PE. Method of analysis for cyclically loaded R.C. plane frames
including changes in geometry and non-elastic behaviour of elements under com-
bined normal force and bending. Symposium on the Resistance and Ultimate
Deformability of Structures Acted on by Well Defined Repeated Loads. International
Association for Bridge and Structural Engineering, Zurich, Switzerland, 1973;
15-22.
[46] Fragiadakis M, Papadrakakis M. Modeling, analysis and reliability of seismically
excited structures: computational issues. Int J Comput Methods 2008:483–511.
[47] Billah AHMM, Kabir MR, Alam MS. Comparative collapse performance assessment
of bridge pier under near-fault and long-duration ground motions. 16th world
conference on earthquake engineering, Santiago, Chile. January 2017.
[48] Caltrans. Seismic design criteria version 1.7. Sacramento, CA: California
Department of Transportation; 2013.
[49] Aviram A, Mackie K, Stojadinovic B. Guidelines for nonlinear analysis of bridge
structures in California. Technical Report 2008/03. Pacific Earthquake Engineering
Research Center, University of California, Berkeley.
[50] Muthukumar S, DesRoches R. A Hertz contact model with non-linear damping for
pounding simulation. Earthq Eng Struct Dyn 2006;35.
[51] Constantinou MC, Tsopelas PC, Kasalanati A, Wolff E. Property Modification Factors
for Seismic Isolation Bearings. Report No. MCEER-990012, Multidisciplinary Center
for Earthquake Engineering Research 1999, State University of New York,
Buffalo, NY.
[52] Wei B, Li C, Jia X, He X, Yang M. Effects of shear keys on seismic performance of an
isolation system. Smart Structures and Systems 2019;24(3):345–60.
[53] Wen YK. Method for random vibration of hysteretic systems. J Eng Mech ASCE
1976;102(2):249–63.
[54] Park YJ, Wen YK, Ang H-S. Random Vibration of Hysteretic Systems under Bi-
Directional Ground Motions. Earthq Eng Struct Dyn 1986;14(4):543–57.
[55] Constantinou MC, Adnane MA. Dynamics of Soil-Base-Isolated Structure Systems:
Evaluation of Two Models for Yielding Systems. Report to the National Science
Foundation 1987, Department of Civil Engineering, Drexel University,
Philadelphia, PA.
[56] Billah AHMM, Alam MS. Probabilistic seismic risk assessment of concrete bridge
piers reinforced with different types of shape memory alloys. Eng Struct
2018;162:97–108.
[57] Billah AHMM, Alam MS. Performance-based prioritization for seismic retrofitting of
reinforced concrete bridge bent. Struct Infrastruct Eng 2014;10(8):929–49.
[58] Constantinou MC, Whittaker AS, Kalpakidis Y, Fenz DM, Warn GP. Performance of
seismic isolation hardware under service and seismic loading. Multidisciplinary
Center for Earthquake Engineering Research, Buffalo, NY, 2007. Report No.
MCEER-07-0012.
[59] Berton S, Infanti S, Castellano MG, Hikisaka H. Self-centring capacity of seismic
isolation systems. Struct Control Health Monitor 2006;14(6):895–914.
[60] Wei B, Zhuo Y, Li C, and Yang M. Parameter Optimization of a Vertical Spring-
Viscous Damper-Coulomb Friction System. Shock and Vibration, 2019, 5764946, 19
pages.
[61] Li S, Wei B, Zuo C, and He X. A Numerical Investigation on Scaling Rolling Friction
Effects in Shaking Table Model Tests. Shock and Vibration, 2019, 7473031, 14
pages.