Application of shape memory alloy bars in self-centring precast

segmental columns as seismic resistance

Ehsan Nikbakht1*, Khalim Rashid1a, Farzad Hejazi2b and Siti. A Osman1c
1
Department of Civil and strructural Engineering, University of Kebangsaan Malaysia, 43600, Malaysia
2
Department of Civil and Structural Engineering, University of Putra Malaysia, Serdang, Malaysia

Abstract. The objective of this study is to analytically investigate the performance of the self-centring precast
segmental bridge columns with shape memory alloy (SMA) starter bars under nonlinear-static and lateral
seismic loading. For this purpose a 3D finite element model for hybrid post-tensioned bridge column has been
developed. The precast post-tensioned segmental bridge columns possessing a central tendon and adequate
transverse confinement provided by the steel tube jacketing around the first segment are as self-centring bridge
columns in this study. This is shown that the self-centring bridge columns have a high lateral seismic demand
due to their low energy dissipation which is undesirable. In order to eliminate this deficiency while keeping the
residual displacement small, SMA starter rebar which has a unique behaviour against seismic loading with the
ability of recovering large deformations over the unloading stages is applied in this system. The effect of post-
tensioning (PT) forces of the central strands and the effect of the SMA bar size are investigated by nonlinear-
static analysis. In time-history analysis, the bridge columns with SMA bars and the self-centring bridge columns
are subjected to the four different earthquake records in order to examine and compare their lateral seismic
response. The results indicate that in high seismicity zones, the bridge columns with SMA bars at a higher level
of PT forces have a superior performance against earthquake loading.

Keywords: Shape memory alloy; precast segmental bridge columns; seismic loading; post-tensioning forces;
finite element method

1. Introduction

Recently, precast segmental bridge construction has attracted the interest of many researchers. This type of
bridge has the advantages of the short period of construction without cutting off the main roads accompanying
with higher quality of construction. Moreover, their reparability after severe earthquake due to the minor
induced damages and residual displacement makes them more economic than the conventional cast-in place
bridge components. In particular, precast segmental bridge columns between prefabricated bridge components
have been the appealing subject for many researchers over the past few years due to the key role of bridge
columns in providing emergency needs for the victims of earthquake during and after seismic event. The
rocking mechanism of the precast segmental columns prevents the formation of the plastic hinge at the basement
which leads to less damages and cracks in this area. However, despite of the high ductility, the stiffness and
strength of these types of bridge columns, low energy dissipation in this system is unavoidable. Some
researchers have proposed the solution of selecting appropriate combination of mild steel reinforcements and
post-tensioning strands for increasing the energy dissipation of this system and lowering the residual
displacement [1-2]. The concept of hybrid system was introduced by Stanton et al. [3]. They investigated the
hybrid concept for reinforced concrete frame. Their research showed that the combination of mild steel
reinforcement and post-tensioning strands leads to a flag shape hysteretic performance and achievement of
higher energy dissipation and low residual displacement against the seismic loading.

* Tel.: +60 30 89215045

E-mail address: enikbakht@siswa.ukm.edu.my (E. Nikbakht).
 The bonded and unbonded precast post-tensioned bridge columns have been investigated experimentally and
analytically in recent years [4-7]. Sakai and Mahin [8] investigated the self-centering bridge columns with
mitigated residual displacements against seismic loading. They showed that RC bridge columns with the central
tendon react against seismic loading at the unloading stage in such a way that they have negligible residual
displacement. Jeong-Il et al. [9] conducted analytical and experimental work to investigate self-centring RC
bridge columns under earthquake loading. ElGawady and sha’lan [10] investigated hysteretic performance of
precast segmental bridge columns with the central post-tensioning strands and concrete-filled steel tube. They
showed that this system has high ductility without experiencing large damages with small residual
displacements.

