Journal of Constructional Steel Research 79 (2012) 195203

Contents lists available at SciVerse ScienceDirect

Journal of Constructional Steel Research

Static and dynamic cyclic performance of a low-yield-strength steel shear

panel damper
Zhang Chaofeng a,, Zhang Zhisheng b, Zhang Qiuju a
a
b

School of Mechanical Engineering, Jiangnan University, China

School of Mechanical Engineering, Southeast University, China

a r t i c l e

i n f o

Article history:
Received 20 April 2012
Accepted 30 July 2012
Available online 31 August 2012
Keywords:
Low-yield-strength steel
Shear panel damper
Static and dynamic tests
Strain rate
Temperature
Lowcycle fatigue

a b s t r a c t
Low-yield-strength steel 100 (LYS100) is widely applied to design a metallic shear panel damper for its high
ductility. A low-yield-strength steel shear panel damper (LYSPD) with the maximum shear strain of 70% is
developed and veried by static incremental cyclic loading in previous research. In this paper, further research on the performances of the developed LYSPD including the fatigue characteristic is carried out by static
and dynamic constant cyclic tests. Four different shear strain amplitudes (20%, 30%, 40% and 50%) are selected
in both static and dynamic tests. Two frequencies, 0.5 Hz and 1 Hz, are adopted for each of the four amplitudes in dynamic tests. Large differences such as stress softening, fatigue cycle deterioration, and temperature increase caused by high strain rate and internal friction are observed in dynamic tests. The test results
suggest that the seismic performance of the LYSPD may be overestimated by static tests and the dynamic
tests are essential to guarantee the reliability of the LYSPD.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
Shear panel dampers (SPD) are widely developed to dissipate the
energy and to reduce or avoid the damage of the primary structures
such as bridge and building, as demonstrated in Fig. 1 [1,2]. Generally
speaking, the passive metallic damper for seismic applications in structure must exhibit: a) adequate elastic stiffness to withstand small earthquake and wind; b) yield strength not exceeding that of the structure;
c) high energy absorbability; d) stable hysteretic force-displacement
response which can be modeled easily; e) good low-cycle fatigue performance. As the low-yield-strength steel 100 (LYS100, yield strength
100 N/mm2) possesses such merits as low yield strength, large deformation capacity and good low-cycle fatigue performance, it is undoubtedly a proper metallic material for SPD [35].
Failure in the LYSPD caused by repeated loads is typically attributed to
the accumulation of a small number of cycles of large amplitude strains,
which are normally considered within the plastic range for the gross
section. Tensioncompression cyclic tests are conducted on LYS100 for
investigating the stressstrain behavior within the plastic range and the
potential low-cycle fatigue characteristics in most researches [6,7]. However, the performance of the LYSPD would be affected by the stress concentration resulting from the emergence of the out-of-plane buckling,
the formation of tension eld and the material deterioration caused by
welding and so on. That means the fatigue performance of material
LYS100 can't reect the real fatigue performance of the LYSPD. Therefore,
the LYSPD mechanical properties and low-cycle fatigue characteristics
Corresponding author. Tel.: +86 183 5150 0918.
E-mail address: zcf830703@163.com (C. Zhang).
0143-974X/$ see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jcsr.2012.07.030

should be veried by tests directly. On the other hand, for the limitation
of test equipments, the full-scale LYSPD cyclic tests are concentrated on
static tests currently [8,9]. However, the temperature of the steel
LYS100 will increase for the high speed loading and internal friction.
And the stressstrain relation would also be affected by high temperature
[10,11]. To understand the performance of the LYSPD under dynamic
loading and improve the credibility of application, the dynamic cyclic
tests are put forward in this paper.
Recently, extensive experimental research has been undertaken at
the Seismic Research Center of Aichi Institute of Technology in Japan
in order to investigate the energy dissipation capacity (deformation
capacity) of shear panels made from LYS100. A compact LYSPD serving as the bridge damper, with 70% shear strain (horizontal displacement/height) which is the largest deformation capacity at present in
the world [12], is developed and veried by 5% shear strain static
incremental hysteretic loading. Here, further investigation on the
static and dynamic hysteretic performance of the developed LYSPD
is attached importance on. Useful information taken as the preliminary design references such as mechanical properties and low-cycle
fatigue are provided, compared and discussed based on constant static and dynamic tests.
2. Specimen details and test setup
2.1. Specimen
Tensile coupon tests for LYS100 are conducted and the obtained
stressstrain curves are shown in Fig. 2. The yield strength y dened
by the 0.2% offset value of LYS100 0.2 is 100 N/mm2 and the elongation

196

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

(a) Bridge

(b) Building

moment when the horizontal force acts on the LYSPD. The shear panel
is welded to the top and bottom beams. Except that the material of the
link is SM490, the other material of link mechanism is made from
SS400. The LYSPD is xed on the bottom x plate by 12 M24 high strength
bolts and the seismic function could be recovered after the earthquake as
soon as possible because it is very easy to be replaced or re-centered.

