You are on page 1of 8

Engineering Structures 46 (2013) 294301

Contents lists available at SciVerse ScienceDirect

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

Low-cycle fatigue testing of extruded aluminium alloy buckling-restrained braces


Chun-Lin Wang a,b, Tsutomu Usami b,, Jyunki Funayama b, Fumiaki Imase b
a
b

International Institute for Urban Systems Engineering, Southeast University, Sipailou 2#, Nanjing 210096, China
Department of Civil Engineering, Meijo University, Tempaku-ku, Nagoya 468-8502, Japan

a r t i c l e

i n f o

Article history:
Received 9 July 2011
Revised 5 February 2012
Accepted 17 July 2012
Available online 13 September 2012
Keywords:
Seismic design
Buckling-restrained brace
Aluminium alloy
Low-cycle fatigue
Extrusion manufacture
Fatigue curve

a b s t r a c t
Aluminium alloys have recently been employed to manufacture buckling-restrained braces (BRBs) with
the aim of improving BRB durability in corrosive environments. Based on the ease with which aluminium
alloys are extruded, the extruded aluminium alloy BRB is proposed to avoid the welded and relatively
complex BRB end used in previous BRB research. This experiment, including 10 nearly identical specimens with or without stoppers, was performed to address low-cycle fatigue performance. According to
the test results, the extruded BRB possessed a stable and repeated hysteretic performance, and the fracture location was random in the yielding portion of the brace. The failure of the extruded BRB was
regarded as a brittle fracture compared to the typical failure of a steel BRB. The comparison between
specimens with and without stoppers showed that the stoppers had no clear inuence on the cumulative
inelastic deformation, provided that the BRB was horizontally placed and the strain amplitude was lower
than 2%. The low-cycle fatigue damage evaluation formula for the extruded BRB is recommended as a reference for strain-based damage assessment.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
A rational approach for the seismic design or retrot of structures is based on the introduction of energy dissipation devices,
which act as ductile fuses during an earthquake to minimise the
damage of both structural and non-structural elements. A
buckling-restrained brace (BRB), which is a type of metallic yielding-based device, can yield under both tension and compression
without buckling and exhibits a stable elasticplastic hysteretic
behaviour. Previous research and engineering applications show
that BRBs improve the seismic performance of conventional buildings and bridges when they replace diagonal braces.
Experimental and numerical studies at the component and
frame levels have been conducted to promote the application of
different types of BRBs. For example, Chou and Chen [1] addressed
experimental and numerical research on a BRB that employed two
restraining systems to restrain the buckling of a core plate. Sabelli
et al. [2] investigated the seismic responses of buckling-restrained
braced frames (BRBFs) with several important parameters. Kim
et al. [3,4] and Wu et al. [5] provided design procedures to quantify
the responses of BRBFs. Usami et al. [6] and Chen et al. [7] numerically studied the effect of BRBs on the seismic performance of a
retrotted steel bridge. In particular, Mazzolani [8] tested three
sub-structures of a real reinforced concrete building that was
upgraded with BRBs and other energy-dissipation techniques to
Corresponding author. Tel./fax: +81 52 838 2363.
E-mail addresses: usamit@meijo-u.ac.jp (T. Usami), chunlin@seu.edu.cn (C.-L.
Wang).
0141-0296/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.engstruct.2012.07.016

calibrate numerical models. The research to date has accelerated


the application of BRBs in civil engineering.
Recently, there has been a focus on the use of aluminium to improve the durability of metallic energy-dissipation devices in corrosive environments. Aluminium oxide is naturally generated on
the surface of a metal and is more stable than pure aluminium.
Thus, it can act as a natural lm to protect the metals body against
corrosion, which is very important for the successful application of
BRBs in saltwater environments [9].
Besides its excellent corrosion resistance, aluminium alloy offers a wide range of useful properties: it is light and has only
one-third the density of steel, it is easily fabricated, and its recyclability provides both economic and environmental benets. For
these reasons, aluminium has been employed to develop shear
panels or links as energy-dissipation dampers in some studies
[10,11]. Nevertheless, contact between aluminium and steel
should be avoided. Steel bolts should be lmed by uorine resin
during the assembly of aluminium alloy BRBs. In this experimental
study, normal high-strength steel bolts were used for convenience.
An economic analysis conrmed that the major costs of this assembly were related to the aluminium alloy, but the total cost, including life cycle costs, was deemed acceptable.

