Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

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

Pre-tensioned SMA cable tests and its application in a novel self-centering

viscoelastic damper
Hui Qian a, Dexin Wei a, Yifei Shi a, *, Zongao Li a, Hongnan Li b
a
School of Civil Engineering, Zhengzhou University, Zhengzhou, 450001, China
b
School of Civil Engineering, Dalian University of Technology, Dalian, 116024, China

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

Keywords: The elastic modulus of superelastic NiTi shape memory alloy (SMA) is between normal structural steel and
Shape memory alloy (SMA) viscoelastic materials. To enhance the initial stiffness and self-centering capability of viscoelastic (VE) dampers,
SMA cable an innovative self-centering high-damping rubber damper (SC-HRD), which is parallel composed of an SMA-
Viscoelastic (VE) damper
cable-controlled self-centering system and a traditional high-damping rubber damper, was proposed and
Residual deformation
Self-centering
investigated in this study. The influence of initial strain, loading amplitude, and strain rate on the behavior of
Parametric analyses SMA cables after training and pre-tensioning was also tested for the SC-HRD application. Configuration and
working principles of the SC-HRD were subsequently illustrated, followed by numerical studies on the behavior
of the novel damper. Parametric analyses were then carried out on several design parameters of the SC-HRD to
optimize its mechanical performance. Results show that the initial stiffness and residual self-centering strength of
the SMA cables, which are essential for the SC-HRD application, can be enhanced with the increasing initial
strain and loading strain rate. Compared with traditional VE dampers, the SC-HRD has better initial stiffness and
self-centering capacities. Initial strain and relative area ratio of the pre-tensioned SMA cables in the proposed SC-
HRD are suggested to be designed as about 2% and 0.123%, respectively, to achieve appropriate initial stiffness,
energy dissipation, and residual deformation of the proposed SC-HRD.

1. Introduction considered in the VE dampers, but it is still a challenge to balance their

stiffness and damping contributions. Moreover, under rare earthquakes,
Various dampers have been developed and implemented in seismic the seismic response and plastic damage of primary structures can be
regions [1], which can be grouped into two broad categories [2]: controlled by the VE dampers to some extent because of their high
displacement-dependence dampers, such as metallic yielding dampers, damping properties [8,9]. But, rare earthquakes can also cause signifi­
and friction dampers, as well as velocity-dependence dampers, such as cant post-earthquake residual deformations on the structures, which are
viscous or viscoelastic (VE) dampers. Compared with metallic yielding difficult to alleviate by VE dampers, to some extent that the buildings
dampers, which do not provide significant energy dissipation at could be decommissioned if the residual deformations are too excessive
low-level vibrations [3], VE dampers can effectively dissipate the input [10,11].
energy for structures at all displacement levels even when the structures Shape memory alloys (SMAs) are a class of smart materials having
are subjected to frequent wind loads. However, the vibrations and two well-known characteristics, i.e., the unique shape memory effect
inter-story drifts of the structures under frequent earthquakes may be (SME) and superelasticity (SE). Due to reversible phase transformations
significantly increased due to the taller and slenderer design of the between two solid phases, SMA materials can return to their original
structures and the insufficient initial stiffness of the VE dampers. shape after undergoing large deformations [12,13]. The superelastic
A number of investigations have made attempts to improve the initial SMA materials possess various advantageous properties, such as excel­
stiffness and damping properties of the VE dampers by changing the lent self-centering capacity, satisfactory energy dissipation, high fatigue,
rubber matrix [4,5] the polymeric material [6], and the rubber inter­ and corrosion resistance. Moreover, the initial stiffness of superelastic
penetrating polymer network [7]. Various VE materials have been SMA materials is between the normal structural steel and the VE

* Corresponding author.
E-mail addresses: [email protected] (H. Qian), [email protected] (D. Wei), [email protected] (Y. Shi), [email protected] (Z. Li), hnli@dlut.
edu.cn (H. Li).

https://doi.org/10.1016/j.soildyn.2023.107850
Received 18 October 2022; Received in revised form 9 February 2023; Accepted 20 February 2023
Available online 24 February 2023
0267-7261/© 2023 Elsevier Ltd. All rights reserved.
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

materials. In this context, the realization of a parallel connection be­

tween the VE damper and SMA tensile elements through a certain
structural measure can be a novel way to appropriately enhance the
stiffness and self-centering capabilities of the VE damper while retaining
its high damping property.
The superelastic SMA materials have already been developed in
many areas of structural seismic application, such as sliding bearings
[14,15], restrainers [16–19], and in-span hinges [20], as well as utilized
to replace some of the reinforcements in concrete beams [21–23] col­
umns [24,25], joints [26], and shear walls [27]. There are also many
developments of SMA-based braces/dampers for seismically resistant
structures. Fraternali et al. [28] presented a novel application of ten­
segrity systems as lightweight and noninvasive braces. The analyzed
braces consist of D-bar systems equipped with shape memory alloy ca­
bles with superelastic response. Santos et al. [29] examined the effi­
ciency of three-dimensional D-bar tensegrity structures as the bracing
devices for seismic control under compressive loads. They show that, by
introducing a pre-strain in the superelastic SMA cables, it is possible to
achieve a wide shaped hysteresis, which yields to a significant amount of Fig. 1. Test results of the commercial superelastic SMA wire.
equivalent viscous damping (up to 30%). Besides, it has been revealed
that the braces have limited energy dissipation capability if the SMA
elements are used as the sole energy dissipation source [30], so extra
energy dissipation components are recommended. The most commonly
considered strategies are adding frictional devices [12,31,32] or using
SMAs in parallel with a BRB [33,34]. The effectiveness of the SMA-based
BRBs/braces in the mitigation of seismic response and post-earthquake
residual deformation has already been confirmed by many studies
[35–37]. However, there are only a few reports on the application of
SMA materials as self-centering elements and initial stiffness enhancers
in the VE damper or braces. Ozbulut et al. [38] first introduced a
superelastic viscous damper (SVD) that strategically combines a visco­
elastic device and SMA cables in parallel. The SMA cables were wound
around the outside of the damper, and the SVD exhibits improved
self-centering and energy-dissipating capabilities compared to only
viscoelastic or only SMA-based devices. Fang et al. [39] proposed an
innovative SMA-VED self-centering brace. The VE dampers were exter­
nally installed on two sides of a SMA brace. The seismic mitigation ef­
fects of the SMA-VED self-centering brace on steel frames were
numerically verified.
Fig. 2. Construction and dimensions of the prepared SMA cable.
In this study, an innovative self-centering high-damping rubber
damper (SC-HRD) composed of a novel self-centering system (SCS) and a
traditional high-damping rubber damper (HRD) was proposed and 2. Experimental tests of pre-tensioned SMA cables
investigated to appropriately enhance the initial stiffness and self-
centering capability of VE dampers while retaining their high- 2.1. Materials and cable specimens
damping property. The HRD was internally installed in the SCS, and
SMA cables were selected as the self-centering elements in the SCS The superelastic NiTi SMA wires with a diameter of 0.4 mm used in
because of their moderate elastic modulus, sufficient bearing capacity, this study were obtained commercially from Xian Saite Metal Materials
excellent superelastic behavior, satisfactory self-centering and equiva­ Development Co., Ltd. The chemical composition of the SMA wires is
lent viscous damping (HRD) abilities, as well as their relative conve­ 55.86 wt% nickel and 44.05 wt% titanium alloys. A cyclic tensile test
nience in anchoring [34,40–42]. Moreover, the SMA cables should be with incremental loading amplitudes was previously conducted to verify
appropriately trained and pre-tensioned to enhance and stabilize the the superelasticity of the SMA monofilament wires at a room tempera­
mechanical properties of the proposed SC-HRD. But there only a few ture of 23 ◦ C ± 1 ◦ C. Fig. 1 shows the tested results of the SMA wires,
reports indicate that a much higher initial stiffness of SMA cables can be which exhibit typical flag-shaped stress-strain curves of the superelastic
achieved in a pre-tensioned state [43]. However, the cyclic behavior of SMA wires. With the increase of the loading amplitude, there are a
pre-tensioned SMA cables may be unpredictable and susceptible because certain accumulation of residual strain and an obvious downward
of the unique superelasticity of SMA materials. So, the influence of movement of the transformation plateaus. The SMA wires were then
initial strain, loading amplitude, and strain rate on the mechanical machined into a conventional double-layered 7 × 7 cable construction
properties of the pre-tensioned SMA cables was also tested in this study [40–43] in Henan Juli cable manufacturing Co., Ltd., as shown in Fig. 2.
for their SC-HRD application. The hysteresis behaviors of the To investigate the influence of initial strain, loading amplitude, and
pre-tensioned SMA cables and a traditional HRD were subsequently strain rate on the cyclic behavior of pre-tensioned SMA cables, a total of
simulated, and the finite element (FE) model of the proposed SC-HRD 15 SMA cable specimens were cut from the same coil of the prepared
was further established and verified. Based on the FE model, the ef­ SMA cable. Two ends of each specimen were firmly constrained to avoid
fects of initial strain and relative cross-sectional area ratio of the unraveling. Then the cable specimens were annealed in an electrical
pre-tensioned SMA cables as well as the displacement amplitude and furnace at 300 ◦ C for the same duration of 10 min to reduce internal
loading frequency on the mechanical properties of the SC-HRD were stress, set shape, and enhance the self-centering capability of the SMA
parametrically analyzed and discussed. cables [43]. Transformation temperatures of the SMA after annealing