Superstructure

2.2. Test setup

Pier

As shown in Fig. 4, one head of the 100 t MTS dynamic actuator is

xed on the back strength wall and another head is connected with the
beam that can move horizontally. Two counterforce devices are installed
on the beam to provide the force acting on the topside of the LYSPD.
Fig. 1. Application of the SPD.

2.3. Measurement equipment

reaches 60%, which is about 3 times larger than thatp

of SS400 ordinary
steel. The corresponding yield shear strength y 0:2 = 3 is 57.7 N/mm2.
The shape and size of specimen are shown in Fig. 3(a). Except the
welding seam (width 10 mm) located at left and right sides of the
panel, the pure size of the shear panel that cuts from 24 mm to
12 mm at the center of the original panel is characterized with height
d = 120 mm and width w = 160 mm respectively. The yield shear
force F y y  w  t is 110 kN. The width/thickness ratio Rw is dened as follows
s


12 12 y

k2 E

2.4. Test plan

an expression widely known where E is the modulus of elasticity (=

206 10 3 N/mm 2), is Poisson's ratio ( = 0.3), t (= 12 mm) and
w are the thickness and the width of the shear panel respectively.
The coefcient k, known as shear buckling coefcient, depends on
2

k 5:35 4:0  w=d w=d1

2

k 5:35  w=d 4:0 w=d1

2
3

as w=d 1:29 is larger than 1 in this study, the width/thickness ratio

R of the shear panel is only 0.065 according to Eqs. (1) and (3). 0.065
is a fairly small value compared with that of in the previous researches. Under this circumstance, the out-of-plane buckling is not
prone to producing. At the same time, the cracks at four panel corners
resulting from the stress concentration under the hysteretic loading
are observed.
Ribs are widely applied to the design of the shear panels or other
structures for the purpose of improving the stress concentration located
at the corners. Although the stress concentration could be improved
when the shear angle is small, the rib ends are easy to crack under the
hysteretic loading with large shear angle for the material deterioration
resulting from welding. Moreover, the stress concentration at panel
four corners would be more serious for the existence of welding intersections. Taking these two points into consideration, two stiffeners
(thickness t = 24 mm, height h = 50 mm) at the up and down sides of
the panel are shaped from original panel instead of welding. Local plastic hinges at the corners of effective panel area are separated from the
rib ends by this optimization. In addition, in order to prevent the stress
concentration caused by sharp variable cross-section, the arcs with
47 mm radius are introduced to connect two different panel thicknesses. The ribs are the thin panels that are also made from LYS100
with length 266 mm, width 72 mm and thickness 12 mm. The deformation capacity of the shear panel can reach 70% shear strain by the optimization [12].
As shown in Fig. 3(b, c), a link mechanism (test xture) is also
designed to simulate the separation function and suppress the bending

The horizontal displacement/panel effective height is dened as shear

strain . The waveforms are applied with constant amplitudes of 20%,
30%, 40% and 50% shear strain respectively in both static and dynamic
tests. One example of the loading pattern when the shear strain is 50%
(displacement 60 mm) is illustrated in Fig. 5. The cycles are fully reversed
with strain ratio R=min /max=1 and only varies in shear strain
amplitude and shear strain rate v. The shear strain rate v of
quasi-static loading is 0.4%/s (0.5 mm/s). Two kinds of frequencies
0.5 Hz and 1 Hz that represent the typical vibration frequencies of bridge
systems are selected as the dynamic loading frequencies [13].
The notation of the static specimen is composed of the abbreviation
of static (ST) and shear strain amplitude (20% simplied as 20) such as
ST20. The notation of the dynamic specimen is characterized by the abbreviation of dynamic (D), frequency (0.5 Hz simplied as 05) and
shear strain amplitude (20% simplied as 20) such as D05-20. The detail
loading information of the twelve specimens is listed in Table 1.

500

SS400
400

Stress (kN/mm )

w
Rw
t

In this experiment, force is measured through the actuator load cell

(precision 0.5 kN) and displacement is measured through laser extensometers (laser extensometer precision 0. 05 mm). In order to eliminate
the errors caused by test equipment to secure the accurate data of damper
horizontal displacement, two laser extensometers (LEX1 and LEX2) are
used to measure the horizontal displacement at the LYSPD's top and bottom beams. The difference of the LEX1 and LEX2 values is taken as the
pure horizontal displacement of the LYSPD. Furthermore, the temperature of the LYSPD is measured by non-contact temperature sensor
TH6300R on real-time.