2. Previous research
Although the low-cycle fatigue performance of aluminium
alloys has been widely veried in the literature [1216], little
attention has been paid to the use of aluminium and its alloys in

295

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

BM
Angle

Weld toe

(b)

(a)

Fig. 1. Congurations of the (a) welded and (b) bolt-assembled aluminium alloy BRBs [17].

Die
Semi-finished

Stopper
Rib

(a)

(b)

(d)

(c)

Fig. 2. Extrusion process of the aluminium alloy BM: (a) step I; (b) step II; (c) step III; (d) step IV.

(Unit: mm)

87.5

87.5

L=1360

240

190

190

240

190

12

30

190
180

45

1565
2015

45

180

(b)

(a)

Stopper

Rib

(c)
Fig. 3. Extruded BM (a) cruciform section; (b) brace member; (c) photo.

the manufacture of BRBs. As a result, there are few detailed test


data on aluminium alloy BRBs. In the past 2 years, one group of
authors [17]has conducted several experiments on welded and
bolt-assembled aluminium alloy BRBs, whose congurations are
shown in Fig. 1.
According to the test results, the low-cycle fatigue life of the
welded BRB was remarkably inuenced by the welding of the ribs,
which were used to improve the out-of-plane stiffness of the unrestrained region of the brace member (BM). Consequently, the boltassembled aluminium alloy BRB was proposed to avoid the rib
welds. Further experiments showed that the performance of the
bolt-assembled BRB with spot-welded stoppers was affected by
the spot-weld of the stoppers, which were used to prevent the
slip-off movement of the restraining members (RMs). Therefore,
a bolt-assembled aluminium alloy BRB without stoppers was

subsequently tested, and the corresponding results showed that


it exhibited a stable hysteretic behaviour throughout the duration
of the test. The performance of the bolt-assembled aluminium
alloy BRB without stoppers partially satises the demands proposed for steel high-performance BRBs (HPBRBs) used in bridge
engineering, as suggested by authors [7,18]. The HPBRB is expected
to survive after three times of strong earthquakes without severe
damage, and it does not need to be replaced during the typical life
cycle of a bridge.
Although a bolt-assembled BRB has good hysteretic performance, the conguration of this BRBs end is relatively complex,
as shown in Fig. 1b. Considering that an aluminium alloy is easily
extruded, the extrusion process is employed to fabricate BMs, as
shown in Fig. 2. A semi-nished product can be obtained through
the use of a die on the nished cruciform section, and this

296

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

Table 1
Geometric dimensions and structural properties of the BMs.
Series

Specimens

L (mm)

B (mm)

t (mm)

A (mm2)

P0 (kN)

d0 (mm)

EA-WS

EA-WS-R1
EA-WS-1.0
EA-WS-1.5
EA-WS-2.0
EA-WS-2.5
EA-WS-D1
EA-WS-D2

1360

99.9
99.9
100.1
100.0
99.8
100.1
100.0

10.0
10.1
10.1
10.2
10.2
10.2
10.1

999
1009
1011
1020
1018
1021
1010

471
465
466
462
461
462
466

185.4
187.3
187.6
189.3
188.9
189.5
187.5

3.94
3.94
3.94
3.94
3.94
3.94
3.94

EA-NS

EA-NS-R1
EA-NS-1.5
EA-NS-2.0

100.1
100.1
100.2

10.0
10.1
10.1

1001
1011
1012

471
466
466

185.8
187.6
187.8

3.94
3.94
3.94

Note: L is the nominal length of the BMs yield portion; B is the width; t is the thickness; A is the sectional area; k is the slenderness ratio on the weak axis; P0 is the Ar0; and d0
is the Le0. The symbols L, B, and t are illustrated in Figs. 3b and 4.

