Engineering Structures 162 (2018) 97–108

Contents lists available at ScienceDirect

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

Probabilistic seismic risk assessment of concrete bridge piers reinforced with T

different types of shape memory alloys
⁎
A.H.M. Muntasir Billah , M. Shahria Alam
School of Engineering, The University of British Columbia, Kelowna, BC V1V1V7, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: Shape memory alloy (SMA) has emerged as an alternative to conventional steel reinforcement for improving the
Shape memory alloy seismic performance of bridges during an extreme earthquake. This paper presents the probabilistic seismic risk
Damage states assessment of concrete bridge piers reinforced with different types of SMA (e.g. Ni-Ti, Cu-Al-Mn, and Fe-based)
Incremental dynamic analysis rebars. To achieve this objective, the bridge piers are designed following a performance-based approach. Ground
Residual drift
motions with different probable earthquake hazard scenarios at the site of the bridge piers are considered.
Fragility curve
Probabilistic seismic demand models are generated using the response parameters obtained from incremental
dynamic analysis. Considering maximum drift and residual drift as demand parameters, fragility curves are
developed for five different SMA-RC bridge piers. Finally, seismic hazard curves are generated in order to
compare the mean annual rate of exceedance of different damage states of different bridge piers. It is observed
that all the bridge piers perform according to the design objective, and the performance of SMA-RC piers is
significantly affected by the type of SMA used. The results show that all the SMA-RC piers have very low
probability of collapse at maximum considered earthquake level. It is found that the bridge pier reinforced with
FeNCATB-SMA (SMA-3) performed better as compared to the other SMA-RC piers.

1. Introduction deformation as an indicator for defining the overall seismic perfor-

Current seismic design guidelines, followed in North America [1,2]
and Europe [3], allow bridges other than life line bridges to undergo
large inelastic deformation while maintaining the load carrying capa-
city without being completely collapsed during a design level earth-
quake. However, past experiences (Kobe 1995, Northridge 1994) have
shown that bridges undergoing large lateral drift are prone to large
residual deformation which renders the bridges to be unusable and
require major rehabilitation or replacement. In order to maintain the
structural integrity and functionality of a bridge after an earthquake, it
is necessary that the bridge components avoid excessive residual de-
formation or permanent damage [4]. Bridge pier is one of the most
critical components of a bridge since the overall seismic response of a
bridge is largely dependent on the response of the piers. The extent of
residual or permanent deformation sustained by the bridge piers pre-
scribes the likelihood of allowing traffic over the bridge and dictates the
amount of repair works and expected loss. As an example, after the
devastating 1995 Kobe earthquake, Japan demolished more than 100
reinforced concrete bridge piers as those piers experienced a residual
drift ratio over 1.75 percent [4]. Evidences from recent earthquakes and
field reports demonstrated the importance of considering residual
mance of a structure. Observations from recent earthquakes (Kobe
1995, Northridge 1994) and a desire to develop innovative structural
systems with improved post-earthquake functionality have motivated
researchers to pioneer and test different novel structural systems. For
example, to reduce the residual displacement of bridge piers, re-
searchers have recommended innovative unbonded post-tensioned RC
bridge columns [5,6] and Shape Memory Alloy (SMA) reinforced con-
crete (RC) bridge piers [7,8].
Over the last few years, researchers have experimentally and nu-
merically investigated the potential application of shape memory alloys
in civil engineering applications and found promising results
[7–9,47,48]. The research outcomes have motivated bridge owners to
apply SMA in bridge construction. Washington State DOT has decided
to incorporate SMA in bridge pier construction part of the SR 99
Alaskan Way Viaduct Replacement program based on experimental
results of Nakashoji and Saiidi [48]. However, the bridge was not de-
signed using a performance-based approach rather following the con-
ventional design method with an aim to increase post-earthquake
functionality of the bridge. Cruz and Saiidi [9] tested a four span bridge
with Ni-Ti SMA in the plastic hinge region of the bridge pier and found
improved performance of the bridge in terms of residual drift and

⁎
Corresponding author at: Bridge Engineer, Parsons, Burnaby, BC V5H4M2, Canada.
E-mail addresses: muntasir.billah@alumni.ubc.ca (A.H.M.M. Billah), shahria.alam@ubc.ca (M.S. Alam).

https://doi.org/10.1016/j.engstruct.2018.02.034
Received 5 June 2017; Received in revised form 29 January 2018; Accepted 8 February 2018
Available online 16 February 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
A.H.M.M. Billah, M.S. Alam
energy dissipation. Billah and Alam [8] assessed the seismic vulner-
ability of a SMA reinforced bridge pier and compared with that of
conventional bridge pier. However, all the previous studies were fo-
cused on the application of Ni-Ti SMA. More recently, Billah and Alam
[10,11] developed performance-based damage states and a perfor-
mance-based seismic design guideline for bridge piers reinforced with
different types of superelastic SMA in the plastic hinge region. While
the previous studies proved the potential of using this smart material in
bridge piers and proposed some design guidelines, adoption of these
guidelines and successful implementation require a complete perfor-
mance-based evaluation of this structural system in light of perfor-
mance-based earthquake engineering (PBEE). To this end, it is neces-
sary to investigate the ability of such novel structural system in
reducing the failure probability as well as the annual rate of exceeding
some structural demand parameters given an earthquake scenario.
The objective of this paper is to perform fragility based probabilistic
seismic risk assessment of concrete bridge piers reinforced with dif-
ferent types of SMA rebars in the plastic hinge region. No baseline
comparison is provided with a typical reinforced concrete pier as the
focus of this study is to compare the seismic performance of concrete
bridge piers reinforced with different types of SMA rebars. Fig. 1 il-
lustrates the methodology adopted in this study. First, the bridge piers
are designed following the performance-based design guideline pro-
posed by Billah and Alam [11]. Later, a detailed description of the finite
element model is provided to elucidate the details of bridge pier
models. Instead of using code-specified design level earthquakes, this
study considered three different earthquake scenarios which resembles
the regional seismicity of Vancouver, British Columbia (BC), where the
bridge is located. The performance and hazard assessment is conducted
by considering shallow crustal, mega-thrust interface, and deep in-slab
earthquake events [12]. Next, incremental dynamic analysis (IDA) [13]
are conducted on each SMA-RC bridge pier model using 30 selected
ground motions scaled to the conditional mean spectra of crustal, in-
slab and interface earthquakes. The performance parameters of interest,
which are maximum and residual drift in this study, are recorded for
each motion. Next, the seismic performances of five different SMA-RC
bridge piers are evaluated and compared using fragility curves. The
fragility curves are developed using the Probabilistic Seismic Demand
Model (PSDM). Finally, a probabilistic risk assessment is conducted to
evaluate the mean annual frequency of exceeding different damage
levels in terms of the selected demand parameters.
Previously, Billah and Alam [11] developed a performance based
seismic design guideline for Shape Memory Alloy (SMA) reinforced
concrete bridge pier. This current paper focuses on probabilistic seismic
risk assessment of different SMA-RC bridge pier designed per the pre-
vious paper. The earlier paper [11] proposed the design guideline and
the current paper shows how the designed pier performs under different
earthquake scenarios. The current paper also focuses on probabilistic
risk assessment to evaluate the mean annual frequency of exceeding
different damage levels in terms of the selected demand parameters.
2. Probabilistic seismic performance assessment
A commonly used method for probabilistic seismic performance
assessment is the Pacific Earthquake Engineering Research (PEER)
Centre PBEE methodology [14] which attempts to address structural
performance in terms of life safety, capital losses and functional losses
[15]. This PBEE methodology is composed of hazard analysis, structural
analysis, damage analysis, and loss analysis. However, most of the ap-
plications of PBEE have been focused on buildings and few of them
focused on bridge structures [5,15]. Moreover, no study has been
conducted to date for probabilistic seismic performance assessment of
SMA reinforced bridge piers. Previous studies have developed perfor-
mance-based seismic design guidelines for SMA-RC bridge piers [11] or
evaluated seismic performance of SMA-RC piers [9]. This study is in-
tended to elucidate the potential benefit and compare the performance
98
Engineering Structures 162 (2018) 97–108
of different SMA-RC bridge piers in light of PBEE. This study conducted
three steps of PBEE involving hazard, structural and damage analyses.
However, the loss analysis was not performed because of limited in-
formation regarding the cost of different types of SMA considered in
this study. This study developed fragility curves and seismic hazard
curves for different SMA-RC bridge piers considering maximum and
residual drift as engineering demand parameters. The developed fra-
gility curves express the probability of reaching or exceeding certain
damage states corresponding to a certain intensity of ground motion.
The hazard curves relate the mean annual rate of exceeding certain
damage states.
Instead of proposing a new methodology for the fragility assess-
ment, this study offers critical insight on the performance-based eva-
luation of SMA-RC bridge piers using fragility curves. In this assessment
different types of SMAs and uncertainties in the seismic hazard of the
site are considered. Details of different methods of fragility curve de-
velopment can be found in [16]. In this study, the fragility curves are
developed using a probabilistic seismic demand model (PSDM) and
limit state model. The PSDM which relates the median demand to the
intensity measure (IM) is developed using the results obtained from IDA
and the power law function [17]. The PSDM provides a logarithmic
correlation between median demand and the selected IM:
EDP = a (IM )b (1)
In the log transformed space, Eq. (1) can be expressed as
ln(EDP ) = ln(a) + bln(IM ) (2)
where a and b are unknown coefficients which can be estimated from a
regression analysis of the response data collected from IDA. Effective-
ness of a demand model is determined by the ability of evaluating Eq.
(2) in a closed form. In order to accomplish this task, it is assumed that
the EDPs follow log-normal distributions. The dispersion (βEDP|IM) ac-
counting for the uncertainty in the relation which is conditioned upon
the IM, is estimated using Eq. (3) [18]:
N
∑i = 1 (ln(EDP )−ln(aIM b))2
βEDP / IM =
N −2 (3)
where N is the number of simulations. With the probabilistic seismic
demand models and the limit states corresponding to various damage
states, it is now possible to generate fragilities (i.e. the conditional
probability of reaching a certain damage state for a given IM) using Eq.
(4) [19].
⎡ ln(IM )−ln(IMn ) ⎤
P [LS / IM ] = Φ ⎢ ⎥
βcomp (4)
⎣ ⎦
Φ [.] is the standard normal cumulative distribution function and
ln(Sc )−ln(a)
ln(IMn ) =
b (5)
ln(IMn) is defined as the median value of the intensity measure for the
chosen damage state, a and b are the regression coefficients of the
PSDMs, and the dispersion component is presented in Eq. (6) [19].
βEDP / IM + βc2
βcomp =
b (6)
where Sc is the median and βc is the dispersion value for the damage
states of different components of a bridge.
By combining the seismic responses obtained from IDA, in terms of
maximum and residual drift, with the seismic hazard curve, it is pos-
sible to calculate the annual rate of exceeding various levels of demand
for each EDP monitored. Using the uniform seismic hazard curve for the
site under consideration, and maximum and residual drift responses
obtained from IDA, the maximum and residual drift hazard of the SMA-
RC piers are calculated based on the convolution integral presented by
Deierlein et al. [20]
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