LYS100
300

200

100

20

40

Strain (%)
Fig. 2. Tensile coupon experiment results.

60

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

(a) Dimension of shear panel

(b) Panel with link mechanism

197

(c) Side view

Fig. 3. Specimen details (units: mm).

3. Results and discussion

3.1.1. Damper force history of the rst cycle

3.1. Hysteretic characteristic

3.1.1.1. Strain hardening under virgin loading. In general, responding to the

rst reversal of the loading pattern (Fig. 5), the cyclic response of the
shear panels follows hysteretic stressstrain behavior starting with a positive shear excursion that includes the transition from elastic to plastic.
Strain hardening, when stress increases with the increase of shear strain,
is ordinarily observed in most researches in this stage under virgin loading. Therefore, the forces at the rst reversal of the loading pattern are
usually the rst quarter cycle maximum damper forces (MDFq, the rst
MDF at the positive direction) in the hysteretic curves. The MDFq of static
tests is illustrated as in Fig. 8 and it increases by 33% averagely when the
shear strain amplitude increases from 20% to 50%. All of the useful test results are also listed in Table 2.

The static and dynamic force-shear strain hysteretic curves are

shown in Fig. 6. X axis represents shear strain amplitude while Y
axis represents shear force. The rst row of Fig. 6 is the static hysteretic curves and the shear strain amplitude is in ascending order.
The dynamic results of frequency 0.5 Hz and 1 Hz follow as the
second row and the third row respectively. The shape of static hysteretic curves is demonstrated as spindle while the appearance of
dynamic hysteretic curves is showed as rectangular approximately. Since the maximum damper force (MDF) at the maximum
damper displacement (MDD) in each cycle is slightly larger than
the damper force when the displacement is 0 in static test, this difference is negligible. The maximum damper force can be taken as
the effective damper force of the plastic range and both static and
dynamic hysteretic curves could be nearly modeled as perfect
elasticplastic.
Except the rst cycle, the MDFs at two loading directions are nearly
the same from the second cycle in all of the specimens which exhibit
good symmetry without the occurrence of Bauschinger effect. Therefore,
the MDF histories at the positive loading direction, within each cycle
recorded during the 12 individual tests of three types of loading frequencies, are plotted in Fig. 7 for comparison. The damper forces all increase in
the MDFs of both static and dynamic tests during the rst cycle. And the
MDFs also increase with the increase of the shear strain. Subsequently,
the distinct difference of MDF histories between static and dynamic
tests is observed from the second cycle. The MDFs show as a constant
in static tests with the increase of cycles before cracking (Fig. 7(a))
while they decrease linearly with the increase of cycles in dynamic tests
(Fig. 7(b) and (c)).

3.1.1.2. Strain rate hardening. Besides strain hardening observed in the

virgin loading stage, the strain rate hardening also emerges. The MDFqs
of dynamic tests are generally higher than those of static tests (Fig. 8
dash line). The strain rate hardening is distinct when the shear strain is
30% and 40% while it is small when the shear strains are 20% and 50%.
Compared with the MDFq of ST20, the strain rate hardening does not appear in D05-20 while it is observed in D10-20. It can be inferred that the
strain rate hardening is not signicant when the strain rate is low. In addition, there are no MDFq differences between static and dynamic tests
when the shear strain is 50%. It can be considered that there is small
room for the strain rate hardening when the shear strain amplitude is
too large. This phenomenon where the damper force remains as a constant when the shear strain is larger than 50% is also observed in static
incremental hysteretic loading tests of previous study [12]. The MDFs
of D10-30 and D10-40 resulting from strain rate hardenings increase
nearly 20% in the dynamic loading tests on the basis of the results of
the static tests.

Loading beam

100t MTS Actuator

Specimen

Fig. 4. Test setup.

198

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

80

first reversal

Displacement(mm)

60
40
20
0
0

10

20

30

-20

Time(t)

-40
-60
-80

second reversal

veried by two specimens D05-40 and D10-20. In these two experiments, two different shear strain amplitudes 20% and 40% are chosen
respectively while the shear strain rates are the same. However, the
same temperature increase rate of these two specimens is observed.
Subsequent cycles of the same shear strain amplitude could result in
the MDFs less than the previous cycles from the second cycle, corresponding to the stress softening caused by temperature in dynamic
tests. As the damper force deterioration is affected by high temperature
based on different shear strain rates, it is more convenient to establish
the relation between shear strain rate and damper force deterioration
directly. The relation between damper force decrease rate and shear
strain rate from the second cycle is shown in Fig. 10. The linear relation
between damper force decrease rate and shear strain rate can be represented as follows:
F v 0:17v 4:840%=secbv b200%=sec; 20%bb50% :

Fig. 5. Loading pattern ( = 50%).