2
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

SMA cable is much lower than the room temperature, and its interme­
diate phase (R phase) is not obvious. This indicates that the SMA cables
after this annealing procedure are still in a relatively deeply superelastic
state at room temperature and probably exhibit a satisfactory super­
elasticity and self-centering capability.
The total length and free length of each cable specimen before cyclic
tensile training are about 250 mm and 120 mm, respectively. Epoxy and
screwed iron ferrules were used to coat the two ends of each cable
specimen, as shown in Fig. 4, to alleviate the stress in gripping regions
and to facilitate the assembling of SMA cables into the loading devices or
equipment. The free length is defined as the length of the region between
the two coated ends of each specimen, which can be freely stretched
under test loads. Table 1 summarizes the details of each cable specimen
for the training process, pre-tensioning procedures, and cyclic tensile
tests. The specimens (A1-A3) prepared for testing the basic mechanical
properties of the conventional 7 × 7 × 0.4 mm SMA cables were virgin
without being trained before their cyclic tests; while the other speci­
Fig. 3. DSC result of the SMA cables after 300 ◦ C annealing. mens, which were prepared for the investigations on the cyclic behavior
of the pre-tensioned SMA cables, were suffered two different procedures
of cyclic tensile training with different constant strain amplitudes for 50
laps and several different procedures of pre-tensioning before their
official cyclic tensile tests.

2.2. Testing procedure and setup

The SMA cable specimens were tested by applying uniaxial cyclic

loads using a universal test machine (UTM) with a load capacity of 200
kN. The load was recorded by a load cell built in the UTM, and the stress
of the cable specimen was calculated by dividing the load by the sum of
the cross-sectional area of the 49 (7 × 7) monofilament SMA wires [43].
The local field strain data near the mid-length region of the specimens
was provided by employing a physical extensometer with a gauge length
of 50 mm, as shown in Fig. 4(b). The screwed ferrules of the two ends of
the specimens were screwed in the fixtures, and the fixtures were tightly
clamped by the top and bottom hydraulic wedge grips. All the cable tests
were conducted at a room temperature of 23 ± 2 ◦ C, and the
strain-controlled loading with the constant strain rate for each specimen
(summarized in Table 1) was applied.
First, three virgin SMA cable specimens (A1, A2, and A3) were tested
by employing two typical loading protocols, i.e., the cyclic loading with
incremental strain (for A1) and the cyclic loading with constant strain
(for A2 and A3), to investigate the basic mechanical properties of the
Fig. 4. Details of SMA cable specimens and the test setup. prepared SMA cable and to optimize the training procedures. For A1, the
loading process starts with an initial amplitude of 1% and proceeds with
are exposed by a differential scanning calorimeter (DSC) in a tempera­ a strain amplitude incremental interval of 1% until the strain amplitude
ture range of − 120 ◦ C–100 ◦ C. Fig. 3 shows the results of the DSC. It can reaches 10%. Two cycles were repeated at each strain amplitude. For
be observed that the martensite starts temperature (Ms ) of the annealed specimens A2 and A3, constant strain amplitudes of 6% and 7% were

Table 1
Test details for the prepared SMA cable specimens after training and pretensions.
Test Specimen designation Training laps Initial strain (%) Loading protocol Strain rate (s− 1) Maximum strain (%)

ID
4
A1 V-NP-I 0 0.0 Incremental 6.94 × 10− 10
4
A2 V-NP-C6 0 0.0 Constant 6.94 × 10− 6
4
A3 V-NP-C7 0 0.0 Constant 6.94 × 10− 7
4
B1 T6-NP-I 50 0.0 Incremental 6.94 × 10− 8
4
B2 T7-NP-I 50 0.0 Incremental 6.94 × 10− 8
4
B3 T7-2.0-I 50 2.0 Incremental 6.94 × 10− 10
4
C1 T7-2.0-C-R5 50 2.0 Constant 6.94 × 10− 4.5
3
C2 T7-2.0-C-R10 50 2.0 Constant 1.39 × 10− 4.5
3
C3 T7-2.0-C-R15 50 2.0 Constant 2.78 × 10− 4.5
3
C4 T7-2.0-C-R20 50 2.0 Constant 5.56 × 10− 4.5
4
D1 T7-1.0-D 50 1.0 Diminishing 6.94 × 10− 4.5
4
D2 T7-1.5-D 50 1.5 Diminishing 6.94 × 10− 4.5
4
D3 T7-2.0-D 50 2.0 Diminishing 6.94 × 10− 4.5
4
D4 T7-2.5-D 50 2.5 Diminishing 6.94 × 10− 4.5
4
D5 T7-3.0-D 50 3.0 Diminishing 6.94 × 10− 4.5

3
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 5. Stress-strain curves of the SMA cable specimens.

applied, respectively, and both were repeated 300 times. It is believed 7% for 50 cycles before their official tensile tests. For B1 and B2, the
that these loading protocols can adequately demonstrate the hysteretic loading processes of the official tests start with an initial amplitude of
behavior of the SMA cables under seismic excitations and provide 0.5% and proceed with a strain amplitude incremental interval of 0.5%
enough degradation information for the training procedures of the SMA until the strain amplitude reaches 8%. For B3, an initial strain of 2% was
cables. pre-tensioned on the cable specimen. Moreover, the official loading
Second, incremental loading protocols were employed to investigate process of B3 starts with an initial amplitude of 3% and proceeds with a
the effect of training amplitude, loading amplitude, and initial strain on strain amplitude incremental interval of 1% until the strain amplitude
the mechanical properties of three SMA cables after training (samples reaches 10%, while the unloading process of each stretch cycle always
B1, B2, and B3). The B1 was trained by applying cyclic tensions with ends with the strain returning to the initial strain of pre-tensioning (2%).
constant strain amplitudes of 6% for 50 cycles, while the B2 and B3 were At last, after training and pre-tensioning, nine specimens (C1–C4 and
trained by applying cyclic tensions with constant strain amplitudes of D1-D5) were tested to investigate the influences of the strain rate of

4
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

conducted with a constant loading strain amplitude of 4.5% for 20 cy­

cles, and the unloading process of each stretch cycle ended with the
strain returning to the 2%. Four different loading strain rates (shown in
Table 1) were considered for C1–C4. For D1-D5, different pre-tensioning
procedures with five different initial strains (1.0%, 1.5%, …, and 3.0%)
were conducted. Then the diminishing loading protocols were employed
for the specimens D1-D5, which start with the maximum strain ampli­
tude (4.5%) as the initial amplitude and then proceed with a strain
amplitude diminishing interval of 0.5% until the strain amplitude rea­
ches only 0.5% larger than the initial pre-strain.