Design and modeling of SMA-RC bridge piers

Design of bridge pier FE modeling of bridge pier

Selection of ground motions

Spectral Acceleration
Spectral Acceleration

UHS
CMS-Crustal
CMS-Inslab
CMS-Interface
Vibration Period (s) Vibration Period (s)
Generation of CMS Selection and scaling of ground motions

Incremental dynamic analysis

1 0.1
Maximum deformation
Displacement (m)

0.5 0.05
Acceleration (g)

0 0
Residual deformation
-0.5 0.05
-0.05

-1 -0.1
0 20 40 0 10 20 30 40
Time (sec) Time (sec)
Ground motion scaled to different Maximum and residual drift
intensities

Probabilistic seismic risk assessment

Annual rate of
exceedance

DS-1
P [DSIPGA]

DS-2
LN (IM)

DS-3
DS-4

y = 1.03x + 0.36
R² = 0.80
LN (EDP) PGA (g) EDP
PSDM Fragility Curves Hazard Curves

Fig. 1. Flowchart of the methodology for seismic risk assessment of SMA-RC bridge piers.

λEDP (edp) = ∫ P (EDP > edp |IM )|dv (IM )| (7)

In this equation, IM denotes the intensity measure of the ground
motion, EDP refers to the engineering demand parameter (maximum
and residual drift), λEDP(edp) represents the mean annual frequency of
exceeding a predefined engineering demand parameter, edp.
3. Design of SMA-RC bridge piers
In this study five concrete bridge piers reinforced with five different
SMAs are designed following the performance-based design guidelines
proposed by Billah and Alam [11]. The bridge piers are assumed to be
located at Vancouver, BC with the site soil class-C (stiff soil). The bridge
pier is a part of three span concrete girder bridge. A constant mass of 85
ton was considered at the top which represents the seismic mass con-
tribution from the superstructure on the pier. The corresponding design
spectrum is selected, with a 2% probability of exceedance in 50 years
that corresponds to a return period of 2475 years, according to the
CHBDC-2014 [1]. The bridge is considered to be a lifeline bridge ac-
cording to the bridge classification described in CHBDC-2014 [1]. For
the selected seismic hazard level (2% in 50 years), the bridge should be
operational (repairable damage) with limited service to meet the per-
formance requirements. Since this design method starts with selecting a
target residual drift, to meet the performance objectives and develop a
comparable design of five different SMA-RC bridge piers, a target re-
sidual drift of 0.6% is selected. This target residual drift is selected
based on the performance based damage states proposed by Billah and
Alam [10]. No other design guidelines except the Japanese code for

99
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

Table 1
Properties of different types of SMA.

Alloy SMA ID E (GPa) εs (%) fy (MPa) fp1 (MPa) fT1 (MPa) fT2 (MPa) Ref

NiTi45 SMA-1 62.5 6 401.0 510 370 130 [26]

NiTi45 SMA-2 68 8 435.0 535.0 335 170 [43]
FeNCATB SMA-3 46.9 13.5 750 1200 300 200 [44]
CuAlMn SMA-4 28 9 210.0 275.0 200 150 [45]
FeMnAlNi SMA-5 98.4 6.13 320.00 442.5 210.8 122 [46]

fy (austenite to martensite starting stress); fP1 (austenite to martensite finishing stress); fT1 (martensite to austenite starting stress); fT2 (martensite to austenite finishing stress), εs
(maximum superelastic strain); and E (modulus of elasticity).

Table 2 LP P ⎞ L
Material properties for SMA-RC bridge pier. = 1.05 + ⎛⎜0.25 ⎟ + ⎛0.08 ⎞ + (0.0002f y − SMA )−(0.16ρl )−(0.019f c′ )
d f ′c Ag ⎝ d ⎠
⎝ ⎠
Material Property
−(0.24ρs ) (8)
Concrete Compressive Strength (MPa) 42.4
Corresponding strain 0.0029 where Lp is the plastic hinge length, d is the diameter of the pier, L/d is
Tensile strength (MPa) 3.5 the aspect ratio, P/fc′Ag is the axial load ratio, ρl = longitudinal re-
Elastic modulus (GPa) 23.1 inforcement ratio, ρs = transverse reinforcement ratio, fy-SMA = yield
Steel Elastic modulus (GPa) 200 strength of SMA rebar and fc′ = concrete compressive strength. This
Yield stress (MPa) 475 equation showed reasonable accuracy in predicting the plastic hinge
Ultimate stress (MPa) 692 length measured from experimental investigations.
Ultimate strain 0.14
Plateau strain 0.016
4. Finite element modeling of bridge piers