Unbonding material

Bolt

Table 2
Chemical compositions of aluminium alloys.

(Unit: mm)

Type

d=1

d0=2

25
HS63S-T5
A6061S-T6

Chemical compositions (%)


Si

Fe

Cu

Mn

Mg

Cr

Zn

Ti

0.57
0.08

0.21
0.26

0.01
0.01

0.02
0.02

0.76
2.62

0.01
0.18

0.01
0.02

0.02

25

48

Restraining member
(RM)

Brace member
(BM)

48

200
Fig. 4. Details of the cross-section.

procedure is used to manufacture the BM. It is important that a BM


with ribs and stoppers has no welds, although the extrusion process has a small negative effect on fatigue performance, as discussed in the literature [19].
Based on previous research of extruded aluminium alloy BRBs,
an experiment involving 10 specimens was conducted at the Advanced Research Centre for Seismic Experiments and Computations (ARCSEC) of Meijo University to address their low-cycle
fatigue performance and to evaluate the effect of stoppers. Details
of the test program and results are summarised below.

3. Test program
3.1. Dimensions of specimens
Fig. 3a and b give the nominal dimensions of the dies cruciform
section and the nished BM. As shown in Fig. 3c, there is no weld in
the BM, but it has stoppers at its centre and ribs in its unrestrained
region. Table 1 lists the measured dimensions and structural properties of the BM. As shown in Fig. 4, the BM was inserted between a
pair of RMs, and the unbonding material, a type of butyl rubber
with a thickness of 1 mm, was used to minimise friction. Gaps with
widths d = 1 mm and d0 = 2 mm were provided between the BM
and the RMs.

Aluminium alloy at plates were selected for manufacturing the


specimens RMs by considering the advantages and disadvantages,
such as lightness and contact corrosion with other materials, of
various aluminium alloys. The nominal dimensions of the RMs
are shown in Fig. 5. The RM dimensions employed in the literature
[17] were used again in this experiment, and the overall buckling
of the specimen and the relative displacement between the two
RMs were not observed in any of the tests. The frictional force between the two RMs was relied upon to prevent relative displacement because each high-strength bolt was tightened by an
electric wrench with a torque of 170 Nm.
3.2. Adopted material characteristics
A new aluminium alloy, HS63S-T5, provided by Nippon Light
Metal Company, LTD., was used for the BM. This material was developed to increase the strength of the A6063 aluminium alloy. An
A6061S-T6 aluminium alloy was employed to fabricate the RMs.
Table 2 lists the chemical compositions of two materials, and Table
3 lists the data of the coupon tests performed on the materials.
Fig. 6 presents the stressstrain curves of the HS63S-T5 and
A6061S-T6 alloys obtained from coupon tests, which show that
the A6061S-T6 has a notably higher yield stress than the HS63S-T5.
3.3. Labelling
Table 1 lists the tests specimens. The stoppers of the EA-NS-R1,
EA-NS-1.5 and EA-NS-2.0 specimens were ground to evaluate their
effects on the BRBs performance. Each specimen was labelled so
that the type and the testing protocol could be clearly identied.

95

48

75

200

200

75

48

(Unit: mm)

25

50

28@50
1500
Fig. 5. Dimensions of the RMs.