2.3. Test results and discussion

2.3.1. Stress-strain responses and performance indexes

The stress-strain responses of the SMA cable specimens under the
above-mentioned cyclic loading protocols are shown in Fig. 5. All the
cable specimens experienced anticipative axial elongations with no
Fig. 6. Illustration of tensile performance indexes for the SMA cables. failure during their loading procedures. For the no-trained cable speci­
mens, flag-shaped behavior with relatively satisfactory recoverability of
loading and the initial strain of pretension on the mechanical properties axial tensile strain is generally shown. However, the “yield” strength and
of the pre-tensioned SMA cables. The nine specimens were both trained transformation plateaus of the cables are less recognizable than those of
by applying cyclic tensions with constant strain amplitudes of 7% for 50 the corresponding monofilament SMA wire, which is similar to the ob­
cycles in advance. For specimens C1–C4, which were both pre-tensioned servations of Ozbulut [41,42], and Fang [43] et al. Moreover, more
with an initial strain of 2%, the official loading processes were remarkable degradations of the “yield” strength and more accumula­
tions of the residual strain are induced for the cable specimens compared

Fig. 7. Strength and stiffness properties of cable specimens: (a) yield strength, (b) peak strength, (c) elastic modulus, (d) EVD, (e) strength of upper and lower
platforms, (f) residual strain,(g) yield strength and peak strength of untrained SMA cable,(h) residual strain of untrained SMA cable, and (i) EVD of untrained
SMA cable.

5
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

with the monofilament SMA wires, which may indicate that the struc­
tural relaxation of the tested cables has a more important impact than
the transformation induced fatigue of the material itself.
To systematically and quantitatively evaluate the cyclic tensile per­
formance of the cable specimens, indexes such as the “yield” strength
σ AM
s , maximum stress σ max , elastic modulus EA , residual strain εres , and
equivalent viscous damping ratio (EVD) ξeqv were further calculated and
discussed. The EVD, which is commonly utilized for equating the energy
dissipated per vibration cycle of a nonlinear system and an equivalent
linear system, can be calculated as:
W
ξeqv = (1)
2πσ max εmax

where the dissipated energy W is the enclosed area of the stress-strain

curve (as shown in Fig. 6) for each loading cycle. The εmax is the
maximum strain of the SMA cable. The other performance indexes are
also illustrated in Fig. 6.

2.3.2. Basic mechanical properties of the virgin SMA cable

To make full use of the mechanical properties of SMA material, we
should try to choose a larger strain amplitude for constant amplitude Fig. 8. Properties of cables with different training amplitudes: (a) yield
cyclic tensile training to obtain a larger and more stable working strain strength, (b) peak strength, (c) EVD, (d) Residual strain.
range after training. It can be observed from Fig. 5(a1) that, when the
strain amplitude is greater than 7%, the tensile response of the SMA amplitude of the SMA cable rises from 6% to 7%, the upper platform
cable begins to appear an obvious martensite hardening phenomenon. strength decreases by 62.9 MPa on average at the same number of
This means that if the constant strain amplitude of the cyclic tensile loading turns, similarly, the lower platform strength decreases by 41.8
training is designed greater than 7%, the accumulation rate of the SMA MPa on average.
cable’s residual deformation may be obviously larger than that of the Fig. 7(f) shows the accumulated residual strains of the specimens as a
cables with an amplitude smaller than 7% [44], which also may lead to a function of cycle numbers under constant cyclic loading. The residual
smaller working strain range after training. Therefore, in this paper, two strains of the two SMA cables with different training amplitudes were
strain amplitudes of 6% and 7% were selected for the constant cyclic compared and analyzed. The residual strain data of the SMA cable from
tensile tests to compare and investigate the post-training mechanical 50 to 300 turns of training were linearly fitted to the slope of the fitted
properties of the SMA cables, and to provide material parameters and line as the cumulative rate of residual strain. The residual strain accu­
verifications for the numerical studies of the proposed SC-HRD in the mulation rate of the SMA cable was faster during the first 50 cycles of
following studies. cyclic loading, and subsequently, the residual strain accumulation rate
Fig. 7 shows the cyclic tensile performance indices of the untrained slowed down. In this regard, the residual strain accumulation rates of the
cable specimens to investigate the basic mechanical properties of the SMA cables with training amplitudes of 6% and 7% were 2.480E-5 and
virgin SMA cable. It can be observed that the constant loading procedure 1.892E-5 for 50 to 300 cycles of loading, respectively.
leads to a more pronounced decrease in the strength and stiffness pa­ For the untrained specimens, the mechanical performance indexes
rameters in the first few cycles, and then the rate of decrease of the are shown in Fig. 7(g–i). It can be observed that, within a certain strain
parameters slows down. In other words, the strength and stiffness pa­ range, the peak strength increases continuously with the increase of the
rameters stabilize when the cable is trained within a certain maximum maximum strain, while the yield strength shows a decreasing trend. The
strain range, and only after a larger strain amplitude is reached does it energy dissipation increases with the number of loaded cycles, and the
cause further significant stiffness degradation or an increase in peak energy dissipation capacity decreases when the strain reaches a certain
stress. A linear fitting of the data was performed for the SMA cable value.
training range of 50–300 turns for 6% training amplitude and 7%
training amplitude, respectively. The peak strength reduction rates of 2.3.3. Properties of cables with different training amplitudes
the two cable specimens undergone training amplitudes of 6% and 7% Fig. 8 shows the comparison of the mechanical properties of SMA
are − 0.103 and − 0.052, respectively. This indicates that the increase in cables stretched at variable amplitudes under two training procedures. It
training amplitude can slow down the decrease in peak strength of the can be observed that the yield strength of the Sample T6-NP-I was larger
SMA cable to some extent. than that of the Sample T7-NP-I at all the strain amplitudes. The yield
As shown in Fig. 7(c and d), both the initial elastic modulus and strength of the SMA cable was stabilized at 110 MPa and 160 MPa at the
energy dissipation of the two cable specimens undergoing different training amplitude of 6% and 7%, respectively. Peak strength in the first
training amplitudes decreased rapidly during the first 50 loading laps. 15 turns of the loading range of 6% training amplitude of the SMA cable
Then, they remain almost constant at a larger number of cycles of larger than that of the SMA cable with a 7% training amplitude, but
loading. At the same number of loading turns, the elastic modulus and when the maximum strain of the loading exceeds the training amplitude
EVD of 7% training amplitude decreased by 0.92 GPa and 0.2% on of both shows the opposite result. This indicates that the training of the
average compared to 6% training amplitude. It is noted that due to the SMA cable will have a certain degree of strength reduction in the range
non-synchronization of stress development within the cross-section, the of the training amplitude.
SMA cables exhibit smaller EVD values than the monofilament SMA Likewise, the energy dissipation of the SMA cable showed a same
wires which typically have an EVD of 7% or above. The EVD of the variation with the peak strength. The energy dissipation of Sample T6-
current cable specimens ranges between 1.4% and 6% depending on NP-I was higher than that of Sample T7-NP-I by 22.5%,16.4%,7.44%,
several factors including loading protocol and maximum strain. From and 8.42% at the 5th, 7th, 9th, and 10th turn of loading, respectively.
Fig. 7(e), the training amplitude of the SMA cables has a certain influ­ This indicates that within the training range, the peak strength of the
ence on the strength of the upper and lower platform. When the training SMA cable will decrease but energy consumption capacity will increase.