highway bridge design provide any residual drift limit of bridge piers.
Lee and Billington [5] considered 1% residual drift to be large enough
for a bridge replacement. Existing literature [7,8] also suggest that a
residual drift larger than 1% can be characterized as a collapse damage
state. The height and diameter of all the bridge piers are assumed to be
5 m and 1 m, respectively. Five different SMAs having different com-
binations of alloys and mechanical properties are selected which are
shown in Table 1. The material properties of concrete and steel re-
inforcement are listed in Table 2 which is based on the experimental
study by Saiidi and Wang [49]. The final design yielded all the bridge
piers to be reinforced with 28 longitudinal SMA rebars of different
diameter in the plastic hinge region and the remaining portion was
reinforced with 28–25 M steel (diameter 25.2 mm) rebar. To meet the
current seismic design requirements, shear reinforcement was provided
using 15 mm spirals at 75 mm pitch. The shear capacity of the pier was
checked using modified compression field theory [21]. The bridge piers
are specified as SMA-RC-1 (reinforced with SMA-1), SMA-RC-2 (re-
inforced with SMA-2), and so on. SMA-RC-1 and SMA-RC-2 are re-
inforced with 28–25 mm SMA-1 and SMA-2 bars, respectively. SMA-RC-
3 is reinforced with 28–22.5 mm SMA-3 bars whereas SMA-RC-4 is
reinforced with 28–35 mm SMA-4 bars, and SMA-RC-5 is reinforced
with 28–30 mm SMA-5 bars. The sizes of the rebars were selected in
such a way that the axial forces developed in the rebar are almost si-
milar. The bridge piers were designed in such a way that they have
comparable moment capacities. From sectional analysis of all five SMA-
RC bridge piers it was found that all the sections have similar initial
stiffness and comparable moment capacities. Since SMA-5 has higher
elastic modulus, SMA-RC-5 showed higher initial stiffness which is
1.78, 1.72, 2.21, and 3.87 times higher than that of SMA-RC-1, SMA-
RC-2, SMA-RC-3, and SMA-RC-4, respectively. Moment-curvature re-
sponse of all the sections revealed that this design process led to com-
parable moment capacities for the five different SMA reinforced bridge
piers. Fig. 2 shows the cross section and elevation of the bridge pier. In
this study, the plastic hinge length of the SMA-RC bridge piers are
calculated using the plastic hinge expression (Eq. (8)) proposed by
Billah and Alam [22]
In this study, the bridge piers are modeled using a fiber based finite
element program Seismostruct [23] to explicitly model the concrete,
SMA and reinforcing steel materials. This program is able to accurately
predict the large displacement and collapse behavior of frame struc-
tures under static and dynamic loading considering both geometric
nonlinearity (P-Δ effect) and material inelasticity [24]. Several re-
searchers [25,26] as well as the authors of this study [8,10] have de-
monstrated the accuracy of the program in predicting the seismic re-
sponse of bridge structures and SMA reinforced concrete members. The
fiber sections of confined and unconfined concrete are simulated using
the Mander et al. [27] (1988) concrete constitutive model. The long-
itudinal reinforcement is modeled using the Menegotto–Pinto [28]
(1973) steel model with Filippou [29] hardening rules. The superelastic
SMA is modeled based on the constitutive relation developed by Aur-
icchio and Sacco [30]. In this study the effect of temperature (Shape
memory effect) was not considered rather the effect of load history and
stress/strain (superelasticity) was considered. The materials are chosen
such that their austenite finish temperature, Af is below the ambient
temperature so that the SMAs will always remain superelastic [26]. The
1-D superelastic SMA model that is considered in this study is shown in
Fig. 3. The parameters used to define the SMA model are fy (austenite to
martensite starting stress); fP1 (austenite to martensite finishing stress);
fT1 (martensite to austenite starting stress); fT2 (martensite to austenite
finishing stress), superelastic plateau strain length or maximum residual
strain, εl; modulus of elasticity, Ea.
In this study, the seismic motion is considered only in the direction
which is parallel to the longitudinal axis of the bridge. To simulate the
boundary condition of the actual pier under seismic excitation, the
column rotation is fixed at the top but movement is free. The numerical
model of the pier consist of inelastic fiber beam/column elements with
couple of springs defined using “zero length link elements” in
Seismostruct. The link elements are used to capture slippage of rebar
inside the coupler and bond-slip of rebar (slip of SMA rebar in concrete)
(Fig. 2c). Mechanical couplers are used to connect SMA with steel re-
bars inside the foundation and above the plastic hinge length [26].
However, the rebar and coupler inside the foundation is not shown in
Fig. 2b for clarity. The slippage of SMA rebar inside the couplerwhich is
represented by introducing a zero-length rotational spring at the
bottom of the column section (Fig. 2c). A zero length rotational spring is

100
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

(a) (b) (c)

75 mm clear cover Steel
28 SMA
S rebar IInelastic beam-column
reinforced element
section

15M spiral at
75 mm pitch SMA
reinforced
section
Zero length bond-slip
spring
Zero llengthh rotational
Z i l
1 mm ĭ
spring

Fig. 2. (a) Cross section, (b) elevation and (c) finite element model of SMA-RC bridge pier.

characterizes the behavior of response curve which are, the yielding

strength Fy (150 kN); the initial stiffness Ky (42,750 kN m); the post
yielding to initial stiffness ratio a (0.0197); the outer loop stiffness
degradation factor β0; and the inner loop stiffness degradation factor
β1. Different parameters required for describing the response curve are
obtained from the moment-rotation relationship of SMA-RC pier. The
stiffness degradation factors β0 and β1 are specified as 0.1 and 0.9,
respectively, to make the unloading stiffness close to the initial stiffness
during the cyclic loading. Moment-rotation relationship of the bar-slip
rotational spring is derived based on the calibration procedures de-
scribed in [31]. The stress-slip relationship of bars inside the coupler
and the details of the splicing can be found in [31]. This study em-
ployed another zero-length inelastic spring to simulate the bond-slip
behavior of SMA rebar in concrete. The bond slip spring is modeled
based on the experimental bond stress-slip relation developed by Billah
and Alam [32]. Using the modified Takeda hysteresis curve, described
by Otani [33] which follows the unloading rules proposed by Emori and
Schonobrich [34], the bond-slip spring was modeled.
Fig. 3. 1D-Superelastic model of SMA [29]
5. Seismic hazard and selection of ground motions
used to This rotational spring represents the effect of rebar slippage in
the mechanical couplers inside the foundation. However, the rebar and For the considered location of the bridge pier, seismic hazard is
coupler inside the foundation is not shown in Fig. 2b for clarity. It calculated using the probabilistic seismic hazard analysis (PSHA).
should be noted that the slippage will be majorly contributed from the Current Geological Survey of Canada model, as described in Atkinson
coupler inside the foundation. However, since the moment/force de- and Goda [12], is used for assessing the seismic hazard of Vancouver. In
mand above the base, i.e. beyond the plastic hinge, is not significant, this study, the hazard curves are obtained considering both peak
negligible slip is expected in that region. The slippage rotational spring ground acceleration (PGA) and spectral acceleration at the first mode
is modeled by defining a link element in SeismoStruct. The link element period (Sa, T1) as intensity measures (IMs). Here, both PGA and Sa, T1
connects two initially coincident structural nodes, and requires the are selected for the seismic hazard curves as these two IMs are com-
definition of an independent force-displacement (or moment-rotation) monly available for the site under consideration. Based on the eigen-
response curve for each of its local six degrees- of-freedom (F1, F2, F3, value analysis, the fundamental periods of all the bridge piers are found
M1, M2, M3). In order to define the link element, modified Takeda to be around 0.7 s. Fig. 4 illustrates the seismic hazard curves for the
hysteresis curve was adopted. Five parameters are necessary to location of bridge pier.
For probabilistic seismic performance assessment, selection of
Annual frequency of exceedence

Annual frequency of exceedence

1.E+00 1.E+00
(a) (b)
1.E-01 1.E-01

1.E-02 1.E-02
2% in 50 years 2% in 50 years
1.E-03 1.E-03

1.E-04 1.E-04

1.E-05 1.E-05

1.E-06 1.E-06
1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01
PGA (g) Sa (g)
Fig. 4. Seismic hazard curve for site soil class C in Vancouver (a) Peak ground acceleration and (b) spectral acceleration.