Despite of the appropriate performance of the post-tensioned precast segmental columns in low seismicity
zones, low energy dissipation of this system against seismic loading has limited the application of this system in
high seismicity zones due to increasing the seismic lateral demands. The low energy dissipation of this system is
mainly because of the re-centring capability of post-tensioning strands which remains elastic over the whole
stages of the loading and unloading. So far, some researchers have conducted analyses in order to increase
energy dissipation of such system against seismic loading. As an alternative solution the super elastic shape
memory alloy with its unique performance can be applied in this system. In seismic structures, Saiidi and Wang
[11] investigated the effect of SMA reinforcements at the conventional RC columns which can improve the
energy dissipation with low residual displacement. Alam et al. [12] investigated the analytical prediction of the
crack width and load-deflection of the reinforced SMA beam-column against seismic loading. Montasir and
Alam [13] investigated the seismic performance of the monolithic conventional columns with various
configuration of the hybrid SMA, fiber reinforced polymer and mild steel reinforcements. Roh and Reinhorn
[14] examined analytically the hysteretic behaviour of precast segmental bridge columns with shape memory
alloy bars. They modelled the uniaxial behaviour of the SMA bars with four springs as of post-yielding,
hysteretic, slip-lock and gap-closing. Roh et al. [15] formulated moment-curvature relationship and provided
analysing tools including a modified four-spring model for the SMA bars.

In high seismicity zones, the application of the shape memory alloys in the precast segmental bridge columns
can eliminate the lack of confidence on their application through removing the deficiency of the lack of energy
dissipation and keeping them functional and repairable after severe earthquakes. However, there are very few
researches regarding with the application of the super elastic SMA bars in the post-tensioned precast bridge
columns. More experimental and analytical studies including closed form, fibre element and finite element
analysis are required to establish the guidelines against severe earthquake loading.

2. Objectives and research significance

According to the International Code Council [16] the behaviour of monolithic conventional system should be
emulated for precast construction in seismic zones except for competence performance of non- emulative
precast construction. Precast segmental bridge columns have shown appropriate performance with the benefit of
reparability due to the minor damages against seismic loading. However, their application in high seismicity
zones is limited because of their low energy dissipation compared to the emulative monolithic conventional
bridge columns. This study analytically shows that by applying SMA bars in the self-centring precast post-
tensioned bridge columns, the lack of energy dissipation of this system against seismic loading will be improved
while keeping the induced damages and residual displacement negligible. Moreover, the precast post-tensioned
segmental bridge columns at a high level of PT forces exhibit large lateral seismic demand due to the high level
of stiffness and strength without sufficient energy dissipation which leads to the concrete failure at the hinge and
joints of the segments area. This study investigates the application of the super elastic SMA starter bars on the
behaviour of the precast post-tensioned segmental bridge columns at different level of PT forces. The primary
objectives of this study are as follows:
1. Analytical investigation of the post-tensioned precast bridge columns with superelastic SMA bars under
cyclic loading and realistic ground motions.
2. Investigation performance of this system with superelastic SMA bars at different PT forces of 40 % and 70
% of the yielding stress of the central strands against seismicity.
3. Study the effect of the superelastic shape memory alloy bars on the self-centring capability of this system
and to examine their energy absorption, residual displacement, strength and seismic demand response.
3. Superelastic shape memory alloys

Since the discovery of shape memory alloy, this material has limitedly been applied in structural engineering
due to lack of knowledge on their performance, and the absence of reliable analytical methods for predicting and
analysing their performance. However, over the past 20 years they have found increasing application in
construction industry owing to numerous research efforts providing more understanding of possible applications
in different fields such as seismic structures. For instance, some of the applications of SMAs in civil structures
can be found in [17-19]. However, the price of the SMAs is expected to decrease in the future as their
applications are dramatically increasing.
Shape memory alloy are a unique material which can recover the large deformation by heating (the shape
memory effect) or removal of the stress (the super elastic effect). The ability of the SMA to recover the large
deformation upon unloading or heating is the super elasticity behaviour. The mechanism of this super elasticity
is due to the reversible phase transformation of austenite and martensite [20]. Des Roches et al. [20] illustrate
the deformation of typical shape memory alloy at different temperatures and stresses (Fig. 1). At the temperature
below the martensite finish temperature (Mf), SMA shows shape memory effect. After undergoing large
deformation as the result of applied stresses, SMA recovers from deformations by heating above the austenite
finish temperature and exhibits super elastic behaviour by heating the material at a very high temperature (Md),
the SMA shows ordinary plastic deformation with a higher strength.