3.1.1.3. Cyclic hardening. The unloading from the rst reversal to the second reversal of the loading pattern (Fig. 5) follows the elastic slope and
initiates plastic behavior in negative shear at a cyclic negative shear
force that is larger than the MDFq. At the second reversal (Fig. 5), the
unloading again follows the elastic slope and initiates plastic behavior at
a cyclic positive MDF that is almost the same with the cyclic negative
MDF completing the rst cycle. That means the cyclic hardening mainly
occurs during the rst reversal and the second reversal. The MDFs at the
end of the rst cycle (Fig. 8 solid line) are about 3050% larger than that
of the MDFq s (Fig. 8 dash line).
The strain hardening and cyclic hardening based on static test results have been already incorporated into the hysteretic model analysis
for predicting the damper strength accurately in most cases while the
strain rate hardening is seldom paid attention to. It is because the dynamic test devices are not popularized and the static test devices are
dominated in the test systems at present. However, seen from Fig. 8,
the strain rate hardening is still obviously at the end of the rst cycle.
Thus, it is essential to be taken into the consideration of MDF in the dynamic loading unless the loading speed is fairly small or the stress is
fully exerted under the large shear strain in the static tests.
3.1.2. MDF history from the second cycle
Typical structural components can dissipate the heat more efciently under low strain rate or proximity to elastic range. After reaching a
MDF in the rst cycle, the remaining cycles follow the trends of no
damper force increase caused by cyclic hardening and no damper
force deterioration caused by the slight developed warm temperatures
until failure in static tests. However, rapid working of the material
through plastic strains can result in temperature increase that can inuence the mechanical properties of the LYSPD. The temperature increase
caused by strain rate under the constant strain tests is shown in Fig. 9.
Temperature increase rate is linear with the shear strain rate while it
has weak or no correlation with the shear strain amplitude. It could be
Table 1
Test plan.

Static (ST)

Dynamic
(D05)

Dynamic
(D10)

Specimen

f (Hz)

T (Sec)

(%)

v (%/s)

ST20
ST30
ST40
ST50
D05-20
D05-30
D05-40
D05-50
D10-20
D10-30
D10-40
D10-50

0.4

0.5

1.0

20
30
40
50
20
30
40
50
20
30
40
50

40
60
80
100
80
120
160
200

3.1.3. Discussion on the damper hysteretic characteristic

After the displacement-base development obtained by improving
the deformation capacity of the LYSPD, the major challenge is the design
of rational and effective model for analyzing the hysteretic performance
that is capable of predicting the structural response. Two simplied
modes such as bi-linear force-displacement relation model and equivalent linearization model have been widely incorporated in several codes
like Japan Road Association [13], AASHTO [14] and Eurocode 8 [15].
3.1.3.1. Force-displacement bi-linear model. In both static and dynamic
tests, the majority of stress hardenings include strain hardening, cyclic
hardening, and strain rate hardening occurring within the rst cycle
and reaching the MDFs of each test. These hardenings can be simplied
and the force-displacement hysteretic curve could be modeled as perfect elasticplastic briey, as shown in Fig. 11. The elastic slope is decided according to the unloading part of hysteretic curve.
As mentioned above, the MDF in static incremental hysteretic curves
is a constant around 610 kN (5.5 times of Fy) when the shear strain is
larger than 50%. In addition, the MDFs are narrowed to a small range
around 610 kN when shear strain amplitude is larger than 30% in dynamic tests with different loading frequencies (Fig. 8 solid line). Therefore, the effective damper force of the plastic range of the supposed
bi-linear model could be set as 610 kN approximately. On the other
hand, it is observed that the MDFs decrease dramatically after the second cycle in constant dynamic tests when the shear strain amplitude
is 50%. Therefore, the damper hysteretic curves could be supposed as
perfect elastic-plastic without force deterioration if the cycle number
is no more than 3 cycles when the shear strain is larger than 50% in
the random wave. Furthermore, heat will be more easily be dissipated
in the random wave, and heat is absolutely overestimated in the constant dynamic tests. Thus, whether the force deterioration should be
taken into consideration or not needs to be further veried by random
wave dynamic tests of the LYSPD which will be conducted in the next
research.
3.1.3.2. Equivalent linearization model. It is generally accepted that the
KelvinVoigt model is able to model the damping characteristics of hysteretic type dampers and to describe the energy dissipation of the specimen vibrator system. For practical use it is sometimes more preferable
to express the device properties in an equivalent linearization system
[16,17]. This is basically a single degree of freedom oscillator with an effective stiffness Kef and an equivalent damping ratio eq.
A typical force-displacement hysteretic curve is shown in Fig. 12,
the Kef and eq could be obtained by static and dynamic constant experimental hysteretic curves. Kef and eq are dened as follows,