101
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

5 5
(a) UHS
(b)

Spectral Acceleration (g)

Spectral Acceleration (g)
CMS-Crustal 2
2
CMS-Inslab 1
1 CMS-Interface

0.5 0.5

0.1 0.1

0.05 0.05
0.05 0.1 0.2 0.5 1 2 5 0.05 0.1 0.2 0.5 1 2 5
Vibration Period (s) Vibration Period (s)
5 5
(c) (d)
Spectral Acceleration (g)

Spectral Acceleration (g)

2 2
1 1
0.5 0.5

0.1 0.1
0.05 0.05
0.05 0.1 0.2 0.5 1 2 5 0.05 0.1 0.2 0.5 1 2 5
Vibration Period (s) Vibration Period (s)
Fig. 5. (a) Comparison of UHS, CMS-Crustal, CMS-Interface, and CMS-Inslab at T1 = 0.7 s, (b–d) comparison of response spectra of the selected records with the target spectra for
individual

Table 3
Selected earthquake ground motion records.

No. Eq. Name Record ID Event ID Type Mw Epi. Dis (km) PGA (g) PGV (cm/s) Source

1 Northridge 953 127 Crustal 6.69 17.15 0.46 54 PEER

2 Duzce, Turkey 1602 138 Crustal 7.14 12.04 0.72 59 PEER
3 Hector mine 1787 158 Crustal 7.16 11.66 0.31 34 PEER
4 Imperial Valley 169 50 Crustal 6.53 22.03 0.28 28 PEER
5 Kocaeli, Turkey 1158 136 Crustal 7.51 15.37 0.3 54 PEER
6 Landers 900 125 Crustal 7.28 23.62 0.21 38 PEER
7 Loma Prieta 752 118 Crustal 6.93 15.23 0.48 34 PEER
8 Manjil, Iran 1633 144 Crustal 7.37 12.56 0.52 47 PEER
9 Chi Chi, Taiwan 1485 137 Crustal 7.62 26 0.47 39 PEER
10 Kobe, Japan 1106 129 Crustal 6.9 0.96 0.71 78 PEER
11 Tohuku, Japan 27,538 368 Inslab 6.8 111.88 0.85 23 K-KIK
12 27,451 368 Inslab 6.8 114.01 0.48 16 K-KIK
13 27,454 368 Inslab 6.8 112.09 0.48 12 K-KIK
14 9813 184 Inslab 7 117.21 0.75 19 K-KIK
15 9837 184 Inslab 7 52.16 0.72 15 K-KIK
16 9831 184 Inslab 7 79.59 0.58 20 K-KIK
17 20,480 294 Inslab 6 52.26 0.15 13 K-KIK
18 19,650 285 Inslab 6.2 79.79 0.14 10 K-KIK
19 Tokachi-oki, Japan 6306 148 Inslab 6.8 58.31 0.41 33 K-KIK
20 6267 141 Inslab 6.8 46.89 0.39 25 K-KIK
21 Tokachi-oki, Japan 19,085 276 Interface 7 76.98 0.66 24.60 K-KIK
22 19,004 276 Interface 7 93.02 0.34 20.18 K-KIK
23 11,026 194 Interface 7.9 119.95 0.56 36.6 K-KIK
24 11,025 194 Interface 7.9 62.65 0.38 60.15 K-KIK
25 21,598 301 Interface 7.1 97.14 0.38 13.28 K-KIK
26 Tohuku, Japan 169 – Interface 9 83.70 1.75 7.090 K-KIK
27 175 – Interface 9 71 0.96 44.43 K-KIK
28 237 – Interface 9 69.14 0.90 56.84 K-KIK
29 323 – Interface 9 62.49 0.67 27.09 K-KIK
30 168 – Interface 9 66.35 0.62 28.47 K-KIK

102
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

4 3 3 3 4
(a) (b) (c) (d) (e)
3 2 2 2 3
2 1 1 1 2
LN (PGA)

1 1
0 0 0
0 0
-1 -1 -1
-1 -1
-2 -2 -2
-2 -2
-3 -3 -3
-3 -3
y = 1.02x + 0.47 -4 y = 1.05x + 0.39 -4 y = 1.03x + 0.36 -4 y = 1.04x + 0.35 y = 1.05x + 0.54
-4 -4
R² = 0.71 R² = 0.78 R² = 0.80 R² = 0.73 R² = 0.79
-5 -5 -5 -5 -5
-4 -2 0 2 -4 -2 0 2 -4 -2 0 2 -4 -2 0 2 -4 -2 0 2
LN (MD) LN (MD) LN (MD) LN (MD) LN (MD)

Fig. 6. Comparison of the PSDMs for (a) SMA-RC-1, (b) SMA-RC-2, (c) SMA-RC-3, (d) SMA-RC-4 and (e) SMA-RC-5 considering maximum drift as EDP.

appropriate ground motions which are representatives of the seismic 6. Fragility analysis of different SMA-RC bridge piers
hazard of the site under consideration is very important. In this study,
the ground motion records are selected for seismic fragility and hazard This section describes the development of PSDMs, characterization
assessment of SMA-RC bridge piers located in site soil class-C of damage states, and fragility curve generation for different SMA-RC
(VS30 = 550 m/s), in Vancouver, BC, Canada. For the seismicity in bridge piers considering two different demand parameters. The devel-
Vancouver, consideration of shallow crustal, subcrustal, and mega- oped PSDMs and fragility curves are used to examine the impact of
thrust Cascadia subduction events are important. These ground motions different SMA properties on the seismic demand and to estimate the
have very different characteristics due to different source and path ef- relative vulnerability of different SMA-RC bridge piers.
fects [35]. In this study, the ground motions are selected by developing
conditional mean spectrums (CMS) for the three different earthquake
6.1. Probabilistic seismic demand model
scenarios (crustal, inslab and interface) that significantly contribute to
the seismic hazard of Vancouver. The CMS for three different earth-
Selection of an appropriate intensity measure (IM) and an effective
quake events are developed following the method described in Baker
engineering demand parameter (EDP) is one of the most challenging
et al. [36]. In this study, the model proposed by Baker and Cornell [18]
tasks for probabilistic seismic performance and vulnerability assess-
is used for the inter-period correlation of crustal events while for inslab
ment of structures as it dictates the reliability of the vulnerability as-
and interface events Goda and Atkinson [37] model is adopted. Fig. 5a
sessment. An appropriate EDP selection is a function of the structural
shows the uniform hazard spectra (UHS) for site soil class-C along with
system and desired performance objectives. In this study, maximum
the target CMS for crustal, inslab and interface events at T1 = 0.7 s.
drift (MD) of the bridge pier, which represents different performance-
Here, the vibration period of 0.7 s is considered since all five SMA-RC
based limit states, is considered as one of the EDPs. A review of recent
bridge piers have their fundamental period of vibration around 0.7 s.
literature [5,8] revealed that residual drift (RD) should be considered as
The UHS and the CMS of three events correspond to 2% probability of
an EDP to fully characterize the seismic performance of structures in
exceedance in 50 years which represents a return period of 2475 years.
light of the performance-based earthquake engineering. Accordingly,
From Fig. 5 it can be observed that the UHS and all the CMS has similar
this study considered residual drift as the second EDP for the com-
spectral acceleration at the target vibration period of 0.7 s. In this study
parative seismic vulnerability assessment of different SMA-RC bridge
30 ground motion suits (10 from each earthquake scenario) are selected
piers. Selection and definition of an appropriate IM has been a deba-
representing crustal, inslab and interface earthquakes in the site under
table issue among the researchers. Some researchers suggested accel-
consideration. These records are selected from PEER NGA and K-NET/
eration based IMs such as PGA [38] or spectral acceleration at the first
KiK-NET database. The records are selected in such a way so that they
mode (Sa-T1) [39] while other suggest velocity based IMs (e.g. peak
have similar spectral shape as the target CMS and the period range of
ground velocity, PGV, and spectrum intensity, SI) [40]. Because of the
interest are considered as 0.2T1 to 2T1. Similarity in the spectral shape
efficiency, practicality, sufficiency, and hazard computability of PGA,
is determined by selecting the record with the smallest average misfit
many researchers [38,41] have suggested PGA as the optimal intensity
between the target CMS and the ground motion corresponding to that
measure for fragility assessment of bridges and bridge piers. Accord-
particular event (i.e. inslab, crustal or interface). The selected records
ingly, for the purpose of this study, PGA is selected as the optimal IM.
corresponding to different earthquake types along with the target CMS
Incremental Dynamic Analyses (IDAs) are performed using the se-
and UHS are shown in Fig. 5b–d. The records selected for performance
lected 30 earthquake records for the five different SMA-RC bridge piers.
assessment of the SMA-RC bridge piers are listed in Table 3.
The maximum drift and the residual drift monitored from IDA are in-
corporated into a PSDM which establishes a linear regression of demand
(EDP)–intensity measure (IM) pairs in the log-transformed space. This