6
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 9. Mechanical properties of cables with different strain rates. (a)Peak strength; (b) Elastic modulus (b) Self-centering strength.

Fig. 10. Mechanical properties of cables with different initial strains. (a)Peak stress; (b) Self-centering strength; (c) Elastic modulus.

Fig. 8(d) shows the maximum residual strains at maximum loading 2.3.4.2. Effects of initial strain. Fig. 10 studies the influence of prestress
strains of 1.65% and 2.84% for the SMA cables with training amplitudes on the peak strength, reset force, and elastic modulus of SMA cable. The
of 6% and 7%, respectively. This indicates that an increase in the initial pre-strain has a significant effect on the peak strength of the SMA
training amplitude will improve the residual strain of the cable at larger cable, with a pre-stress of 3.5% compared to 1%, which increases the
strain amplitudes. peak strength by 234.8%. The self-centering force of the cable is stabi­
lized at 51 MPa, 65 MPa, 108 MPa, 130 MPa, 175 MPa, and 235 MPa
2.3.4. Properties of pretensioned SMA cables respectively when the prestressing force is from 1% to 3.5%.
The elastic modulus can be significantly increased by applying a
2.3.4.1. Effects of strain rate. To investigate the effect of loading rate on certain initial pre-strain. The elastic modulus of the pre-stretched SMA
the mechanical properties of SMA cables, experiments were conducted cables ranged from 15 to 24 GPa at initial strains of 1.0%–3.5%, where
on SMA cables with 2% pre-strain and different loading rates after 100 higher initial pre-strains increased the elastic modulus of the SMA ca­
turns of training at 7% strain amplitude. As can be seen from Fig. 9(a), bles. Therefore, in the proposed SC-HRD, the SMA cable can be used in
the peak stresses were stabilized at 367 MPa, 377 MPa, 396 MPa, and the pre-stretched condition so that a higher initial stiffness can be
401 MPa with the increase of loading rate, respectively. The peak obtained.
strength of the SMA cable increased by 9.26% when the loading rate was To sum up, the specimens show stable and superior strength and
increased from 5 mm/min to 20 mm/min. stiffness, and desired self-centering capability and energy dissipation,
Fig. 9(b) shows that the loading rate can appropriately increase the after the cyclic tensioning at 7% strain for more than 100 times.
elastic modulus of SMA cables, which ranges from 6.5 GPa to 8.5 GPa. Therefore, the specimen under cyclic tension for 100 times at 7% strain
Fig. 9(c) shows the “self-centering strength” of the pre-stretched SMA is used as the self-centering element of the proposed SC-HRD in the
cables with different strain rates versus the number of cycles. The self- following theoretical and numerical studies.
centering strength is estimated as the stress of SMA cables when the
strains are unloaded to their respective initial strains. With the increase
in the number of cycles, the self-centering strength of SMA cable with
different strain rates tends to decrease, but the effect of strain rate on the
self-centering force of cable is not obvious.

Fig. 11. The name of each part of the damper and the connection method.

7
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 13. SMA model for superelasticity in software ABAQUS.

4. Numerical investigations

The finite element software ABAQUS was used to numerically

simulate the SMA cable and the high-damping rubber and compare the
experimental data to obtain the appropriate parameters to ensure the
validity of the damper model. The finite element model of the damper
was established for subsequent parametric analysis.

Fig. 12. Working principle diagram of the SC-HRD. 4.1. Modeling of SMA cables

3. Proposed SMA-cables-based SC-HRD The hysteresis properties of the shape memory alloy were simulated
using the superelastic intrinsic model available in the ABAQUS material
3.1. Structure of proposed SC-HRD library. As shown in Fig. 13, the flag model requires only a limited
number of control parameters that can be calibrated based on the test
The self-centering high-damping rubber damper (SC-HRD) is a results of pre-trained and pre-tensioned SMA cables. The SMA filament
hybrid passive control device that strategically combines a viscoelastic rope parameters are referenced to Ref. [34], and the parameters are
device and shape memory alloys in parallel. Fig. 11 illustrates the 3D adjusted during the simulation to achieve a better match between the
renderings of the SC-HRD. In the SC-HRD, the SMA elements are pri­ simulation of the filament rope and the experiment. Solid cells were used
marily employed to achieve re-centering capability while a viscoelastic to model the SMA cable with a diameter of 2.8 mm, a length of 120 mm,
component is used to dissipate input seismic energy. a mesh type of C3D8R, and a linear expansion coefficient of Ni–Ti of
The core of the damper consists of a steel plate fixed by vulcanization 1.1E-05 and the pre-stressing method was applied by the cooling
with high-damping rubber. The inner core is placed in the inner cylin­ method.
der. In this case, the right end of the inner barrel is slotted and welded to In this paper, the simulation is carried out with the equivalent area of
the flange of steel plate 2, and the left end is slotted and in contact with the SMA cable, and only the overall mechanical properties of the SMA
the flange of steel plate 1 and can move against it. Similarly, the flange of cable are considered. The simulated working condition is the variable
steel plate 1 is welded to the slotted left end of the outer barrel. The two amplitude stretching of the SMA cable with different pre-strain after 100
end plates are assembled with holes in the middle of the end plate on turns of training with 7% amplitude strain. The parameters of the cable
steel plates 1 and 2 respectively, and the SMA cable is fixed to the two are shown in Table 2, where Esma is the initial elastic modulus of the
end baffles with anchors. superelastic SAM; σAMs and σ AM
f are the beginning and end stresses of the
martensitic phase transformation induced by the gravitational force of
3.2. Working principle of proposed SC-HRD the superelastic SMA, respectively, while σ MA
s and σ MA
f are the beginning
and end stresses of the inversion of the unloading phase of the supere­
The damper is mainly subjected to axial force in the process of use, lastic SMA, respectively. εL is the maximum accumulated residual strain
and the damper components are set up with a high degree of symmetry. of SMA material during the simulation; α is the tensile-compression
As shown in Fig. 12, when the damper is pressed, steel plates 1 and 2 hysteresis correlation coefficient of hyperelastic SMA. A comparison of
drive the rubber to produce inward displacement, thus causing the the simulated results with the experimental example is shown in Fig. 14.
rubber to be sheared. At the same time, steel plate 1 drives the outer
barrel to the left, and steel plate 2 drives the inner barrel to the right and 4.2. Highly damped rubber simulation validation
acts on the end plate, and the two-produce relative displacement thus
causing the SMA cable to be pulled. When the damper is stretched, steel The hysteresis performance of the highly damped rubber was simu­
plate 1 and 2 drive the rubber to move to both sides, also making the lated numerically using ABAQUS. The simulation of the viscoelastic
rubber shear. At the same time, steel plate 1 drives the outer cylinder to material constitutive model was based on the generalized Maxwell
the right, and steel plate 2 drives the inner cylinder to the left and acts on model, including the hyperelastic constitutive and viscoelastic consti­
the end plate, the two-produce relative displacement so that the SMA tutive models. As shown in Fig. 15, the model is modeled using the
cable is pulled. Due to the restoring force of the prestressed SMA cable ABAQUS standard analysis module, in which the rubber is 400 mm in
and the elastic restoring force of the high-damping rubber, SC-HRD has a length and width and 25 mm in thickness, the steel plate is 15 mm thick,
good restoring effect when unloading. and the rubber is bonded to the steel plate on both sides. The steel plate