50

297

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301


Table 3
Material properties of aluminium alloys.
Alloy Type

E (GPa)

r0.2 (MPa)

r0 (MPa)

e0.2 (%)

e0 (%)

ru (MPa)

eu (%)

Objective

HS63S-T5
A6061S-T6

64.9
72.1

206.3
273.8

185.6
246.5

0.52
0.58

0.28
0.34

230.3
300.9

8.02
7.82

0.35
0.33

For BM
For RMs

Note: E is the initial Youngs modulus; r0.2 is the 0.2% proof stress; r0 is the 0.9 r0.2; e0.2 is the 0.2% proof strain; e0 is the deformation corresponding to the stress r0; ru is the
ultimate tensile strength; eu is the tensile strain at fracture; and m is the Poisson ratio.

3.4. Testing setup


350

As shown in Fig. 7, an extruded specimen was horizontally pinned between two rigid pillars. The force was applied by two parallel jacks. The axial displacement of the restrained yield portion was
monitored by eight displacement transducers, which were
mounted on both ends of the specimen.

300

A6061S-T6

Stress (MPa)

250

300

200

HS63S-T5

250

150

3.5. Testing protocol

200
150

100

The experiment employed several tensile and compressive reversed cyclic testing patterns controlled by the axial strain of the
specimens. As shown in Fig. 8a, the rst testing pattern (R1) was
a stepwise incremental cyclic protocol adopted in the EA-WS-R1
and EA-NS-R1 specimen tests. The strain amplitude of each cycle
increased from e0 to 10 e0 (approximately 2.8%) with the increment
e0 and then maintained this amplitude until failure.
As shown in Fig. 8b, the second testing pattern included two
stages. The rst stage was composed of one 0.5 e0 strain amplitude
cycle and one e0 strain amplitude cycle, which were used to evaluate the equipment system. In the second stage, the constant strain
amplitude specied in Table 4 was imposed cyclically until the
specimen failed.
The third testing pattern (D1 or D2) was used to directly evaluate the damage index based on Miners Law. As shown in Fig. 8c,
two cycles of the e0 strain amplitude were rst imposed as an evaluation procedure. In the testing of specimen EA-WS-D2, the imposed strain amplitude was 1% (4 cycles), 2% (4 cycles) and 2.5%
(employed until specimen failure), where as in the testing of the
EA-WS-D1, the imposed strain amplitude was 1% (8 cycles), 1.5%
(2 cycles), 2% (1 cycle) and 2.5% (employed until specimen failure).
Because the engineering strain, e, is dened as the relative displacement divided by the original length of the BMs yield portion,
the strain control becomes equivalent to the displacement control
during the testing. Thus, these tests are conducted by controlling
the axial displacement.

100
50

50

0
0.0 0.1 0.2 0.3 0.4 0.5

Strain (%)
Fig. 6. Stressstrain curves from the coupon tests.

Jacks
Specimen

Displacement

Data acquisition system

transducers

Fig. 7. Testing setup.

The rst part of the label indicates the type of the BRB, where EA
refers to the extruded aluminium alloy BRB; the middle part indicates whether the BRB contains stoppers, where WS and NS refer to specimens with and without stoppers, respectively; and the
last part indicates the strain amplitude of the testing pattern,
where 1.0 refers to a constant 1% strain amplitude and R1,
D1, andD2 indicate the variable strain amplitude, which is explained in the following section.

/2

40
30
20
0

4. Test results
4.1. Stressstrain relationships
Fig. 9 provides stressstrain curves for the extruded specimens.
The tensile states of the BRBs are displayed in the positive

n
2.5%
2%
1%
0

0
0.5 0
0

1%
2%
2.5%

/2

(a)

0
0

0.5 0

0
2 0
3 0
4 0

n2
n1

(b)

(c)

Fig. 8. Testing patterns: (a) stepwise incremental strain amplitude (R1); (b) constant strain amplitude; (c) variable strain amplitude (D1 or D2).