2 2 2 2 2
(a) (b) (c) (d) (e)
1 1 1 1 1

0 0 0 0 0
LN (PGA)

-1 -1 -1 -1 -1

-2 -2 -2 -2 -2

-3 -3 -3 -3 -3

-4 -4 -4 -4 -4

-5 y = 1.02x - 0.92 -5 y = 1.05x - 0.99 -5 y = 1.05x - 1.21 -5 y = 1.04x - 1.06 -5 y = 1.05x - 0.85
R² = 0.71 R² = 0.80 R² = 0.80 R² = 0.73 R² = 0.80
-6 -6 -6 -6 -6
-4 -2 0 2 -4 -2 0 2 -4 -2 0 2 -4 -2 0 2 -4 -2 0 2
LN (RD) LN (RD) LN (RD) LN (RD) LN (RD)
Fig. 7. Comparison of the PSDMs for (a) SMA-RC-1, (b) SMA-RC-2, (c) SMA-RC-3, (d) SMA-RC-4 and (e) SMA-RC-5 considering residual drift as EDP.

103
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

Table 4 Table 6
PSDMs for different EDPs. Residual drift damage states of SMA-RC bridge pier.

Pier Type Maximum Drift Residual Drift Damage State Functional Description Residual
Level Drift, RD (%)
a b β EDP| IM a b β EDP| IM
Slight Fully No structural realignment is 0.33
SMA-RC-1 1.6 1.02 0.71 0.4 1.02 0.71 (DS = 1) Operational necessary
SMA-RC-2 1.48 1.05 0.58 0.37 1.05 0.58 Moderate Operational Minor structural repairing is 0.62
SMA-RC-3 1.44 1.03 0.56 0.3 1.05 0.55 (DS = 2) necessary
SMA-RC-4 1.42 1.04 0.59 0.35 1.04 0.58 Extensive Life safety Major structural realignment is 0.87
SMA-RC-5 1.71 1.05 0.68 0.43 1.05 0.71 (DS = 3) required to restore safety
margin for lateral stability
Collapse Collapse Residual drift is sufficiently 1.22
linear regression model is used to determine the slope, intercept, and (DS = 4) large that the structure is in
danger of collapse from
dispersion of the EDP-IM relationship. Fig. 6 shows the PSDMs of dif-
earthquake aftershocks
ferent SMA-RC piers in terms of maximum drift. Each figure also depicts
the corresponding linear regression equation and R2 value. From Fig. 6,
it is evident that all the PSDMs have a R2 value higher that 0.7 which all of them focus on Ni-Ti SMA, performance-based damage states for
indicates a strong correlation between the considered EDP and IM (MD- SMA-RC bridge piers developed by [10] has been considered in this
PGA). Similarly, the PSDMs for different SMA-RC bridge piers in terms study. Billah and Alam [10] developed performance-based damage
of residual drift is shown in Fig. 7. The R2 values shown in these figures states for five different SMA-RC bridge piers in terms of maximum and
also reveal a strong correlation between this EDP-IM pair (RD-PGA). residual drift as well as considering different seismic hazard levels.
Using the linear regression model expressed in Eq. (2), the regression These damage states were developed based on extensive numerical
coefficients for various SMA-RC bridge piers in terms of different EDPs analysis and the details can be found in [10]. Table 5 and 6 show the
are computed and shown in Table 4. The parameters listed represent maximum and residual drift damage states adopted in this study and
the regression parameters from Eq. (2) along with the dispersion. From the definition of associated functional level.
the results, it is evident that in the case of maximum drift, the SMA-RC- As indicated by the closed form of fragility function in Eq. (4), a
1 bridge pier yielded an increase in dispersion in the demand (βD|IM), reliable capacity limit state model is required for developing depend-
while the SMA-RC-3 exhibited a reduction in dispersion in the demand. able fragility curves. For the selected demand parameters, each limit
On the other hand, it is evident from the regression model that the state model is assumed to follow a two-parameter lognormal distribu-
SMA-RC-5 tends to increase the median value and the slope (b) of the tion (median SC and dispersion βC). Table 7 lists parameter values used
demands placed on the piers. This can be attributed to the higher elastic to define the limit state models on the basis of the maximum drift (%)
modulus and lower yield strength of SMA-5. It reveals that SMA-RC-3 and residual drift (%). The component limit states developed by [10]
and SMA-RC-4 are effective in reducing the maximum drift of the has been used in this study. Since the study of [10] only provides the
bridge pier which is exhibited by a reduction in the parameters af- median values (SC), a prescriptive approach described by Nielson
fecting both the intercept (a) and slope (b) of the regression model. (2005) is followed to define dispersions of limit state models (βC). The
Similar observation can be made from the regression coefficients of RD- dispersion values are calculated using the following equation provided
PGA relationship. From Table 4 it is evident that SMA-RC-3 is the most by Nielson [19].
effective case in reducing the residual drift demand. This can be at-
tributed to the higher recovery strain of SMA-3, which eventually helps βc = ln(1 + COV 2)
reduce the residual deformation of SMA-RC-3.
In this equation the COV values for different limit states are calcu-
lated based on the probabilistic distribution of different limit states
6.2. Characterization of damage states described in [10]. The COV values were found to be 0.21, 0.26, 0.45
and 0.52 for DS-1, DS-2, DS-3 and DS-4, respectively. These values
An important aspect of PBEE for fragility curve development is the yielded in similar dispersion values (βC) as described by other re-
definition of appropriate damage states in relation to the functionality searchers [19].
of the structure. Four damage states as defined by HAZUS-MH [42] are
commonly adopted in the seismic vulnerability assessment of en- 6.3. Fragility curves
gineering structures, namely, slight, moderate, extensive, and collapse
damages. Damage states are often developed based on expert judgment, Using the linear PSDMs developed in previous section and limit
post-earthquake survey, and experimental results. However, in absence state models presented in Table 7, fragility curves are developed for
of sufficient experimental results or post-earthquake reconnaissance different SMA-RC bridge piers for each EDP using the closed form of
report, analysis based methods are often adopted for developing da- fragility function shown in Eq. (4). Fragility curves for the two different
mage states that corresponds to different functional level. Since very EDPs are shown in Figs. 8 and 9. The relative vulnerability of different
limited experimental results are available on SMA-RC bridge piers and SMA-RC bridge piers are compared in terms of reducing their

Table 5
Maximum drift damage states of SMA-RC bridge pier.