Fig. 1. Stress-strain diagram for deformation of Ni-Ti shape memory alloy, Des Roches et al. [20]

4. Numerical analysis

In order to analyse the post-tensioned precast segmental bridge columns exploiting the shape memory alloys
subjecting to the cyclic and pseudo-dynamic loading firstly the analytical results are validated and verified with
the experimental results conducted by Hewes and Pristley [5]. For the analysis, 3D nonlinear finite element
ANSYS software is applied. The details of the modelling of the samples including the geometry, properties and
real constants of the materials, modelling of super elastic SMA bars and loading program are explained as
follows.

4.1 Unbonded PT segmental sample

Unbonded PT segmental bridge column, which is self-centring bridge column has four column segments. The
segments and the footing are connected together through the 27 numbers of unbonded PT strands with diameter
of 12.7 mm. The geometry and dimension of the sample is shown in Fig. 2. There is a steel tube jacketing with 6
mm thickness around the first segment. The steel tube jacketing is started from 25 mm above the footing. In this
sample, there are eight longitudinal reinforcements which are discontinuous through the joints of the segments.
The material properties of the sample are shown in table. 1.
4.2 Loading program procedure

There are four loading stages in the analysis; first post-tensioning of the strands. A 2230 kN 40 % of
yielding stress of the strands is applied. In the second stage, the PT strands are locked at the ends of the
strands. As the third loading stage 890 kN axial loading equivalent to 0.08 f'cAg, representing the bridge deck
weight is applied on the top of the column. In the fourth stage of loading, the lateral cyclic loading is applied.
The cyclic loading program is shown in fig. 3. The lateral loading is imposed as 0.6 %, 0.9 %, 1.6 %, 2.2 %,
3.0 % and 4.0 % drift displacement. In each drift, three cycles of loading are repeated. Two criteria of yielding
PT strands and the 1 % residual drift displacement have to be checked during the analysis. Once each of these
limits is reached, the analysis should not continue due to the functionality failure.

Fig. 2. Unbonded PT segmental bridge columns

Fig. 3. Lateral loading program for the models

 Table1. Properties of the precast PT segmental samples
Specimen Segmental column

Prestressing steel Material 27 D12.7

Yielding stress (Mpa) 1890
Initial stress (MPa) 756

Longitudinal reinforcement Material 8 D12.7

Yielding stress (MPa) 410

Transverse reinforcement Material D10@75

Yielding stress (MPa) 410

Steel tube jacketing Thickness (mm) 6

Yielding stress (MPa) 317
Strength of concrete (MPa) 41.4

4.3 Types of elements

The details of applied elements and their relevant properties and coefficients are as follows:
A Solid65 element is used for modelling of concrete. This element is defined with eight nodes and three
degrees of freedom at each node. This element has the capability of cracking in tension, crushing, and nonlinear
plastic deformation. The compressive stress-strain of concrete is obtained using Eqs. (1-3) in [21].

f = (1)

= (2)

′
= 4700 (3)

where 0 is the strain at the ultimate compressive strength of concrete, is elastic modulus of the concrete
(MPa) up to 30 % of compressive strength of concrete.

The Link8 element is used for the mild steel reinforcement. It has the capability of plasticity, swelling, stress
defining and large deflection. Shell181 is used for the steel tube jacketing. This element is appropriate for
analysing thin and moderate shells.
During the occurrence of segment uplift, the stiffness of contact elements at joints has to be set zero and no
penetration should be considered while closing the segments at the unloading stages. For this purpose, the
Contact174 and Target170 with unilateral flexible surface to surface and penalty method contact algorithm are
employed to model the contact between segments. According to ACI 318-02 [22], coefficient of friction of 0.5
is used for the friction between surfaces of two adjacent segments.
The Solid185 element is used for the PT strands. This element has the capability of large deflection, plasticity
and large strain, and the Prets179 element is applied for distributing the pretension force in the tendon. This
element is an element with one degree of freedom in one translational direction. It acts between meshed solid
elements. In fact, there are three defined nodes for distributing the pretension force by this element: two created
coincident nodes, and a third node through which the direction of force should be specified.
Bilinear elastic-perfectly plastic stress-strain has been assumed for the longitudinal, transverse and post-
tensioning strands.
 4.4 Modelling of super elastic SMA bars