MDF MDF 


K ef 
MDD MDD 

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

(c) D10-20

800

800

400

400

400

0
-400
-800
-60

-30

30

0
-400
-800
-60

60

30

-400
-800
-60

60

-30

(f) D10-30

800

800

400

400

400

-800
-60

-30

30

0
-400
-800
-60

60

F orce (kN)

(e) D05-30

800

Shear Strain (%)

-30

30

-800
-60

60

-30

800

400

400

400

30

0
-400
-800
-60

60

F orce (kN)

(i) D10-40

800

F orce (kN)

(h) D05-40

-30

Shear Strain (%)

-30

30

-800
-60

60

800

800

400

400

400

30

Shear Strain (%)

60

0
-400
-800
-60

-30

30

60

(l) D10-50
F orce (kN)

F orce (kN)

(k) D05-50

-30

-30

Shear Strain (%)

800

-800
-60

60

-400

(j) ST50

-400

30

Shear Strain (%)

Shear Strain (%)

800

-800
-60

60

-400

(g) ST40

-400

30

Shear Strain (%)

Shear Strain (%)

(d) ST30

-400

F orce (kN)

-30

Shear Strain (%)

F orce (kN)

F orce (kN)

Shear Strain (%)

F orce (kN)

F orce (kN)

(b) D05-20

800

F orce (kN)

F orce (kN)

(a) ST20

199

30

60

0
-400
-800
-60

Shear Strain (%)

-30

30

60

Shear Strain (%)

Fig. 6. Static and dynamic force-shear strain hysteretic curves.

eq

1 W

4 W

where MDF+ and MDF are the maximum damper force at positive
and negative directions, MDD+ and MDD are the maximum damper
displacement at two opposite directions. W is the energy loose per
cycle represented by the area enclosed inside the hysteresis loop. W
is the energy stored in an elastic spring with a stiffness Kef and MDD.
The equivalent stiffness Kef decreases when LYSPD undergoes
larger displacement in both static and dynamic tests. Furthermore,
the equivalent stiffness Kef is also likely to decrease if the damper
force declines in the dynamic loading. On the other hand, whether
or not in the quasi-static tests with the strain rate closes to 0 or in

the dynamic tests with the strain rate changes from 40%/s to 200%/
s, the equivalent damping ratio eq keeps as a constant value around
0.53. It conrms that the equivalent damping ratio eq of the LYSPD
is strain rate-independent. Compared with other developed metallic
shear panel dampers, the developed LYSPD in this research can provide higher damping ratio which means better seismic performance.
3.2. Low-cycle fatigue'
3.2.1. Shear strain-fatigue cycle curve
The fatigue failure criterion should establish correlation between the
criterion parameter and the number of cycles to fracture or to macrocrack initiation. In previous static constant hysteretic experimental
studies, the number of fatigue cycles to fracture is set as the MDF

200

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

800

(a) Static MDF history

MDF (D05)

MDF (D10)
600

MDF (kN)

Damper force (kN)

600

400

MDF (ST)

400

MDFq (D10)
MDFq (D05)

200
200

ST50

ST40

MDFq (ST)

ST30 ST20
0

0
0

10

20

30

40

Damper forcee (kN)

600

400

200

D05-30

30

40

50

Cycles (N)

Damper force (kN)

800

400

200

D10-30

D10-20

0
0

10

20

30

0:66

0:78

0:81

294N70% DO5

Besides the shear strain amplitude, the fatigue cycles also decrease
with the increase of the shear strain rate in dynamic tests. The relation
between shear strain rate and fatigue cycles is illustrated in Fig. 14. The
nonlinear relation between shear strain rate and fatigue cycles is observed when the shear strain amplitude is 20% while the linear relations
are observed in other three shear strain amplitudes (30%, 40% and 50%).
At the same time, the inuence of the strain rate gets smaller and smaller with the increase of the shear strain amplitude. Although the strain
rates are the smallest ones when the shear strain amplitude is 20%,
the fatigue cycle deteriorations in dynamic tests are the most dramatic
ones. Compared with the static tests, the fatigue cycles of D05-20 and
D10-20 decrease by 33% and 40% respectively which decrease more
than that of other dynamic specimens.