Damage State Performance Level Functional Level Description Maximum Drift, MD (%)

SMA-RC-1 SMA-RC-2 SMA-RC-3 SMA-RC-4 SMA-RC-5

Slight (DS = 1) Cracking Fully Operational Onset of cracking 0.28 0.30 0.28 0.28 0.28
Moderate (DS = 2) Yielding Operational Theoretical first yield of longitudinal 1.68 1.66 2.28 1.74 1.10
rebar
Extensive (DS = 3) Initiation of Local Mechanism Life safety Onset of concrete spalling 2.66 2.69 1.64 2.52 1.97
Collapse (DS = 4) Strength Degradation Collapse Crushing of core concrete 5.05 5.51 7.65 5.56 4.73

104
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

Table 7
Limit state capacity of SMA-RC bridge pier in terms of maximum and residual drift.

Damage state Maximum Drift

SMA-RC-1 SMA-RC-2 SMA-RC-3 SMA-RC-4 SMA-RC-5

Sc βc Sc βc Sc βc Sc βc Sc βc

DS-1 0.28 0.21 0.30 0.21 0.28 0.21 0.28 0.21 0.28 0.21
DS-2 1.68 0.26 1.66 0.26 2.28 0.26 1.74 0.26 1.10 0.26
DS-3 2.66 0.43 2.69 0.43 1.64 0.43 2.52 0.43 1.97 0.43
DS-4 5.05 0.50 5.51 0.50 7.65 0.50 5.56 0.50 4.73 0.50

Residual Drift

Sc βc Sc βc Sc βc Sc βc Sc βc

DS-1 0.33 0.21 0.33 0.21 0.33 0.21 0.33 0.21 0.33 0.21
DS-2 0.62 0.26 0.62 0.26 0.62 0.26 0.62 0.26 0.62 0.26
DS-3 0.87 0.43 0.87 0.43 0.87 0.43 0.87 0.43 0.87 0.43
DS-4 1.22 0.50 1.22 0.50 1.22 0.50 1.22 0.50 1.22 0.50

probability of entering into different damage states. Fragility curves for
different SMA-RC piers considering different EDPs are also compared by
evaluating the relative change in the median value of the fragility
curves.
The evaluation of the fragility curves offered a valuable insight on
the performance of different SMAs in reducing the probability of da-
mage considering the maximum drift. Fig. 8 presents the fragility curves
of the five bridge piers considering maximum drift as the EDP. From
Fig. 8a it is evident that all the piers have similar probability of cracking
damage irrespective of the intensity level. However, the effect of dif-
ferent SMAs is more pronounced in the other damage states. As de-
picted in Fig. 8b, SMA-RC-5 is more likely to experience yielding at a
lower intensity while SMA-RC-3 showed much better performance as it
showed only 47% probability of yielding even at a PGA of 2 g. This can
be attributed to the very high yield strength of SMA-3 as compared to
other SMAs. However, an interesting behavior is observed in spalling
damage state where SMA-RC-3 has the highest probability of damage
and SMA-RC-2 has the lowest. This can be attributed to the capacity
limit states of spalling damage state considered in this study where
SMA-RC-3 has the lowest drift limit before entering the spalling damage
state. In general, all the SMA-RC piers show better performance at
collapse/crushing damage state as evident from the probability of col-
lapse at maximum considered earthquake (MCE) level, which usually
corresponds to 2% probability of exceedance in 50 years (PGA 0.46 g),
which is only 0.5%, 0.1%, 0.08%, 0.3%, and 0.8% for SMA-RC-1, SMA-
RC-2, SMA-RC-3, SMA-RC-4, and SMA-RC-5, respectively.
Plots of the fragility curves for the bridge piers for residual drift as
the EDP are shown in Fig. 9, and illustrate the relative vulnerability of
the five bridge piers over a range of earthquake intensities and damage
states. Unlike maximum drift fragility curves, there are marked

1 1
(a) (b)
0.8 0.8
P [Yielding I PGA]
P [Cracking I PGA]

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
PGA (g) PGA (g)
1 1
(c) (d)
0.8 0.8
P [Spalling I PGA]

P [CrushingI PGA]

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
PGA (g) PGA (g)
Fig. 8. Fragility curves for the five SMA-RC bridge piers for: (a) slight, (b) moderate, (c) extensive and (d) collapse damage state considering maximum drift.

105
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

1 1
(a) (b)
0.8 0.8

P [DS-2 I PGA]
P [DS-1 I PGA]
0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
PGA (g) PGA (g)
1 1
(c) (d)
0.8 0.8
P [DS-3 I PGA]

P [DS-4I PGA]
0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
PGA (g) PGA (g)
Fig. 9. Fragility curves for the five SMA-RC bridge piers for: (a) slight, (b) moderate, (c) extensive and (d) collapse damage state considering residual drift.

differences in fragilities of different bridge piers in terms of residual and 1.21 g–3.05 g for DS-2 and DS-3, respectively. On the other hand, at
drift at all damage states. Irrespective of damage states, the SMA-RC-3 DS-4, only SMA-RC-5 has a median PGA lower than 3 g, while the other
showed lower probability of exceeding certain damage level. This can four SMA-RC piers have median PGA around 3.5 g and SMA-RC-3 has as
be attributed to the higher recovery strain of SMA-3 which reduced the high as 3.98 g. This can be attributed to the fact that except SMA-RC-5
residual drift in the bridge pier by bringing back the pier close to its all other SMA-RC piers have collapse drift limit (DS-4) over 5% whereas
original position at the end of ground motion. Moreover, none of the the same for SMA-RC-5 is 4.73%. However, in terms of residual drift no
bridge piers showed 50% probability of exceeding DS-2, for which the such big difference is observed at any damage state for different SMA-
bridge piers are designed, even at a PGA of 1 g. It also indicates that the RC piers.
bridge piers are performing according to the design performance ob-
jective. As evident form Fig. 9, the probability of collapse (DS-4) at
maximum considered earthquake (MCE) level is only 1.5%, 0.45%, 7. Seismic demand hazard of different SMA-RC bridge piers
0.2%, 0.68%, and 1.4% for SMA-RC-1, SMA-RC-2, SMA-RC-3, SMA-RC-
4, and SMA-RC-5, respectively. In order to fully implement the performance-based earthquake en-
The different SMA-RC bridge piers are also compared in terms of the gineering (PBEE) methodology for SMA-RC bridge pier, it is necessary
relative change in the median value of the fragility curves which in- to conduct the probabilistic seismic demand analysis (PSDA) in terms of
dicates the PGA associated with a 50% probability of reaching a certain annual rate of exceeding some structural demand parameter such as
limit state. Table 8 compares the median PGA for different damage maximum drift or residual drift. In this study, the annual rate of ex-
states of five different SMA-RC piers in terms of both EDPs. The median ceeding various levels of demand for the five considered SMA-RC piers
PGA in terms of maximum drift for different bridge piers at DS-1 ranges are estimated by aggregating the EDP|IM relationship obtained from
from 0.03 g to 0.05 g. However, at higher damage states, i.e. at DS-2 seismic response analysis with the seismic hazard curve. Using the
and DS-3, the median PGA varies over a wide range from 0.45 g–2.16 g convolution integral presented in Eq. (7), the demand hazard curves for
five different SMA-RC bridge piers are developed in terms of maximum
and residual drift.
Table 8 Fig. 10a and b show the maximum drift and residual drift hazard
Comparison of median PGA (g).
curves for five SMA-RC bridge piers, respectively. The residual drift
EDP Maximum Drift Residual Drift hazard curves depict the annual probability of exceeding different da-
mage states for different SMA-RC piers (shown with vertical dashed
Pier Type Damage State Damage State lines). It should be noted that, the same type of results for different
damage states are not presented for the maximum drift since different
DS-1 DS-2 DS-3 DS-4 DS-1 DS-2 DS-3 DS-4
maximum drift limits were considered for different SMA-RC piers. The
SMA-RC-1 0.031 1.113 2.880 3.502 0.357 1.440 3.486 – probability of collapse (probability of exceeding DS-4) of each bridge
SMA-RC-2 0.053 1.236 3.045 3.562 0.6 1.595 3.482 – pier in terms of maximum drift are summarized in Table 9.. Here, DS-4
SMA-RC-3 0.056 2.16 1.208 3.980 0.882 2.310 3.781 –
is selected to compare the probability of collapse of different SMA-RC
SMA-RC-4 0.047 1.470 3.040 3.48 0.662 1.822 3.610 –
SMA-RC-5 0.037 0.456 1.325 2.88 0.456 1.234 2.695 2.88 piers. Results show that all the bridge piers have very low probability of
collapse while the SMA-RC-3 has the lowest probability of 1.27%.