The material model of shape memory alloy is defined using Auricchio model [23] in the finite element
analysis of ANSYS package. The constants of stress-strain behaviour of this material with their definition are
shown in table. 2.
The Solid185 element is used for modelling of shape memory alloys. The SMA bars are passed through the
corrugated ducts through the first two segments and the footing foundation. The SMA bars are bonded in order
to provide the higher energy dissipation.

Table 2 SMA material properties used for the models

Constants Description Value
354
4 Starting stress value for artensite 414
phase (MPa)
354
6
Final stress value for aretensite phase 530
(MPa)
375
4 Starting stress value at the unloading 380
stage (MPa)
375
8 Final stress value at the unloading 130
stage (MPa)
9:; Maximum residual strain (%) 6.2
E Modulus of the elasticity (MPa) 54200

Fig. 4. Idealized stress-strain of super elastic

SMA [24]

4.5 Concrete failure criteria

Multi linear isotropic Von Mises formula [25] is applied to show the limit of critical stress for principle space
stresses of three axes. Eq. 4 shows the critical stress limit for the yielding of the concrete. The Willam and
Warnke [26] model is used for surface failure criteria i.e. determining the conical surface failure of concrete
(Eq. 5).

1
= [ − 2" + − 3" + − 1"
2 2 2 ]
2 1 2 3 (4)
'( ,(
" = +* + −1 (5)
& ) +" )

where is the critical stress, , 2 and 3 are the principle stresses, - and .- are average stress components,
/
is the ultimate compressive strength of concrete, 0 is the surface apex, 1 2" is a position vector with 2
(angle of similarity) and " is conical failure surface.

4.6 Validation and verification

Fig. 4 shows a comparison of the cyclic lateral force-displacements of the unbonded segmental sample
predicted by finite element method in this study (FEM) and the experiment (EXP). It is clear from the figure that
FE closely follows the experimental results. In Table 3, the lateral force of the experimental and numerical
results is compared for every drift level. The difference between the FE and experimental results is less than 6.6
%, which implies the accuracy of the predicted results in this study.
 According to Hewes and Pristley [27], the unbounded post-tensioned sample was tested up to 3 % lateral drift
(108.9 mm) because concrete spalling occurres in the second segment at the joint of first segment area which is
captured appropriately by analytical method (Fig. 5).

Table3. Comparison of FEM and

experimental results

Lateral force (kN)

Drift (%) EXP FEM Difference
Analysis (%)
0.6 164.7 164.0 0.4
-0.6 -156.2 -165.3 5.8
0.9 186.7 183.6 1.6
-0.9 -175.6 -183.7 4.6
1.2 196.9 193.1 1.9
-1.2 -185.3 -192.8 4.0
1.6 206.1 200.8 2.5
-1.6 -194.9 -201.0 3.1
2.0 210.4 209.6 0.3
-2.0 -199.4 -209.5 5.0
3.0 217.0 220.1 1.4
-3.0 -206.3 -220.0 6.6

Fig. 4. Load-deflection of EXP and FEM

Experimental test by Hewes and Priestley [27] b) FEM modelling (this study)
Fig. 5. Cracks and crushing comparison of experiment and analysis
5. Bridge columns analysed

In this part, three types of precast segmental bridge columns of the self-centring columns, the columns with
SMA starter bars and columns with mild steel starter bars are investigated and compared in terms of stiffness,
energy dissipation and residual displacement. As mentioned before, self-centring bridge columns are the
columns with steel tub jacketing around the first segment and a central PT strands which returns the column to
the original position when unloading occurs.
At the second bridge column, the mild steel reinforcements are applied as starter bars throughout the footing
and the first two segments. In order to achieve equal stiffness and nominal shear strength compared to the self-
centring bridge columns, eight mild steel starter bars with 16 mm diameter are applied instead of steel tube
jacketing in the first sample (Fig. 6). Fig. 7 compares the pushover lateral load-deflection of the self-centring
bridge column and the second type of bridge columns with mild steel starter bars which shows that their
stiffness and strength are very close (The strength of 266.7 kN of the sample with mild steel rebar versus 272.0
kN strength of the Self-center sample).