298

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

Table 4
Test results of the specimens.
Series

Specimen

De/2 (%)

De (%)

Nf

ni

CID (%)

Failure position

EA-WS

EA-WS-R1
EA-WS-1.0
EA-WS-1.5
EA-WS-2.0
EA-WS-2.5
EA-WS-D1

1.0
1.5
2.0
2.5
1.0
1.5
2.0
2.5
1.0
2.0
2.5

2.0
3.0
4.0
5.0
2.0
3.0
4.0
5.0
2.0
4.0
5.0

40
10
5
2

8
2
1
2
4
4
1

31
93
42
30
17
49

1.14
1.40
1.22
1.41
1.01
1.88

Middle
Left
Middle
Right
Middle
Middle

41

1.72

Right

1.5
2.0

3.0
4.0

9
6

34
37
36

1.45
1.09
1.70

Middle
Middle
Middle

EA-WS-D2

EA-NS

EA-NS-R1
EA-NS-1.5
EA-NS-2.0

Note: De/2 is the strain amplitude; De is the strain range; Nf is the number of failure cycles; ni is the occurrence frequency according to Dei range; CID is the cumulative
inelastic deformation; and D is the damage index.

direction. The specimens demonstrated stable and repeated hysteretic curves at all of the testing amplitudes, even though the
maximum strain amplitude was 2.5%.
The hysteretic behaviour of some extruded specimens tested
with comparatively large strain amplitudes, such as EA-WS-2.0

300
EA-WS-1.0

(MPa)

0
-100
-200

Nf=40

-300

200

200

100

100

100

0
-100

Nf=10

-300
3

-100

-3

-2

-1

(%)

Nf=5

-300
3

-3

-2

-1

(%)

200

100

100

100

(MPa)

200

200

-100

0
-100
-200

-200

-3

-2

-1

-2

-1

(%)

-100

100

100

100

(MPa)

200

0
-100
-200

-1

(%)

-3

0
-100
-200

Nf=9

-300

EA-NS-R1
-2

EA-NS-2.0

200

-3

-1

300

200

-300

-2

(%)

EA-NS-1.5

-200

-1

-3

300

-2

EA-WS-D2

(%)

-100

-3

-300

-3

300

Nf=2

-300
3

-200

-300

-300

-2

-1

Nf=6

-300
3

(%)
Fig. 9. Stressstrain relations of the specimens.

(%)

300

EA-WS-D1

EA-WS-R1

-100

(%)

300

300

-200

(MPa)

-1

-200

(MPa)

-2

(MPa)

-3

EA-WS-2.5

200

-200

(MPa)

(MPa)

Failure
point

300
EA-WS-2.0

(MPa)

200
100

300
EA-WS-1.5

(MPa)

300

and EA-WS-2.5, is slightly asymmetric in both tension and compression. For example, the maximum compressive stress of the
EA-WS-2.0 specimen was 4.5% greater than its maximum tensile
stress. In comparison, under the same strain amplitude conditions,
the maximum compressive stress of a steel BRB with an identical

-3

-2

-1

(%)

299

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

between the EA-WS-2.0 and EA-NS-2.0 specimens under the same


2% amplitude. Therefore, it was concluded that the stoppers had
no remarkable inuence on the low-cycle fatigue performance of
the extruded BRB when the BRB was horizontally placed and the
imposed strain amplitude was smaller than 2.0%.

conguration is 5.7% greater than its maximum tensile stress [20].


One reason for the similarity in asymmetry between the steel and
aluminium alloy BRBs can be explained as follows. The shear friction occurred between the BM and the RMs when the BM sustained
the compressive force. This friction was correlated with the friction
coefcient and the compressive force. Additionally, because the
same unbonding material was used, this similarity was observed
in BRB experiments that were identical in every way except for
the choice of material. The rst loop in the BRBs hysteretic behaviour was hardly affected, and the subsequent loops were inuenced
by the strain hardening effect. These results are identical to the test
results of steel BRBs [20] and A5083P-O aluminium alloy BRBs [17].