8
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Table 2
Key parameters of SMA cable models with different pre-stretched schemes.
Scheme Target pre-strain (%) Recovery force after unloading (MPa) Controlling parameters for the superelastic SMA material model in the software ABAQUS

ID Esma (GPa) σAM

s (MPa) σAM
f (MPa) σMA
s (MPa) σMA
f (MPa) εL α

A 1 45 21.00 100 580 300 50 0.025 0

B 1.5 65 21.00 100 600 200 70 0.025 0
C 2.0 108 21.08 180 610 300 130 0.020 0
D 2.5 130 24.0 200 640 310 130 0.018 0
E 3.0 176 27.1 230 700 370 180 0.018 0
F 3.5 235 30.99 260 720 380 235 0.018 0

Fig. 14. Comparison of simulated tensile responses of the pre-tensioned SMA cables with the past test results.

rubber. The comparison of the simulated results with the tested data
reported in a previous study [45] is plotted in Fig. 16.
The comparison of mechanical indexes is shown in Table 3. The
maximum errors of positive peak stress, negative peak stress, hysteresis
loop area, and residual displacement for the highly damped rubber
simulations and experiments were 7.58%, 6.64%, 6.66%, and 6.94% at
loading amplitudes ranging from 25% to 112%, respectively. As can be
Fig. 15. Finite element model of simulated test pieces. seen from Table 4, the maximum errors of positive peak stress, negative
peak stress, hysteresis loop area, and residual displacement for the
is a hexahedral cell with a mesh type of C3D8R, and the rubber is an highly damped rubber simulations and experiments were 3.11%,
incompressible and isotropic material with a mesh type of C3D8H. The 6.25%,1.65%, and 5.26% at loading frequencies ranging from 0.1 Hz to
upper and lower plates are completely fixed, and RP-3 is coupled to the 1 Hz, respectively. The maximum error is not greater than 8%, which
right end face of the middle plate. indicates that the numerical model can capture the nonlinear mechan­
The Mooney-Rivlin model applies to strain amplitudes from 20% to ical performance of the high-damping rubber at different strain ampli­
150%, so the Mooney-Rivlin model is used to simulate the hyper­ tudes and loading frequencies.
elasticity of rubber and the Prony level to simulate the viscoelasticity of

Fig. 16. Comparison of simulation results with experimental data. (a) Comparison of different shear strain data with 0.3hz loading frequency. (b) Comparison of data
with different loading frequencies for 100% shear strain.

9
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Table 3 Table 5
Comparison of model results with experimental results for different loading Q345 steel parameters.
amplitudes. Elastic modulus (MPa) Plastic strain Poisson’s ratio
Yield stress (N/mm2 )
Evaluating Test Model Error (%)
266000 345 0 0.3
indication
Shear strain 266000 450 0.01 0.3
Damper
25%、50%、75%、100%、112%

Positive peak 83.40、124.52、 78.26、115.07、 6.16、7.58、 4.3. Modeling of SC-HRD assemblies

force (kN) 149.82、176.93、 152.42、174.45、 1.73、1.40、
206.08 209.31、 1.57
The high-damping rubber is assembled with SMA cable and auxiliary
Negative peak − 73.17、-113.84、 − 77.70、-113.89、 6.19、
force (kN) 0.038、 equipment such as a steel barrel. The core and sleeve dimensions of the
− 143.66、- − 153.21、- 6.64、1.01、 damper proposed in this paper are shown in Fig. 17. Inner and outer
177.10、 175.30、 sleeve, left and right end plates, upper and lower steel plates, and steel
− 208.51 − 201.17 3.52 plates 1 and 2 are made of Q345 steel. Q345 steel parameters as Table 5.
Hysteresis loop 1038.46、2622.53 991.76、2557.07 4.50、2.50、
area (J) 4337.68、6619.97 4076.51、6179.37 6.02、6.66、
The dimensions of the left and right end plates are 280 × 145 × 20 mm,
8889.56 8727.84 1.82 and the length of the SMA cable is 450 mm. The dampers are modeled by
Residual 4.2、8.7、 4.1、8.8、 2.38、1.15、 using finite element software ABAQUS, and the high-damping rubber
displacement 4.55、6.94、 and SMA cable simulations were consistent with the previous paper. The
(mm) 6.14
maximum elongation of SC-HRD is designed to be 5%.
11、14.4、16.3 10.5、15.4、17.3

5. Optimization analysis of damper parameters

Table 4 Parametric studies were carried out to investigate the effects of the
Comparison of model results with experimental results for different loading initial strain of the pre-tensioned SMA cables, the sectional area of SMA
frequencies.
cables in proportion to those of rubber core, loading frequency and
Evaluating indication Test Model Error (%) displacement amplitude elements on the axial resistance, energy dissi­
Damper pation, and self-centering potential of the proposed SC-HRD. The main
Loading frequency
parameters studied are the hysteresis response, EVD ratio, and residual
0.1 Hz、0.3 Hz、0.5 Hz、1 Hz
deformation of SC-HRD. The objective is to investigate the optimal
Positive peak force (kN) 166.81、171.28 164.41、174.45 1.44、
preload scheme for SMA cables and the hysteresis response under
1.85、
179.72、203.37 188.71、209.69 5.03、3.11 different loading conditions to achieve the desired hysteresis energy
Negative peak force (kN) − 166.70、- − 164.38、- 1.39、 dissipation as well as the appropriate residual deformation response of
169.84 175.30 3.22、 SC-HRD.
− 175.13、- − 187.64、- 7.14、5.31
199.62 210.22
Hysteresis loop area (J) 5062.12、 5141.99、 1.58、
6161.81 6172.88 0.18、 5.1. Effects of SMA cable cross-sectional area on SC-HRD
7388.78、 7506.54、 1.59、1.65
10407.08 10235.24 The ratio Sr of the SMA cable cross-sectional area to the highly
Residual displacement 11.2、13.6、 12.2、14.4、 6.25、 damped rubber area will largely affect the resetting ability and energy
(mm) 5.88、
15.2、18.8 17.0、18.2 5.26、3.19
dissipation of the damper when the rubber thickness is certain. To study
the influence of its ratio on the damping force, reset capacity, and energy
dissipation capacity of the damper, finite element models of the damper
with Sr of 0, 0.062%, 0.092%, 0.123%, and 0.154% were established
respectively. The 0.3 Hz simple harmonic load was used for loading. The
initial strain and the maximum tensile strain of the SMA cables are
designed as the constant values of 2% and 5%. Fig. 18 shows the force-
displacement curves under various Sr for the SC-HRD at strain ampli­
tudes of 50% and 100%. Fig. 19 shows the variation of the peak damping
force, residual strain, and EVD with the Sr .
It can be observed from Table 6 and Fig. 19 that Sr has a significant
effect on the peak damping force, residual strain, and EVD of the SC-
HRD. Compared with that when Sr is 0, the peak strength of SC-HRD
with Sr equal to 0.123% increases 139% and 47.9% at 50% and 100%
of shear strain, respectively, and the residual strain decreases 95.5% and
95.3%, respectively, and the EVD decreases 54% and 63%, respectively.
Meanwhile, the peak strength of SC-HRD with Sr equal to 0.154%
increased by 165% and 68% at 50% and 100% of shear strain, respec­
tively, and the residual strain decreased by 94.3% and 99.2%, respec­
tively, and the EVD decreased by 58% and 66.4%, respectively,
compared with that when Sr was 0. The results show that when Sr rea­
ches 0.123%, the residual strain reduction of the damper is not signifi­
cant, but the EVD is significantly reduced. In other words, when the
initial strain and maximum tensile strain of the SMA cable are designed
as constants, the self-centering capability increases with increasing Sr ,
Fig. 17. Damper core dimensions. but the energy dissipation capability decreases and the economical ef­
ficiency decreases. Therefore, it is crucial to select the appropriate Sr .