Fig. 6. Precast segmental bridge columns with SMA starter bars

Fig. 7. Pushover lateral load-deflection of the sample with mild steel rebar (without steel tube jacketing) and
the Self-centring sample (without starter bars)
The column with SMA starter bars which are replaced with mild steel reinforcements in the second type of
bridge column as the third type of bridge columns are analysed in this study. The geometry and material
properties of the all analysed bridge columns are the same. The samples mentioned above are analysed at
different initial stress levels of 40 % and 70 % of the central PT strands. The bridge columns analysed are
labelled as follows:

The label Self-center is representing the self-centring bridge columns; the bridge columns with the SMA
starter bars and the columns with the mild steel reinforcement are labelled as SMA and MS samples study
respectively.

6. Nonlinear-static analysis

In this part, the bridge columns with SMA bars (SMA sample) and the bridge columns with mild steel bars (MS
sample) were subject to nonlinear-static loading. The effect of the PT forces and SMA bar sizes are investigated
in the following subsections. In order to assess the performance of the SMA bars, the sample with SMA bars
are compared with the samples with mild steel rebar through the analyses.

6.1 Effect of post-tensioning forces

Fig. 8 compares the performance of the three types of samples. The samples are post-tensioned at 40 % and 70
% initial stress levels of central PT strands. The figure indicates that the self-centring sample and MS sample
with 16 mm diameter of mild steel starter bars show very close strength at 40 % initial stress level of PT strands
(Fig.8a). However, the SMA sample exhibits lower strength due to the lower stiffness of the SMA bars.

The stiffness and strength of all samples increase by raising the initial stress level to 70 % (Fig. 8b).The
stiffness reduction of SMA samples at 40 % and 70 % initial stress levels of PT strands are shown in the figure
9. This figure depicts that increasing the initial stress raises stiffness of SMA sample against seismic loading. As
shown in these figures, the SMA sample shows very small residual displacement due to re-centring capability of
SMA bars at unloading stages. The results indicate that increasing the initial stress of PT strands has little effect
on residual displacement of SMA sample against lateral cyclic loading.

Fig. 8. Lateral load-deflection of the samples with and without SMA starter bars and with mild steel
reinforcements at (a) 40 % initial stress level and (b) 70 %initial stress level
 Fig. 9. Stiffness reduction of the sample with SMA bars at 40 % and 70 % initial stress level of PT strands

The equivalent viscous damping of the bridge columns subjected to 40 % and 70 % initial stress level are
shown in Fig. 10. The equivalent viscous damping of the bridge columns analysed has been derived according to
[28]. By increasing the initial stress level to 70 %, the energy absorption of all samples increase due to crushing
of concrete around PT strands at the centre of columns at higher initial stress level.
It can be seen in Fig. 10a that under 40 % initial stress level, the SMA sample has the highest level of
equivalent viscous damping at the first three cycles of loading. However, from the third cycle onwards the MS
sample shows higher amount of energy absorption because of increasing cracks and residual displacements
induced in this sample.

Fig. 10. Equivalent viscous damping of samples at (a) 40 % and (b) 70 % initial stress level

6.2 Effect of SMA bar size

Fig. 11 compares the load-deflection of SMA and MS samples with different diameters of 10, 16 and 32 mm
at 70 % initial stress level of PT strands under cyclic loading. Fig. 11a indicates that MS sample has higher
energy dissipation and shows significantly higher residual displacement compared to SMA sample. However,
this sample exhibits more stiffness and strength (360 kN versus 300 kN). Fig. 11b indicates that SMA and MS
samples with 16 mm diameter rebar show closer stiffness, strength and residual displacement in comparison
with the sample with 32 mm diameter. In Fig. 11c, these two columns with 10 mm diameter of starter bars
exhibit almost the same stiffness, strength, energy dissipation and residual displacement.
Fig. 11. Lateral load-deflections of sample with SMA bars with diameter of, (a) 32 mm, (b) 16 mm and (c)
10 mm