4.3. Failure modes


Fig. 10 presents the failure mode of the extruded BRBs, and
Fig. 11 presents the failure modes of the bolt-assembled aluminium alloy BRBs and the steel BRBs tested with 2% strain amplitude
reported in the relevant literature [17,20]. The rupture of the extruded BM induced the specimens failure, the locations of which
were random. Based on the comparison between specimens with
and without stoppers, it appears that the stoppers did not signicantly affect fracture. The fracture surfaces of the extruded specimens were much rougher than those in the blot-assembled
aluminium alloy BRBs. One possible reason for this difference
could be that the defects of the extrusion process, which can be
considered as an existent structural imperfection, had an inuence
on the failure mode. Thus, the failure of the extruded BRBs was regarded as a brittle fracture and no apparent necking plastic deformation occurred before fracture.
Fig. 12 shows the nal loops of the corresponding stressstrain
relationships. During the testing of the aluminium alloy BRBs, the
actuators force dropped rapidly to zero at the instant of each
BMs rupture when the imposed velocity measured 0.15 mm/s.
However, the failure of the steel BRBs in previous studies was considered to be a ductile fracture because a declination process and a

4.2. Fatigue performance


The cumulative inelastic deformation (CID) indicates the cumulative ductility deformation capacity excluding the elastic portion
before the failure of specimens, which is also listed in Table 4.
The CID values are greater than 30% when the imposed constant
strain amplitude is lower than 2% or the imposed amplitude is
variable.
The specimens with and without stoppers were tested to evaluate the effect of the stoppers on the extruded BRBs performance.
During the test of the EA-WS-R1 and EA-NS-R1 specimens, the
slip-off movement between the BM and the RMs was not clearly observed, and their CID values were nearly equal. Moreover, the number of fatigue cycles Nf dropped from 10 to 9 during a comparison
between the EA-WS-1.5 and EA-NS-1.5 specimens under the same
1.5% amplitude, while Nf increased from 5 to 6 during a comparison

EA-WS-R1

EA-WS-1.0

EA-WS-1.5

EA-WS-2.0

EA-WS-2.5

EA-WS-D1

EA-WS-D2

EA-NS-R1

EA-NS-1.5

EA-NS-2.0

EA-WS-1.5 EA-WS-2.5

EA-WS-1.0
Fixed End

Left

EA-WS-D1

EA-WS-R1
Middle

EA-WS-D2

EA-WS-2.0
Right

EA-NS-2.0

EA-NS-1.5 EA-NS-R1
Fig. 10. Failure modes of the extruded specimens.

Loading End

300

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

De C  Nf k

Based on test results of the extruded BRBs with stoppers under constant strain amplitude (from EA-WS-1.0 to EA-WS-2.5, as listed in
Table 4), the values of C and k are obtained by the least mean square
method. The MansonCofn equation for the extruded BRB can be
expressed as:

Crack

(a)

Fracture

De 0:063  Nf 0:306

(b)

As shown in Fig. 13a, some specimens fall below the fatigue curve
(SN curve) of Eq. (3), which suggests that the extruded BRB is unsafe for structural applications. Combined with the standard error
obtained by the least mean square method, the recommended MansonCofn equation for the extruded BRBs is updated to obtain:

Fig. 11. Failure modes of (a) a bolt-assembled aluminium alloy BRB [17] and (b) a
steel BRB [20].

clear inelastic deformation were observed on the BM, as shown in


Figs. 12c and 11b.