600

D10-40

60

275N70% D10:

(c) D10 dynamic MDF history

D10-50

50

239N70% ST

D05-20

0
20

40

criterion can be unied as the MDF drops to 70% in both static and dynamic constant tests.
The fatigue cycles (N70%) are calculated according to the fatigue criterion mentioned above and plotted in Fig. 13. The symbol represents
the static test results while and represent the dynamic results of
D05 and D10 respectively. The fatigue cycles all decrease with the increase of the shear strain amplitude in all the tests. The corresponding
shear strainfatigue cycle function relations under three different loading frequencies based on test results can be expressed as:

800

10

30

Fig. 8. Strain hardening in the rst cycle.

(b) D05 dynamic MDF history

20

Shear strain (%)

Cycles (N)

D05-50 D05-40

10

50

40

50

Cycles (N)
Fig. 7. MDF history.

Table 2
Test results.
Test type

dropped to 90% in ordinary. It is observed that the damper still has

enough energy dissipation ability when this fatigue failure criterion is
applied to dynamic constant hysteretic tests. Seen from Fig. 7, the dramatic damper force decrease begins at 90% MDF in static tests while it
begins at 70% MDF in dynamic tests. Although the dynamic tests are seldom conducted at present due to the limitation of test equipment and
the lack of similar researches that could be taken as reference, it is believed that the slow dropping of the MDF should not be taken as fatigue
failure criterion in dynamic tests. At the same time, there is no fatigue
cycle difference when the fatigue failure criterion depends on the
force dropping to 90% MDF or 70% MDF. Therefore, the fatigue failure

Static
(ST)

Dynamic
(D05)

Dynamic
(D10)

Specimen

ST20
ST30
ST40
ST50
D05-20
D05-30
D05-40
D05-50
D10-20
D10-30
D10-40
D10-50

First cycle
MDFq (kN)

MDF (kN)

357
342
424
476
352
435
395
488
397
423
477
502

507
533
564
615
521
573
595
616
513
598
623
636

N70% (N)

CPS
(%)

CEA
(kNm)

42
23
15
10.5
28
20
13
9
25
15
11
8

3110
2677
2371
2136
2396
2368
1816
1795
1931
1813
1744
1612

1550
1400
1344
1270
1023
1075
935
968
829
881
910
890

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

800

80

400

60

Force (kN)

temperature increase rate (C/sec)

201

40

-400
20
D05

-800
-100

D10
0

50

100

150

200

-50

50

100

Displacement (mm)
250
Fig. 11. Perfect elasticplastic model.

Shear strain rate (%/sec)

Fig. 9. Temperature increase rate and shear strain rate.

Crack and stress softening resulting from high temperature for the
internal friction are the main two factors that lead to the MDF deterioration and the LYSPD failure. Crack is the dominant factor that
leads to sharp MDF deterioration and the LYSPD failure in static
tests as the temperature increase is negligible. Meanwhile, the stress
softening is another factor that leads to the gradual MDF deterioration and accelerates the LYSPD failure in dynamic tests. The LYSPD
may still fail as crack similar with static failure in dynamic tests before
the stress softening is dominated, because the fatigue cycle is small
when the shear strain amplitude is large. Under this circumstance,
only small fatigue cycle difference will be observed between static
and dynamic tests. It is accordance with the test results when the
shear strain is 50%, only 12 cycles' difference between static test
and dynamic test results is observed. On the opposite, the fatigue
cycles will be affected by stress softening greatly when the shear
strain is small. This is the main reason for the sharp fatigue cycle decrease in dynamic tests when the shear strain amplitude is 20%.
3.2.2. Cumulative plastic shear strain
Many damage accumulation models, proposed in professional literature, can be evaluated by stress, cumulative plastic shear strain (CPS),
cumulative energy absorption (CAE) and local accumulate damage [18]
besides the shear strainfatigue cycle mentioned above. They are formulated on the grounds of experiments or theoretical analysis. As the
stress and local accumulate damage are not easy to be gained in engineer practical use, the CPS and the CAE are attached importance on in

Force decrease rate (kN/sec)