As it is shown in figures, raising the diameter of SMA bars leads to higher values of strength and lower
stiffness. Fig. 12 compares the equivalent viscous damping of samples with different SMA bars. It is shown that
the equivalent viscous damping of the samples with 10 mm and 16 mm are similar. However, the column with
10 mm diameter of starter bars has higher viscous damping over the 4 % drift level compared to other bridge
columns. The figure indicates that increasing the diameter of SMA bars reduces the energy dissipation of the
columns such that the sample with 32 mm diameter starter bars has the lowest value of equivalent viscous
damping.
The column with 10 mm diameter SMA bars undergoes very high level of stress. As a consequence the SMA
bars experience ordinary plastic deformation, similar to the bridge columns with mild steel rebar, despite of its
lower stiffness. Increasing the diameter of the SMA bars slightly decreases the residual displacement of the
analysed samples.

Fig. 12. Equivalent viscous damping of the samples with different diameter of 10 mm, 16 mm and 32 mm SMA
rebar
7. Time-history analysis

In order to examine the dynamic performance of precast segmental columns with SMA bars, time-history
analysis is conducted. The aforementioned precast post-tensioned segmental bridge columns at 40 % and 70 %
initial stress levels of the central PT strands are subjected to four ground motions with different PGA. Fig. 13
shows the earthquake records applied in the analysis. The 5 % damping ratio is applied according to AASHTO
recommendation [29].

Fig. 13. Earthquake records applied in the analysis (a) Northridge, 1994 USA, PGA= 0.36g, (b) Loma Prietta,
1989 USA, PGA = 0.6g, (c) Imperial Valley, 1979 Mexico-California border, PGA = 0.26g and (d)
Coalinga, 1983 USA, PGA = 0.34g. [30]

7.1 Results and discussion

The lateral seismic demands of SMA and Self-centre samples at the 40 % and 70 % initial stress levels of PT
strands against the four earthquake records are compared in Figs 14-17.
Fig. 14a compares the performance of the post-tensioned self-centring bridge columns and SMA sample with
40 % initial stress of PT strands subjected to Loma Prietta earthquake record. The figure indicates that the Self-
center sample exhibits larger lateral deck displacement compared to the sample with SMA rebar. Fig. 14b
compares their lateral deflection at 70 % initial stress level. In this figure, the bridge column with SMA bar
shows lower lateral peak response (209.0 mm versus 236.0 mm). Similar behaviour can be seen in Figs. 15-17
in which the SMA samples exhibit the lower lateral peak displacement.
The moderate earthquakes analyses in Figs. 16 and 17 depict that Self-center and SMA bridge column samples
at 40 % initial stress level demonstrate higher peak lateral deflection. However, in the more severe earthquake
records analyses (Figs. 14 and 15) at 70 % initial stress level, the Self-center and SMA samples show higher
lateral peak displacement compared to 40 % initial stress level. This inconsistency is due to the higher level of
stiffness and strength of the bridge columns under very high level of PT forces with insufficient energy
dissipation. This lack of energy dissipation against seismic loading leads to crushing of concrete in hinge area
and the joints of the column segments, which causes considerable stiffness loss of the columns. Although there
is increasing of energy dissipation in columns with SMA bars, the results indicate that there is still a lack of
energy dissipation for the case of bridge columns at very high level of initial stress.
However, the lower lateral deflection of SMA bridge column samples compared to the Self-center bridge
columns at 40 % and 70 % initial stress levels yielded because of the ability of the SMA bars to recover the
large lateral deformation with higher energy dissipation over the unloading stages. Fig. 18 shows the reduction
of lateral peak response of samples with incorporating SMA bars subjected to the most severe earthquake
records in this study, at 40 % and 70 % initial stress levels. It can be seen from the figure that SMA bridge
column samples at 70 % initial stress level exhibit more appropriate performance against severe earthquake.
Fig. 14. Time-history lateral displacement of (a) samples at 40 % initial stress level and (b) samples at
70 % initial stress level against Loma Prietta EQ.