De 0:060  Nf 0:329

To present the test results of the specimens under the variable


strain amplitude with SN curves in Fig. 13a, the equivalent strain
range Deeq and the equivalent number of fatigue cycles N can be
expressed as [22]:

5. Low-cycle fatigue life


5.1. Fatigue curves
The MansonCofn equation is often used to indicate the relationship between the number of failure cycles Nf and a strain range
and can be given as [21]:

De C e  Nf ke C p  Nf kp

Deeq

where De is the total strain range; Nf is the number of failure cycles;


and Ce, Cp, ke and kp are constants that depend on the material. Because the plastic strain of the BM is much larger than its elastic
strain and the total strain range is directly measured from the
experiment, Eq. (1) can be approximately given by:

500

!k
P
 ni
De1=k
i
N

Fig. 13a shows that the specimens tested with the variable strain
amplitude lie over the recommended SN curve according to Eq.
(4). Fig. 13b compares the fatigue curves of the different BRBs that
have been recently studied [17,20]. The low-cycle fatigue performance of the extruded aluminium alloy BRB is better than that of

500

400

Failure point

400

Failure point

300

200

200

200

100

100

100

-100

(MPa)

300

(MPa)

300

0
-100

-200

-200

-300

-300

-300

-400

-400

-400

-500

-500

-2

-1

0
1
(%)

-3

-2

(a)

-1

0
1
(%)

Failure point

-100

-200

-3

ni

500

400

(MPa)

-500

-3

-2

-1

0
1
(%)

(c)

(b)

Fig. 12. Final loops of the stressstrain relationships of (a) the EA-WS-2.0 specimen, (b) the bolt-assembled aluminium alloy BRB [17] and (c) the steel BRB [20].

0.07

0.08
0.07
0.06
0.05

0.06
: EA-WS-1.0~2.5
: Equation (3)
: Equation (4)

0.04
0.03

: EA-WS-D1
: EA-WS-D2
: EA-NS-1.5
: EA-NS-2.0

0.02

Strain range

Strain range

0.05

Steel BRB
Bolt-assembled BRB

0.04

Extruded BRB
Welded BRB

0.03
0.02
Aluminum alloy
: Extruded BRB
: Welded BRB
: Bolt-assembled BRB
: Steel BRB

0.01
0.01

10

Number of cycles (Nf )

100

10

Number of cycles (Nf )

Fig. 13. (a) SN curves for the extruded BRB; (b) comparison of SN curves.

100

C.-L. Wang et al. / Engineering Structures 46 (2013) 294301

the welded aluminium alloy BRB, but it is lower than that of the
bolt-assembled aluminium alloy BRB or the steel BRB.
5.2. Miners law
The damage caused by one cycle at the Dei strain range is dened as Di = 1/N, where N is the number of failure cycles for
the Dei strain range [9]. Thus, the cumulative damage throughout
the straintime history can be expressed as:

X ni
Nfi

where ni is the occurrence frequency at the Dei strain range, and D is


the cumulative damage index. The relationship expresses the linear
damage rule proposed by Miner for the prediction of aircraft fatigue
life [21]. When the cumulative damage reaches 1.0, the structure
suffers fatigue failure. Based on test results of the specimens with
a constant strain amplitude, the D values of the EA-WS-D1 and
EA-WS-D2 can be given directly as 1.6 and 1.4, respectively. By
combining Eqs. (2) and (7), the relationship between the cumulative
damage index and the strain range can be given as:

D C 1=k 

ni  Dei 1=k C 

ni  Dei m

Thus, the cumulative damage formula for the extruded aluminium


alloy BRB is given as follows:

D 5:1  103 

ni  Dei 3:04

The test results were used to quantify the validity of Eq. (9). The
cumulative damage index is given in Table 4 and indicates that this
evaluation formula for extruded BRBs is conservative and effective.
6. Conclusions
This paper further investigated the potential for the use of easily
extruded aluminium alloy BRBs in structural applications. Tests of
extruded aluminium alloy BRBs with and without stoppers were
performed to understand their low-cycle fatigue performance.
The main results are summarised as follows:
(1) Low-cycle fatigue tests show that extruded BRBs exhibit stable and repeated hysteretic performance and that the CID
value is larger than 30% when the imposed constant strain
amplitude is lower than 2% or when the specimen is under
variable strain amplitude.
(2) The fracture location of extruded BRBs is random in the yield
portion of the BM. Its failure is regarded as a brittle fracture
in comparison to the failure of steel BRBs.
(3) A comparison between the specimens with and without
stoppers shows that the stoppers have no clear inuence
on low-cycle fatigue performance when the BRB is horizontally placed and the imposed strain amplitude is lower than
2%.
(4) The low-cycle fatigue damage evaluation formula for
extruded aluminium alloy BRBs is recommended as a reference for strain-based damage assessment.