10
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 18. Hysteretic curve of SC-HRD under various Sr at (a) 100% and (b) 50% shear deformations.

Fig. 19. Variation of (a) peak damping force,(b) residual strain, and (c) EVD with different Sr .

increase of SMA pre-strain.

Table 6 The simulated data for 50% and 100% strain amplitudes at different
Simulation data of 50% strain and 100% strain at different Sr . SMA wire rope pre-strains are shown in Table 7. The hysteresis curve
Sr exhibits a significant resetting effect as the SMA pre-strain increases. The
0% 0.062% 0.092% 0.123% 0.154%
peak damper force increases 47.7% and 68.5% at 50% and 100% strain
amplitudes when the SMA pre-strain increases from 1% to 3.5%. When
Peak force (kN) 50% 13.73 22.59 28.82 32.82 36.40
the strain amplitude was 100%, the residual strain for the initial pre-
100% 33.84 35.56 42.79 50.05 56.86
Residual strain 50% 74.00 34.30 14.70 3.30 4.20 strain of 1%–3.5% was calculated as 22.2%, 18.2%, 3.8%, 3.4%,
(%) 100% 80.6 35.63 16.65 3.82 0.68 2.9%, and 1.7%. When the strain amplitude was changed from 50% to
EVD (%) 50% 7.95 4.83 4.08 3.66 3.34 100%, the residual strain for SC-HRD increased by − 0.1%, 1.39%,
100% 16.71 7.93 6.92 6.18 5.62 0.42%, 0.07%, 0.15%, and − 2.29%, respectively. This indicates that the
change in strain amplitude does not have a significant effect on the re­
5.2. Effects of initial pre-strain sidual strain of the damper for the same pre-strain. For the 100% strain
amplitude, for example, the calculated EVDs were 6.95%, 6.81%, and
To investigate the effects of SMA pre-strain on the mechanical 6.2% for pre-strains of 1%–2%. When the pre-strain changed from 1% to
properties of SC-HRD, six tests with pre-strain ranging from 1% to 3.5% 2%, the decrease in EVD was only 0.76%. This indicates that the SC-HRD
with 50% and 100% shear strain were simulated at Sr of 0.123%. Fig. 20 has a more desirable resetting capability and energy dissipation capa­
shows the force-displacement curves under various pre-strain for the SC- bility under the initial pre-strain of 2%.
HRD at strain amplitudes of 50% and 100%. In addition, Fig. 21 illus­
trates the variation of various mechanical properties of SC-HRD with the

Fig. 20. Hysteretic curve of SC-HRD under various pre-strain at (a) 100% and (b) 50% shear deformations.

11
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 21. Variation of (a) maximum force, (b) residual strain, and (c) EVD with initial pre-strain.

increases by 7%, the EVD increases by 77.9%, and the residual strain
Table 7 increases by nearly 80 times at 1 Hz loading compared to 0.1H loading
Simulation data of 50% and 100% strain amplitude under different SMA wire frequency. This indicates that the loading frequency has no significant
rope pre-strain. effect on the peak force of SC-HRD, and the increase of frequency is
initial pre-strain (%) beneficial to the energy dissipation of the damper, but the residual strain
1% 1.5% 2% 2.5% 3% 3.5%
increases.

Peak force (kN) 50% 27.74 28.55 32.82 34.22 37.42 40.97
100% 43.60 44.69 50.05 56.43 60.88 73.48 5.4. Effects of displacement amplitude
Residual strain 50% 22.30 16.80 3.40 3.30 2.70 4.00
(%) 100% 22.20 18.19 3.82 3.37 2.86 1.71
EVD (%) 50% 4.05 3.98 3.66 3.62 3.36 3.38 To investigate the effects of displacement amplitude on the me­
100% 6.95 6.81 6.20 5.30 4.81 3.86 chanical properties of SC-HRD, a finite element model was built at a
loading frequency of 0.3 Hz at five different displacement levels

Table 8
Simulation data of 100% strain amplitude at different loading frequencies.
loading frequency (Hz)

0.1 0.3 0.5 1

Peak force (kN) 49.38 50.05 51.48 52.83

Residual strain (%) 0.51 3.50 21.1 41.27
EVD (%) 5.17 6.18 7.14 9.2

Fig. 22. Hysteretic curve of SC-HRD under various frequencies at 100% shear
deformations.

5.3. Effects of loading frequency

To investigate the effects of loading frequency on the mechanical

properties of SC-HRD, four shear tests with loading frequencies from 0.1
to 1.0 Hz with 100% shear strain were simulated. Fig. 22 shows the
force-displacement curves under various frequencies for the SC-HRD at
strain amplitudes of 100%. Moreover, Fig. 23 illustrates the variation of
various mechanical properties of the SC-HRD with loading frequency.
The hysteresis curve becomes full as the loading frequency increases. Fig. 24. Hysteresis curves for SC-HRD under various shear strain amplitudes at
As can be seen from Table 8 and Fig. 23, the peak force of the damper 0.3 Hz.

Fig. 23. Variation of (a) maximum force, (b) residual strain, and (c) EVD with loading frequency.

12
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

Fig. 25. Variation of (a) peak damping force, (b) EVD, (c) residual strain with shear strain.

Table 9 Table 10
Simulated data of different strain amplitudes at 0.3 Hz loading frequency. List of abbreviations.
strain amplitude (%) Abbreviations Full spellings

25% 50% 75% 100% 112% VE Viscoelastic

SMA Shape memory alloy
Peak force (kN) 50% 27.74 28.55 32.82 34.22 37.42
SC-HRD Self-centering high-damping rubber damper
Residual strain (%) 50% 22.3 16.8 3.4 3.3 2.7
SME Shape memory effect
EVD (%) 100% 6.95 6.81 6.20 5.30 4.81
SE Superelasticity
SVD Superelastic viscous damper
VED Viscoelastic damper
corresponding to shear strains ranging from 25% to 112%. Fig. 24 il­ SCS Self-centering system
lustrates the force-displacement curves for the SC-HRD for all strain DSC Differential scanning calorimeter
amplitudes. Fig. 25 shows the variation of the peak damping force, re­ MS Martensite start temperature
sidual strain, and EVD with the strain amplitude. UTM Universal test machine
EVD Equivalent viscous damping ratio
As can be seen from Table 9 and Fig. 25, The peak damping force
increases almost linearly with the displacement amplitude. The peak
damping force increases by 34.9% over the range of available load consideration, the training scheme of the self-centering elements
amplitudes. the EVD also increases with displacement amplitude, but at (SMA cables) in the proposed SC-HRD is suggested as the 7%
higher displacement amplitudes, the increase in EVD slows down and strain amplitude for 50 cycles to optimize and stabilize its me­
shows a decreasing trend. The residual strain increases from 0.7% for chanical properties.
25% strain to 3.3% for 100% strain. Despite the 2.72% increase in re­ (3) The initial stiffness and residual self-centering strength of the
sidual strain, the SC-HRD still has a very good resetting capability. Be­ after-trained SMA cables, which are essential for the SC-HRD
sides, the decrease in residual strain at a strain amplitude of 112% does application, can be enhanced by increasing the initial strain of
not follow the overall pattern, which may be due to errors during the pre-tensioning and the strain rate of loading.
rubber simulation. (4) The incorporating of pre-tensioned SMA cables improved the
initial stiffness and peak load of the SC-HRD by 253%–941% and
6. Conclusions 5.1%–68.0%, respectively, and decreased the residual deforma­
tion by 55.8%–99.2%, compared with the HRD without SMA
In this study, an innovative SMA-cable-controlled self-centering cables. After a comprehensive consideration of the initial stiff­
viscoelastic damper is proposed and investigated. The influence of initial ness, energy dissipation, and residual deformation of the pro­
strain, loading amplitude, and strain rate on the mechanical properties posed SC-HRD, the initial strain and relative area ratio of the pre-
of pre-tensioned SMA cables is also tested. The cyclic behavior of the 7 tensioned SMA cables in the SC-HRD are suggested to be designed
× 7 × 0.4 mm SMA cables after training and pre-tensioning for the SC- as about 2% and 0.123%, respectively.
HRD application was investigated. The finite element model of the (5) The increasing of loading frequency can increase the bearing
proposed SC-HRD was established and verified. Subsequently, para­ capacity and energy dissipation of the SC-HRD but decrease its
metric analyses were carried out on several important design parameters self-centering capacity. Compared with the SC-HRD with a
of the SC-HRD to optimize its mechanical performance. The used ab­ loading frequency of 0.1 Hz, the SC-HRD with a loading fre­
breviations are summarized in Table 10. The following conclusions can quency of 1 Hz exhibits a 7% larger peak load, a 77.95% higher
be drawn from this study: EVD ratio, but a significantly increased residual deformation.