106
A.H.M.M. Billah, M.S. Alam Engineering Structures 162 (2018) 97–108

1.E+00 1.E+00
(a) (b)

Annual rate of exceedance

Annual rate of exceedance
SMA-RC-1 SMA-RC-1
SMA-RC-2 SMA-RC-2
1.E-01 SMA-RC-3 1.E-01 SMA-RC-3
SMA-RC-4

DS-1
SMA-RC-4
SMA-RC-5

DS-2
SMA-RC-5
1.E-02 1.E-02

DS-3

DS-4
1.E-03 1.E-03

1.E-04 1.E-04
0 2 4 6 8 10 0 0.5 1 1.5 2 2.5
Maximum Drift (%) Residual Drift (%)
Fig. 10. Hazard curves for five SMA-RC bridge piers (a) maximum drift and (b) residual drift.

Table 9
Annual rate and probability of collapse (DS-4) in terms of maximum drift.
SMA-RC-1 SMA-RC-2
Annual rate of 1.31E−04 1.88E−04
DS-4
Prob. Of DS-4 in 1.31% 1.88%
100 years
Table 10
Annual rate and probability of DS-2 in terms of residual drift.
SMA-RC-1 SMA-RC-2
Annual rate of 5.35E−04 3.98E−04
DS-2
Prob. Of DS-2 in 5.35% 3.98%
100 years
Among the five different SMA-RC piers, SMA-RC-5 has the highest
probability of collapse which is 31%, 1%, 33% and 23% higher that that
of SMA-RC-1, SMA-RC-2, SMA-RC-3 and SMA-RC-4, respectively. This
is due to SMA-5’s very low yield strength to elastic modulus ratio
(0.0033), which reduced the drift capacity of SMA-RC-5. The prob-
ability of exceeding DS-2 in terms of residual drift are presented in
Table 10. Here, the probability of residual drift exceeding DS-2 is pre-
sented since all the bridge piers were designed considering a target
residual drift of 0.6% which is the limiting value of DS-2. A comparison
of the five bridge piers in terms of exceeding DS-2 reveals that SMA-RC-
3 has the lowest probability of exceeding DS-2 in 100 years which is
only 2.84%. On the other hand, SMA-RC-5 resulted in highest annual
rate of exceeding DS-2 which is 6.53%. A closer look into the annual
rate of exceeding DS-2 for different SMA-RC bridge pier shows that the
annual rate of exceedance is influenced by the superelastic strain of the
SMA rebar.
8. Conclusions
This study conducted a probabilistic performance-based risk as-
sessment of five SMA-RC bridge piers when subjected to three different
earthquake scenarios (crustal, inslab and interface) that significantly
contribute to the seismic hazard of Vancouver. The piers were designed
following a performance-based design guidelines proposed by the au-
thors. In order to ensure a comprehensive seismic performance and risk
assessment, this study developed maximum and residual drift hazard
curves and fragility curves for different SMA-RC bridge piers. The in-
fluence of application of different SMAs and their properties in the
seismic hazard curve was also investigated. Based on the results ob-
tained from the risk assessment, the following conclusions are drawn:
1. The EDPs considered in this study, i.e., maximum drift and residual
drift, are shown to be well correlated with the intensity measure
(PGA) considered in this study.
SMA-RC-3 SMA-RC-4 SMA-RC-5
2. Mechanical properties of different shape memory alloys, specifically
1.27E−04 1.46E−04 1.90E−04 the recovery strain, significantly affects the seismic fragility and risk
of SMA reinforced concrete bridge piers in terms of both residual
1.27% 1.46% 1.90%
and maximum drift.
3. All the SMA-RC bridge piers, in general, are quite effective in con-
trolling the seismic response and reducing the vulnerability which is
exhibited by the low probability (less than 1%) of collapse in terms
of maximum drift at the maximum considered earthquake level.
4. In terms of residual drift, the SMA-RC-3 outperformed all other
SMA-RC-3 SMA-RC-4 SMA-RC-5
SMA-RC bridge piers at all damage states and significantly reduced
2.84E−04 3.39E−04 6.53E−04 the overall vulnerability of the bridge pier. This can be attributed to
the higher superelastic strain and low residual strain of SMA-3.
2.84% 3.39% 6.53% 5. Comparing maximum and residual drift hazard curves for different
SMA-RC piers revealed that in both cases SMA-RC-5 has the highest
probability of exceeding DS-4 and DS-2 as it was evident from the
mean annual rate of exceedance which is 1.90 × 10−4 and
6.53 × 10−4, for maximum and residual drift, respectively.
6. From the hazard analysis of different SMA-RC bridge piers, it is
expected that the SMA-RC bridge piers will incur a lower annual loss
and will provide significant financial benefit in the long run since
these SMA-RC piers showed very low probability of damage.
However, a detailed loss estimation needs to be carried out before
highlighting the potential financial benefit of SMA-RC piers.
The present study was limited to a flexure dominated single bridge
pier commonly found in highway bridges without considering soil-
structure interaction and geometric uncertainties. Incorporating such
effects and considering a bridge as a system will shed light on addi-
tional issues and are likely to change the dynamics of the response of
the entire bridge structure. In future, it will be of great interest to in-
vestigate the response of whole bridge by considering different SMAs.
Moreover, performing further study considering the construction, re-
pair, and maintenance cost of SMA-reinforced bridge, as well as com-
paring the smart bridge with a conventional bridge along with the
development of a loss-hazard relationship will shed more light on the
potential economic benefit of this smart bridge system.
Acknowledgement
The financial contributions of Natural Sciences and Engineering
Research Council of Canada (NSERC) through Discovery Grant and
Industrial Postgraduate Scholarship Program were critical to conduct
this study and are gratefully acknowledged.