Fig. 15. Time-history lateral displacement of (a) samples at 40 % initial stress level and (b) samples at
70 % initial stress level against Northridge EQ.

Fig. 16. Time-history lateral displacement of (a) samples at 40 % initial stress level and (b) samples at
70 % initial stress level against Imperial Valley EQ.

Fig. 17. Time-history lateral displacement of (a) samples at 40 % initial stress level and (b) samples at
70 % initial stress level against Coalinga EQ.
Fig. 18. Reducing lateral peak displacement effect of SMA bars on self-centring columns at 40 % and 70 %
initial stress levels subjected to most intense earthquake records

8. Conclusions

In this study, the performance of precast segmental self-centring bridge columns with SMA bars is
investigated. The previous studies show that the self-centring segmental bridge columns at very high level of
post-tensioning forces subjected to severe earthquake loading shows higher level of stiffness and strength with
inadequate energy dissipation against earthquake loading that leads to the high lateral seismic demand. As an
alternative solution, SMA bars are applied in this system in order to increase its energy dissipation. The analyses
were performed at 40 % and 70 % initial stress levels.
In nonlinear-static analysis self-centring bridge sample, the sample with SMA bars and sample with mild steel
reinforcements are compared. It is found that the bridge column with SMA bars show the highest energy
dissipation up to the first three cycles of loading. However, the column with mild steel starter bars show higher
amount of energy dissipation up to 4% drift displacement due to crushing of concrete and the induced damages
at the hinge area.
Increasing the post-tensioning forces in all samples analysed leads to higher stiffness, strength and energy
dissipation. The samples with SMA bars exhibit very small residual displacement due to the unique re-centring
capability of SMA bars when unloading occurs. The results of nonlinear-static analyses show that an increase in
initial stress level has little effect on residual displacement of the bridge columns with SMA bars.
We compared the effect of SMA bar size on precast self-centring bridge columns. The results indicate that the
samples with larger diameter of SMA bars demonstrate lower stiffness and equivalent viscous damping and
higher strength. Furthermore, small diameter SMA bars sample (10 mm) has very similar behaviour to the
sample with mild steel bars. This indicates that when SMA bars undergo very high level of stress, they
experience ordinary plastic deformation with higher strength, considering their relatively low stiffness.
In time history analyses, self-centring columns and the columns with SMA bars are compared. The results
show that application of SMA bars decrease the lateral seismic demand of precast segmental self-centring bridge
columns in both 40 % and 70 % stress levels of PT strands. Moreover, when the bridge columns with SMA bars
with 70 % initial stress level of PT strands are subjected to intense earthquake records, there is lower lateral
peak displacement compared to the samples at 40 % initial stress level.

Acknowledgements

The authors wish to acknowledge the financial support received from the University of Kebangsaan Malaysia
(Grant No. FRGS:Gup-2012033) in connection with this work.