(1) The mechanical properties of SMA cables can be stabilized by CRediT authorship contribution statement
proper training procedures. When the working strain amplitude
of SMA cables does not exceed the difference between the pre- Hui Qian: Conceptualization, Methodology, Project administration,
training strain amplitude and the maximum accumulated resid­ Funding acquisition, Writing - review & editing. Dexin Wei: Concep­
ual strain during training, the training with constant amplitudes tualization, Methodology, Investigation, Writing - original draft. Yifei
can greatly reduce the residual strain within the subsequent Shi: Conceptualization, Methodology, Project administration, Investi­
designed working strain amplitude of SMA cables. gation, Writing - original draft. Zongao Li: Visualization, Investigation,
(2) Compared with the SMA cable trained at 6% amplitude for 50 Data curation. Hongnan Li: Conceptualization, Writing - review &
cycles, the SMA cable trained at 7% amplitude for 50 cycles has editing.
the similar elastic modulus and EVD ratio, and a lower yield
strength, but the larger peak strength and working strain ampli­
tude as well as the smaller peak strength reduction rate and re­ Declaration of competing interest
sidual strain accumulation rate. After comprehensive
The authors declare that they have no known competing financial

13
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

interests or personal relationships that could have appeared to influence Mater Syst Struct 2017;28(12):1533–52. https://doi.org/10.1177/
1045389X16672593.
the work reported in this paper.
[20] Johnson R, Padgett JE, Maragakis ME, DesRoches R, Saiidi MS. Large-scale testing
of nitinol shape memory alloy devices for retrofitting of bridges. Smart Mater
Data availability Struct 2008;17:035015. https://doi.org/10.1088/0964-1726/17/3/035018.
[21] Abdulridha A, Palermo D, Foo S, Vecchio FJ. Behavior and modeling of superelastic
shape memory alloy reinforced concrete beams. Eng Struct 2013;49:893–904.
Data will be made available on request. https://doi.org/10.1016/j.engstruct.2012.12.041.
[22] Shrestha KC, Araki Y, Nagae T, Koetaka Y, Suzuki Y, Omori T, Sutou Y, Kainuma R,
Acknowledgments Ishida K. Feasibility of Cu-Al-Mn superelastic alloy bars as reinforcement elements
in concrete beams. Smart Mater Struct 2013;22(2):025025. https://doi.org/
10.1088/0964-1726/22/2/025025.
The authors express their gratitude to the National Natural Science [23] Mas B, Biggs D, Vieito I, Cladera A, Shaw J, Martínez-Abella F. Superelastic shape
Foundation of China (No. 51978631) and Supported by the Program for memory alloy cables for reinforced concrete applications. Construct Build Mater
2017;148:307–20. https://doi.org/10.1016/j.conbuildmat.2017.05.041.
Innovative Research Team (in Science and Technology) in the University [24] Saiidi MS, Wang HY. Exploratory study of seismic response of concrete columns
of Henan Province (No. 23IRTSTHN006). We also appreciate the assis­ with shape memory alloys reinforcement. ACI Struct J 2006;103(3):436–43.
tance provided by Mr. Hongyi Qian in Grammar polishing. https://www.webofscience.com/wos/alldb/full-record/WOS:000237129100014.
[25] Saiidi MS, O’Brien M, Sadrossadat-Zadeh M. Cyclic response of concrete bridge
columns using superelastic nitinol and bendable concrete. ACI Struct J 2009;106
References (1):69–77. https://www.webofscience.com/wos/alldb/full-record/WOS:0002622
91800008.
[1] Symans MD, Charney FA, Whittaker AS, Constantinou MC, Kircher CA, [26] Qian H, Li ZA, Pei JZ, Kang LP, Li HN. HN. Seismic performance of self-centering
Johnson MW, McNamara RJ. Energy dissipation systems for seismic applications: beam-column joints reinforced with superelastic shape memory alloy bars and
current practice and recent developments. J Struct Eng 2008;134(1):3–21. https:// engineering cementitious composites materials. Compos Struct 2022;294:115782.
doi.org/10.1061/(ASCE)0733-9445. 2008)134:1(3). https://doi.org/10.1016/j.compstruct.2022.115782.
[2] Soong TT, Dargush GF. Passive energy dissipation system structural engineering. [27] Qian H, Kang LP, Guo YC, Li ZA, Shamim AS. Experimental study on seismic
Chichester: Wiley; 1997. behavior of self-centering shear wall reinforced with shape memory alloy bars and
[3] Guo L, Wang JF, Wang WQ, Wang HL. Experimental, numerical, and analytical engineered cementitious composites. China Civ Eng J 2020;53(10):51–61+71.
study on seismic performance of shear-bending yielding coupling dampers. Eng https://doi.org/10.15951/j.tmgcxb.2020.10.005.
Struct 2021;244:112724. https://doi.org/10.1016/J.ENGSTRUCT.2021.112724. [28] Fraternali F, Santos F. Mechanical modeling of superelastic tensegrity braces for
[4] Xu ZD, Liao YX, Ge T, Xu C. Experimental and theoretical study of viscoelastic earthquake-proof structures. Extr. Mech. Lett. 2019;33:100578. https://doi.org/
dampers with different matrix rubbers. J Eng Mech 2016;142(8):04016051. 10.1016/j.eml.2019.100578.
https://doi.org/10.1061/(ASCE)EM.1943-7889.0001101. [29] Santos FA, Caroco C, Amendola A, Miniaci M, Fraternali F. 3D tensegrity braces
[5] Ranaei Omid, Ali Akbar Aghakouchak. Experimental and numerical study on with superelastic response for seismic control. Int J Multiscale Comput Eng 2022;
developed elastomeric layers based on natural and butyl matrix rubbers for 20(5):53–64. https://doi.org/10.1615/IntJMultCompEng.2022041968.
viscoelastic dampers. Mech Time-Dependent Mater 2021;26(1):211–33. https:// [30] Shi F, Zhou Y, Ozbulut OE, Ren FM. Hysteretic response and failure behavior of an
doi.org/10.1007/S11043-021-09484-2. SMA cable-based self-centering brace. Struct Control Health Monit 2022:e2847.
[6] Kishi H, Kuwata M, Matsuda S, Asami T, Murakami A. Damping properties of https://doi.org/10.1002/stc.2847.
thermoplastic-elastomer interleaved carbon fiber-reinforced epoxy composites. [31] Christopoulos C, Tremblay R, Kim HJ, Lacerte M. Self-centering energy dissipative
Compos Sci Technol 2004;64(16):2517–23. https://doi.org/10.1016/j. bracing system for the seismic resistance of structures: development and
compscitech.2004.05.006. validation. J Struct Eng 2008;134(1):96–107. https://doi.org/10.1061/(ASCE)