107
A.H.M.M. Billah, M.S. Alam
References
[1] Canadian Highway Bridge design code. CAN/CSA-S6-14. Draft version for public
review. Ottawa; ON: National Research Council of Canada; 2014.
[2] AASHTO. AASHTO guide specifications for LRFD seismic bridge design, LRFDSEIS-
2-M, 2nd ed., Washington, DC; 2011.
[3] CEN/TC250. European Standard: Eurocode 8 – design of structures for earthquake
resistance – Part 2: Bridges, European Committee for Standardization, Brussels;
2005.
[4] Kawashima K, MacRae G, Hoshikuma J, Nagaya K. Residual displacement response
spectrum. J Struct Eng 1998;124(5):523–30.
[5] Lee WK, Billington SK. Performance-based earthquake engineering assessment of a
self-centering, post-tensioned concrete bridge system. Earthq Eng Struct Dyn
2011;40:887–902.
[6] Hewes JT, Priestley MJN. Seismic design and performance of precast concrete
segmental bridge columns. Report No. SSRP-2001/25. San Diego: Department of
Structural Engineering, University of California; 2002.
[7] Saiidi MS, O’Brien M, Zadeh MS. Cyclic response of concrete bridge columns using
superelastic nitinol and bendable concrete. ACI Struct J 2009;106(1):69–77.
[8] Billah AHMM, Alam MS. Seismic fragility assessment of concrete bridge piers re-
inforced with superelastic shape memory alloy. Earthquake Spectra
2015;31(3):1515–41.
[9] Cruz NC, Saiidi MS. Shake-table studies of a four-span bridge model with advanced
materials. J Struct Eng ASCE 2012;138(2):183–92.
[10] Billah AHMM, Alam MS. Performance based seismic design of concrete bridge pier
reinforced with Shape Memory Alloy-Part 1: development of performance- based
damage states. ASCE J Struct Eng 2016. http://dx.doi.org/10.1061/(ASCE)ST.
1943-541X.0001458.
[11] Billah AHMM, Alam MS. Performance-based seismic design of Shape Memory Alloy
(SMA) reinforced concrete bridge pier – Part 2: methodology and design example.
ASCE J Struct Eng 2016. http://dx.doi.org/10.1061/(ASCE)ST.1943-541X.
0001623.
[12] Atkinson GM, Goda K. Effects of seismicity models and new ground-motion pre-
diction equations on seismic hazard assessment for four Canadian cities. Bull
Seismol Soc Am 2011;101:176–89.
[13] Vamvatsikos D, Cornell CA. Incremental dynamic analysis. Earthq Eng Struct Dyn
2002;31(3):491–514.
[14] Cornell CA, Krawinkler H. Progress and challenges in seismic performance assess-
ment. PEER Cent News 2000;3(2):1–3.
[15] Aslani H. Probabilistic earthquake loss estimation and loss disaggregation in
buildings [Ph.D. Thesis]. Stanford, CA: John A. Blume Earthquake Engineering
Centre, Dept. of Civil and Environmental Engineering Stanford University; 2005;
2005. p. 382.
[16] Billah AHMM, Alam MS. Seismic fragility assessment of highway bridges: a state-of-
the-art review. Struct Infrastruct Eng 2015;11(6):804–32.
[17] Cornell AC, Jalayer F, Hamburger RO. “Probabilistic basis for 2000 SAC federal
emergency management agency steel moment frame guidelines. J Struct Eng
2002;128:526–32.
[18] Baker JW, Cornell CA. Vector-valued ground motion intensity measures for prob-
abilistic seismic demand analysis. Pacific Earthquake Engineering Research report
2006/08. PEER Center, University of California Berkeley; 2006.
[19] Nielson BG. Analytical fragility curves for highway bridges in moderate seismic
zones [Ph.D. thesis]. Atlanta: Georgia Institute of Technology; 2005.
[20] Deierlein GG, Krawinkler H, Cornell CA. A Framework for performance-based
earthquake engineering. In: Pacific conference on earthquake engineering.
Christchurch, New Zealand; 2003.
[21] Vecchio FJ, Collins MP. The modified compression-filed theory for reinforced
concrete elements subjected to shear. ACI Struct J 1986;83(2):219–31.
[22] Billah AHMM, Alam MS. Plastic hinge length of shape memory alloy (SMA) re-
inforced concrete bridge pier. Eng Struct 2016;117(2016):321–31.
[23] SeismoSoft. SeismoStruct – a computer program for static and dynamic nonlinear
analysis of framed structures, V 6; 2015. < www.seismosoft.com > .
[24] Pinho R, Casarotti C, Antoniou S. A comparison of single run pushover analysis
techniques for seismic assessment of bridges. Earthquake Eng Struct Dyn
2007;36(10):1347–62.
108
Engineering Structures 162 (2018) 97–108
[25] Casarotti C, Pinho R. Seismic response of continuous span bridges through fibre-
based finite element analysis. Earthquake Eng Eng Vib 2006;5(1):119–31.
[26] Alam MS, Youssef MA, Nehdi M. Analytical prediction of the seismic behaviour of
superelastic shape memory alloy reinforced concrete elements. Eng Struct
2008;30(12):3399–411.
[27] Mander JB, Priestley MJN, Park R. Theoretical stress–strain model for confined
concrete. J Struct Eng ASCE 1988;114(8):1804–26.
[28] 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. In: Symposium on the resistance and ultimate
deformability of structures acted on by well-defined repeated loads. Zurich,
Switzerland: International Association for Bridge and Structural Engineering; 1973.
p. 15–22.
[29] Filippou FC, Popov EP, Bertero VV. Effects of bond deterioration on hysteretic
behaviour of reinforced concrete joints. Report EERC 83-19. Berkeley: Earthquake
Engineering Research Center, University of California; 1983.
[30] Auricchio F, Sacco E. Superelastic shape-memory-alloy beam model. J Intell Mater
Syst Struct 1997;8(6):489–501.
[31] Billah AHMM, Alam MS. Seismic performance of concrete columns reinforced with
hybrid shape memory alloy (SMA) and fiber reinforced polymer (FRP) bars. Constr
Build Mater 2012;28(1):730–42.
[32] Billah AHMM, Alam MS. Bond behavior of smooth and sand-coated shape memory
alloy (SMA) rebar in concrete. Structures 2016;5(2016):186–95.
[33] Otani S. A computer program for inelastic response of R/C frames to earthquakes.
Civil Engineering Studies. Report UILU-Eng-74-2029, UIUC, USA; 1974.
[34] Emorsi K, Schnobrich WC. Analysis of reinforced concrete frame-wall structures for
strong motion earthquakes. Structural Research Series No. 434, 1978. University of
Illinois at Urbana-Champaign; 1978.
[35] Goda K, Atkinson GM. Seismic performance of wood-frame houses in south-western
British Columbia. Earthquake Eng Struct Dyn 2011;40:903–24.
[36] Baker JW. The conditional mean spectrum: a tool for ground motion selection.
ASCE J Struct Eng 2011;137:322–31.
[37] Goda K, Atkinson GM. Probabilistic characterization of spatially-correlated re-
sponse spectra for earthquakes in Japan. Bull Seismol Soc Am 2009;99:3003–20.
[38] Padgett JE, Nielson BG, DesRoches R. Selection of optimal intensity measures in
probabilistic seismic demand models of highway bridge portfolios. Earthquake Eng
Struct Dyn 2008;37:711–25.
[39] Mackie KR, Stojadinovi¢ B. Performance-based seismic bridge design for damage
and loss limits States. Earthquake Eng Struct Dyn 2007;36:1953–71.
[40] Bradley BA, Dhakal RP, Cubrinovski M, MacRae GA, Lee DS. Seismic loss estimation
for efficient decision making. Bull NZ Soc Earthquake Eng 2009;42(2):96–110.
[41] Billah AHMM, Alam MS, Bhuiyan AR. Fragility analysis of retrofitted multi-column
bridge bent subjected to near fault and far field ground motion. ASCE J Bridge Eng
2013;18(10):13.
[42] Federal Emergency Management Agency (FEMA). “HAZUS-MH software”,
Washington DC; 2003.
[43] Ghassemieh M, Mostafazadeh M, Sadeh MS. Seismic control of concrete shear wall
using shape memory alloys. J Intell Mater Syst Struct 2012;23:535–45.
[44] Tanaka Y, Himuro Y, Kainuma R, Sutou Y, Omori T, Ishida K. Ferrous poly-
crystalline shape-memory alloy showing huge superelasticity. Science
2010;327:1488–90.
[45] Shrestha KC, Araki Y, Takuya Nagae T, Koetaka Y, Suzuki Y, Omori T, et al.
Feasibility of Cu–Al–Mn superelastic alloy bars as reinforcement elements in con-
crete beams. Smart Mater Struct 2013;22:025025 (12pp).
[46] Omori T, Ando K, Okano M, Xu X, Tanaka Y, Ohnuma I, et al. Superelastic effect in
polycrystalline ferrous alloys. Science 2011;333:68–71.
[47] Asadi H, Bodaghi M, Shakeri M, Aghdam MM. Nonlinear dynamics of SMA-fiber-
reinforced composite beams subjected to a primary/secondary-resonance excita-
tion. Acta Mech 2015;226(2):437–55.
[48] Nakashoji B, Saiidi MS. Seismic performance of square Nickel-Titanium reinforced
ECC columns with headed couplers. Report No: CCER 14-05. NV: University of
Nevada Reno; 2014.
[49] Saiidi MS, Wang H. Exploratory study of seismic response of concrete columns with
shape memory alloys reinforcement. ACI Struct J 2006;103(3):435–42.