40

this research. Furthermore, the calculation of the CPS and CAE is based
on the aforementioned fatigue failure criterion N70%.
The CPSs of the LYSPD, corresponding to each shear strain under static
and dynamic loading, are plotted in Fig. 15. From top to bottom of Fig. 14,
they are the approximate curves of static, D05 and D10 successively. The
approximate curve of static test is shown as index while the approximate
curves of dynamic tests are demonstrated as linear. It is obvious that CPS
decreases with the increase of dynamic loading frequency. Similar with
the downward trend of fatigue cycles between static and dynamic tests,
CPS has dropped dramatically as the shear strain is 20% while it is relatively small as the shear strain is 50%. Furthermore, CPS also decreases with
the increase of shear strain in both static and dynamic tests. However,
the CPS difference between different shear strains is large (974%) in static
tests while it is narrowed to a small value gap (319%) in dynamic tests
(D10). In accordance with the trend from static tests to D10 dynamic
tests, the correlation between the CPS and shear strain amplitude is gradually becoming smaller and smaller with the increase of dynamic loading
frequency.
In simple terms, regardless of shear strain, only the CPS is focused on
the discussion of the LYSPD's fatigue performance in most studies. From
the perspective of security application, the minimum CPS, when the
shear strain is 50% and loading frequency is 1 Hz (D10-50), is taken as
the lower limit value of the LYSPD which equals 1612%.
3.2.3. Cumulative energy absorption
Similar to Fig. 12, the energy of every cycle can be calculated by the
area enclosed inside the hysteresis loop. The accumulation of the total
energy is the energy dissipation capability of the LYSPD. The CAE of all
the tests is plotted in Fig. 16. The overall distribution and trend of the
CAE in static and dynamic tests are almost the same with that of the
CPS. The difference is that CPS is greatly inuenced by shear strain

30

MDF +
20

Kef
10

MDD-

D05

MDD +

D10
0

50

100

150

200

Shear strain rate (%/sec)

Fig. 10. Damper force decrease rate and shear strain rate.

250

MDFFig. 12. Hysteretic damping denition.

202

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

80

5000
4000

ST
40

D05

20

3000
2000

0
0

10

D10

1000

D10
0

ST

D05
CPS (%)

Shear strain (%)

60

20

30

40

50

20

40

60

Shear strain (%)

Fatigue Cycles N70% (N)

Fig. 15. CPS and shear strain.

Fig. 13. Strain and fatigue cycle relation.

amplitude and the approximate curve shows as an index curve, while

the CAE is slightly inuenced by shear strain and the approximate
curve shows as a linear in static tests. Seen from Fig. 15, both in static
and dynamic tests, the CAE of the LYSPD could also be nearly regarded
as a constant in the same loading frequency regardless of different
shear strains. The minimum CAE (830 kNm) in D10 tests (D10-20) is
taken as the lower limit value of the LYSPD.
3.3. Failure mode
Specimen nal failure modes of static and dynamic tests are shown in
Fig. 17. In spite of the rarely low width/thickness ratio of panel, the
out-of-plane shear buckling is still observed under the static cyclical
loading (Fig. 17(a)(d)). The failure modes of static specimens are almost the same. With the increase of shear buckling, the cracks appear
at the panel corners or the welding seams between panel and ribs. Subsequently, the LYSPD collapses with the expanding of the crack.
Instead of shear buckling in static tests, the in-plane shear deformation is dominant in the dynamic tests. Signicant heat developing
during the dynamic tests on the LYSPD, prior to failure, is also observed within the effective panel deformation area. The maximum
temperatures in dynamic tests are all about 550 C. The horizontal
high temperature red band is mainly caused by internal friction
resulting from in-plane shear while the vertical cracks between the
panel and ribs are inuenced by material deterioration resulting from
welding greatly. Based on these 12 specimens, the failure mode distinction is observed between static and dynamic test results.

4. Conclusion
Four static and eight dynamic specimens have been conducted to
verify the cyclical hysteretic behavior, the damper force history and
the fatigue life of the developed LYSPD. These test results suggest that
the LYSPD subjected to dynamic loading would be greatly affected by
strain rate and the main conclusion could be concluded as follow:
(1) The majority of stress hardening occurs within the rst cycle and
reaches the maximum damper force of each test in both static
and dynamic tests. The strain rate hardening occurs within the
rst quarter cycle of virgin dynamic loading and keeps to the
end of the rst cycle.
(2) After reaching a maximum force, stable damper force is kept
unchanged until failure in static tests and the cyclic damper force
deterioration towards failure with the increasing cycles in dynamic
tests.
(3) A linear relation is setup between damper force deterioration and
strain rate which provides the reference on modeling and designing the LYSPD according to the constant dynamic tests.
(4) The LYSPD can provide the stable, high damping ratio around 0.53
regardless of the strain rate which benets for structure seismic
performance.
(5) The fatigue cycles decrease with the increase of frequency and
shear strain amplitude at the same time. The fatigue cycle is
inuenced by strain rate greatly in dynamic tests when the shear
strain amplitude is small while it can be ignored when the shear
strain amplitude is larger than 50%.
(6) The CPS decreases with the increase of shear strain amplitude in

250

200

150

ST
1500

CAE (kN m)

Strain rate (%/sec)

2000

20%
30%
40%
50%

100

D05
1000

D10

500

50

0
0

10

20

30

40

Fatigue Cycles N70% (N)

Fig. 14. Strain rate and fatigue cycle relation.