References
[1] A. Palermo, S. Pampanin, G.M. Calvi, Concept and development of hybrid solutions for seismic resistant
bridge systems, J. Earthquake Eng., 9(6), 2005, 899-921.
[2] Y.C. Ou, P.H. Wang, M.S. Tesaei, K.C. Chang, G.C. Lee, Large scale experimental study of precast
segmental unbonded posttensioned concrete bridge columns for seismic regions, J. Struct. Eng., 136(3), 2010,
255-264.
[3] J.Stanton, W.Stone, G.C. Cheok, A hybrid reinforced precast frame for seismic regions, PCI J. 42(2), 1998,
20-32.
[4] J.T. Hewes, M.J.N Priestley, Seismic design and performance of precast segmental bridge columns, PhD
thesis, 2002, Univ. of California Sandiego.
[5] A. Palermo, S. Pampanin, D. Mariott, Design, modelling and experimental response of seismic resistant
bridge piers with posttensioned dissipating connections, J. Struct. Eng., 133(11), 2007, 1648-1661.
[6] W.P. Kwan, S.L. Billington, Unbonded posttensioned concrete bridge piers I: monotonic and cyclic
analyses, J. Bridge Eng., 8(2), 2003, 92-101.
[7] M.A. Elgawady, H.M. Dawood, Analysis of segmental piers consisted of concrete filled FRP tubes,
Engineering Structures, Vol. 38, 142-152.
[8] J. Sakai, S. Mahin, Mitigation of residual displacement of circular reinforced concrete bridge columns, in:
Earthquake engineering Conf Vancouver,B.C., Canada, 2004.
[9] H.I. Jeong, J. Sakai, S. Mahin, Shaking table test and numerical investigation of self-centring reinforced
concrete bridge columns, PEER report, Pacific earthquake engineering, 2008, Univ. California Berkley.
[10] M.A. ElGawady, A. Sha’lan, Seismic behaviour of self-centring precast segmental bridge bents, J. Bridge
Eng., 16(3), 2011, 328-339.
[11] S. Saeiidi, H.Y. Wang, Exploratory study of seismic response of concrete columns with shape memory
alloys reinforcements, ACI, 103(3), 2006, 435-442.
[12] M.S. Alam, M.A. Youssef, M. Nehdi, Analytical prediction of the seismic behaviour of super elastic shape
memory alloy reinforced concrete elements, Eng. Struct, 30(12), 2008, 3329-3411.
[13] A.H.M Montasir Billah, M.S. Alam, Seismic performance of concrete columns reinforced with hybrid
shape memory alloys (SMA) and fiber reinforced polymer (FRP) bars, Construction and Building Material, Vol.
28, 2012, 730-742.
[14] Hwasung. Roh, A.M. Reinhorn, Hysteretic behaviour of precast segmental bridge piers with superelastic
shape memory alloy bars, Eng. Struct., Vol.32, 2010, 3394-3403.
[15] Hwasung. Roh, A.M. Reinhorn, L. Jeong She, Modelling and cyclic behaviour of segmental bridge
columns connected with shape memory alloy bars, Earthquake Eng. and Eng. Vibration, 11(3), 2012, 375-389.
[16] ICC, International Code Council, International bridge code, 2006, Falls Church.
[17] Jc. Wilson, M.J. Wesolowsky, Shape memory alloys for seismic response modification: a state-of-art
review, Earthquake Spectra, 21(2), 2005, 569-601.
[18] L. Janke, C. Czaderski, M. Motavalli, J. Ruth, Application of shape memory alloys in civil engineering
structures-Overview, limits and new ideas, Materials and Structures, Vol.38, 2005, 578-592.
[19] G. Song, N. Ma, HN. Li, Applications of shape memory alloys in civil structures, Eng. Struct., 28(9), 2006,
1266-1274.
[20] R. Des Roches, J. Mc Cornic, M. Delemont, Cyclic properties of superelastic shape memory alloys wires
and bars. J. Struct Eng., 130(1), 2004, 38-46.
[21] P. Desayi, S. Krishnan, Equation for the stress-strain curve of concrete, J. Am. Conc Inst, Vol.61, 345-350.
[22] American Concrete Institute (ACI), ACI 318-02, Building code requirement for structural engineering,
2002, Famington Hills.
[23] E. Auricchio, R.L. Taylor, J. Lubliner, Shape memory alloys: Macro modelling and numerical simulations
of the superelastic behaviour, Computational Method in Applied Mechanical Eng. Vol.146, 1997, 281-312.
[24] ANSYS, User’s manual, Theory reference for the mechanical APDL and mechanical applications, 2009.
[25] A.J. Wolanski, Flexural behaviour of reinforced and prestress concrete beam using finite element analysis,
Master thesis Marquette university, 2004.
[26] K.J. Willam, E.P. Warnke, Constitutive model for triaxial behaviour of concrete, proceeding of the
international association of bridge and structural engineering conference, Vol.19, 1974, p174.
[27] J.T. Hewes, M.J.N. Priestley, Seismic design and performance of precast segmental bridge columns, Rep.
No. SSRP. 2001, Univ. of California at San Diego.
[28] M.J.N Priestley, F. Seible, G.M. Calvi, Seismic design and retrofit of bridges, Wiley, New York, 1996.
[29] AASHTO, Standard specifications for buildings, 16th ed., Washington,D.C., 1996.
[30] PEER, Pacific Earthquake Engineering Research centre, Univ. California, Berkeley.