[7] Patri M, Reddy CV, Narasimhan C, Samui AB. Sequential interpenetrating polymer 0733-9445. 2008)134:1(96).
network based on styrene-butadiene rubber and poly alkyl methacrylates. J Appl [32] Chen JB, Fang C, Wang W, Liu YQ. Variable-friction self-centering energy-
Polym Sci 2007;103(2):1120–6. https://doi.org/10.1002/app.24338. dissipation braces (VF-SCEDBs) with NiTi SMA cables for seismic resilience.
[8] Christopoulos C, Montgomery M. Viscoelastic coupling dampers (VCDs) for J Constr Steel Res 2020;175:106318. https://doi.org/10.1016/j.jcsr.2020.106318.
enhanced wind and seismic performance of high-rise buildings. Earthq Eng Struct [33] Miller DJ, Fahnestock LA, Eatherton MR. Development and experimental
Dynam 2013;42(15):2217–33. https://doi.org/10.1002/eqe.2321. validation of a nickel–titanium shape memory alloy self-centering buckling-
[9] Fang XJ, Zhou Y, Zhang C. Seismic analysis of frame core tube structure with lead restrained brace. Eng Struct 2012;40:288–98. https://doi.org/10.1016/j.
viscoelastic coupling beam damper. J Civ Eng 2018;51:1–7. https://doi.org/ engstruct.2012.02.037.
10.15951/j.tmgcxb.2018.s2.001 (in Chinese). [34] Shi YF, Qian H, Kang LP, Li ZA, Xia LK. Cyclic behavior of superelastic SMA cable
[10] Di Sarno L, Elnashai AS. Innovative strategies for seismic retrofitting of steel and and its application in an innovative self-centering BRB. Smart Mater Struct 2021;
composite structures. Prog Struct Eng Mater 2005;7(3):115–35. https://doi.org/ 30:095019. https://doi.org/10.1088/1361-665X/ac1907.
10.1002/pse.195. [35] Fahiminia M, Zahrai SM. Seismic performance of simple steel frames with
[11] Mccormick J, Aburano H, Ikenaga M, Nakashima M. Permissible residual buckling-restrained knee braces & SMA to reduce residual displacement. Soil
deformation levels for building structures considering both safety and human Dynam Earthq Eng 2020;137:106268. https://doi.org/10.1016/j.
elements. Proc. 14th World Conf. Earthquake Engineering 2008. Seismological soildyn.2020.106268.
Press of China, Beijing, https://www.researchgate.net/publication/26829499 [36] Ghowsi AF, Sahoo DR. Seismic response of SMA-based self-centering buckling-
8_Permissible_residual_deformation_levels_for_building_structures_considering_b restrained braced frames under near-fault ground motions. Soil Dynam Earthq Eng
oth_safety_and_human_elements. 2020;139:106397. https://doi.org/10.1016/j.soildyn.2020.106397.
[12] Qian H, Li HN, Song GB. Experimental investigations of building structure with a [37] Qiu CX, Du XL, Teng J, Li ZH, Chen C. Seismic design method for multi-story SMA
superelastic shape memory alloy friction damper subject to seismic loads. Smart braced frames based on inelastic displacement ratio. Soil Dynam Earthq Eng 2021;
Mater Struct 2016;25:125026. https://doi.org/10.1088/0964-1726/25/12/ 147:106794. https://doi.org/10.1016/j.soildyn.2021.106794.
125026. [38] Silwal B, Ozbulut OE, Michael RJ. Seismic collapse evaluation of steel moment
[13] Qian H, Li HN, Song GB, Recentering W Guo. Shape memory alloy passive damper resisting frames with superelastic viscous damper. J Constr Steel Res 2016;126:
for structural vibration control. Math Probl Eng 2013:963530. https://doi.org/ 26–36. https://doi.org/10.1016/j.jcsr.2016.07.002.
10.1155/2013/963530. 2013. [39] Lv ZH, Wei PY, Fang C. Seismic vulnerability and risk analysis of SMA-VED hybrid
[14] Liang D, Zheng Y, Fang C, Yam MCH, Zhang CT. Shape memory alloy (SMA)-cable- self-resetting braced steel frames. World Earthq Eng 2022;38(2):10–20. https://
controlled sliding bearings: development, testing, and system behavior. Smart doi.org/10.19994/j.cnki.WEE.2022.0027 (in Chinese).
Mater Struct 2020;29:085006. https://doi.org/10.1088/1361-665x/ab8f68. [40] Reedlunn B, Daly S, Shaw J. Superelastic shape memory alloy cables: Part I -
[15] Zheng Y, Dong Y, Li YH. Resilience and life-cycle performance of smart bridges isothermal tension experiments. Int J Solid Struct 2013;50(20–21):3009–26.
with shape memory alloy (SMA)-cable-based bearings. Construct Build Mater https://doi.org/10.1016/j.ijsolstr.2013.03.013.
2018;158:389–400. https://doi.org/10.1016/j.conbuildmat.2017.10.031. [41] Ozbulut OE, Daghash S, Sherif MM. Shape memory alloy cables for structural
[16] Li S, Dezfuli FH, Wang JQ, Alam MS. Performance-based seismic loss assessment of applications. J Mater Civ Eng 2015;28(4):04015176. https://doi.org/10.1061/
isolated simply-supported highway bridges retrofitted with different shape (ASCE)MT.1943-5533.0001457.
memory alloy cable restrainers in a life-cycle context. J Intell Mater Syst Struct [42] Sherif MM, Ozbulut OE. Tensile and superelastic fatigue characterization of NiTi
2020;31(8):1053–75. https://doi.org/10.1177/1045389X20906018. shape memory cables. Smart Mater Struct 2018;27(1):015007. https://doi.org/
[17] Li S, Dezfuli FH, Wang JQ, Alam MS. Displacement-based seismic design of steel, 10.1088/1361-665X/aa9819.
FRP, and SMA cable restrainers for isolated simply supported bridges. J Bridge Eng
2018;23(6):04018032. https://doi.org/10.1061/(ASCE)BE.1943-5592.0001231.
[18] Zhang YF, Hu XB, Zhu SY. Seismic performance of benchmark base-isolated bridges
with superelastic Cu–Al–Be restraining damping device. Struct Control Health
Monit 2009;16:668–85. https://doi.org/10.1002/stc.327.
[19] Aryan H, Ghassemieh M. A superelastic protective technique for mitigating the
effects of vertical and horizontal seismic excitations on highway bridges. J Intell

14
H. Qian et al. Soil Dynamics and Earthquake Engineering 168 (2023) 107850

[43] Fang C, Zheng Y, Chen JB, Wang W. Superelastic NiTi SMA cables: thermal- [45] Zhou Y, Shi F, Ozbulut OE, Xu HF, Zi DM. Experimental characterization and
mechanical behavior, hysteretic modeling and seismic application. Eng Struct analytical modeling of a large-capacity high-damping rubber damper. Struct
2019;183:533–49. https://doi.org/10.1016/j.engstruct.2019.01.049. Control Health Monit 2018;25(6):e2183. https://doi.org/10.1002/stc.2183.
[44] Qian N, Zhu LH, Han W, Zhao C. Experimental study on mechanical properties of
large NiTi superelastic shape memory alloy bars. Smart Mater Struct 2021;31(1):
015024. https://doi.org/10.1088/1361-665X/ac3c00.

15