50

20

40

Shear strain (%)

Fig. 16. CAE and shear strain.

60

C. Zhang et al. / Journal of Constructional Steel Research 79 (2012) 195203

(a) ST20

(b) ST30

(c) ST40

(d) ST50

(e) D05-20

(f) D05-30

(g) D05-40

(h) D05-50

(i) D10-20

(j) D10-30

(k) D10-40

(l) D10-50

203

Fig. 17. Failure mode.

static tests while it shows weak correlation with shear strain amplitude in dynamic tests. The lower limit value can be conservatively estimated as 1612%.
(7) The CAE shows weak or no correlation with shear strain amplitude in both static and dynamic tests and the lower limit value is
around 830 kNm.
(8) The response of the static specimens is dominated by out-of-plane
deformation while in-plane shear deformation dominates the response in the dynamic specimens.
(9) Whether the damper performance deterioration of dynamic random wave should be taken into consideration or not is depending
on the cycle numbers and shear strain amplitude at the same time.
Attention should be paid to the damper force deterioration when
the shear strain amplitude is small and the cycle number is large
or when the shear strain amplitude is large and even if the cycle
number is only around 3 cycles.
Acknowledgment
The authors acknowledge the supports given by the assistance
coming from the staff of Seismic Research Center at Aichi Institute
of Technology in Japan.
References
[1] Tanaka Kiyoshi, Sasaki Yasuhito, Yoneyama Shin-ichiro. An experimental study on hysteretic performance of shear panel dampers using different strength type of steel under
static loading. J Struct Constr Eng AIJ 1999;520(6):117-24 (In Japanese).
[2] Chan Ricky WK, Albermani Faris, Williams Martin S. Evaluation of yielding shear panel
device for passive energy dissipation. J Constr Steel Res 2009(65):260-8.

[3] De Matteis G, Landolfo R, Mazzolani FM. Seismic response of MR steel frames with
low-yield steel shear panels. Eng Struct 2003;25:155-68.
[4] Hitomi Yasuyoshi, Konomi Hatsunobu Wada Shinzo, Satito Kiichirou, Nakata
Yasuhiro, Iwata Mamoru. Development of the high ductile shear panel. J AIJ
1996;12:118-23.
[5] Shih Ming-hsiang, Sung Wen-pei, GO Cheer-germ. Investigation of newly developed added damping and stiffness device with low yield strength steel. J Zhejiang
Univ Sci 2004;5(3):326-34.
[6] Dusicka Peter, Itani Ahmad M, Buckle Ian G. Cyclic response of plate steels under
large inelastic strains. J Constr Steel Res 2007;63:156-64.
[7] Saeki Eiichiro, Sugisawa Misturu, Yamaguchi Tanemi, Mochizuki Haruo, Wada
Akira. A study on low cycle fatigue characteristics of low yield strength steel. J
Struct Constr Eng AIJ 1995;472:139-47.
[8] Chen Sheng-Jin, Jhang Chyuan. Cyclic behavior of low yield point steel shear walls.
Thin-Walled Struct 2006(44):730-8.
[9] Tanaka Kiyoshi, Sasaki Yasuhito. Hysteretic performance of shear panel dampers
of ultra low yield-strength steel for seismic response control of buildings.
12WCEE 2000; 2000. p. 1-8.1248.
[10] Aoki T, Dang J, Zhang C, Takaku T, Fukumoto Y. Dynamic shear test of low-yield
steel panel dampers for bridge bearing. STESSA 2009; 2009. p. 647-52.
[11] Aoki Tetsuhiko, Zhang Chaofeng, Yuan Huihui. Dynamic loading test of shear
panel damper. EUROSTEEL 2011; 2011. p. 975-80.
[12] Zhang Chaofeng, Zhang Zhisheng, Shi Jinfei. Development of high deformation capacity
low yield strength steel shear panel damper. J Constr Steel Res 2012;75:116-30.
[13] Japan Road Association. Seismic design specications of highway bridges; 1996.
[14] AASHTO. Guide specications for seismic isolation design. Washington (D.C.):
American Association of State Highway and Transportation Ofcials; 1999.
[15] Eurocode 8. Design provisions for earthquake resistance of structures, Part2: bridges. ENV 1998-2. Brussels; 1994.
[16] Hwang JS, Sheng LH. Equivalent elastic seismic analysis of base isolated bridges
with lead-rubber bearings. Eng Struct 1994;16:201-9.
[17] Jangid RS. Equivalent linear stochastic seismic response of isolated bridges. J
Sound Vib 2008;309:805-22.
[18] Borodii MV, Adamchuk MP. Life assessment for metallic materials with the use of
the strain criterion for low cycle fatigue. Int J Fatigue 2009;31:1579-87.