301

Acknowledgement
The study is supported by JSPS Grant-in-Aid for Scientic Research (B) 23360200 and in part by grants from Japan Science
and Technology Agency for Evaluation and Mitigation of Environment Impacts of Earthquake and Typhoon Disaster on Urban Area
and Infrastructures (Project Title: Rened Analysis and Damage
Control of Earthquake Disaster Impact on Bridge Structures), under
the Strategic Japanese-Chinese Cooperative Program on Science
and Technology (S&T) for Environmental Conservation and Construction of a Society with Less Environmental Burden.
References
[1] Chou C-C, Chen S-Y. Subassemblage tests and nite element analyses of
sandwiched buckling-restrained braces. Eng Struct 2010;32:210821.
[2] Sabelli R, Mahin S, Chang C. Seismic demands on steel braced frame buildings
with buckling-restrained braces. Eng Struct 2003;25:65566.
[3] Kim J, Choi H. Behavior and design of structures with buckling-restrained
braces. Eng Struct 2004;26:693706.
[4] Kim J, Seo Y. Seismic design of low-rise steel frames with buckling-restrained
braces. Eng Struct 2004;26:54351.
[5] Wu J, Liang R, Wang C-L, Zhou Z. A modied capacity spectrum method with
direct calculation of seismic intensity of points on capacity curve. J Earthq Eng
2011;15:66483.
[6] Usami T, Lu Z, Ge H. A seismic upgrading method for steel arch bridges using
buckling-restrained braces. Earthq Eng Struct Dyn 2005;34:47196.
[7] Chen X, Ge H, Usami T. Seismic demand of buckling-restrained braces installed
in steel arch bridges under repeated earthquakes. J Earthq Tsunami
2011;5:11950.
[8] Mazzolani FM. Innovative metal systems for seismic upgrading of RC
structures. J Constr Steel Res 2008;64:88295.
[9] Mazzolani FM. Aluminium alloy structures. 2nd ed. London, UK: E & FN Spon;
1995.
[10] De Matteis G, Mazzolani FM, Panico S. Experimental tests on pure aluminium
shear panels with welded stiffeners. Eng Struct 2008;30:173444.
[11] Sahoo DR, Rai DC. Seismic strengthening of non-ductile reinforced concrete
frames using aluminum shear links as energy-dissipation devices. Eng Struct
2010;32:354857.
[12] Cofn Jr LF. Low cycle fatigue: a review. Appl Mater Res 1962;1:12941.
[13] Brown GW, Ikegami R. The fatigue of aluminum alloys subjected to random
loading. Exp Mech 1970;10:3217.
[14] Chung YS, Abel A. Low cycle fatigue of some aluminum alloys. In: Solomon HD,
Halford GR, Kaisand LR, Leis BN, editors. Low cycle fatigue. ASTM STP 942,
Philadelphia: American Society for Testing and Materials; 1988. p. 94106.
[15] Borrego L, Abreu L, Costa J, Ferreira J. Analysis of low cycle fatigue in AlMgSi
aluminium alloys. Eng Fail Anal 2004;11:71525.
[16] Xue L. A unied expression for low cycle fatigue and extremely low cycle
fatigue and its implication for monotonic loading. Int J Fatigue
2008;30:16918.
[17] Usami T, Wang C-L, Funayama J. Developing high-performance aluminum
alloy buckling-restrained braces based on series of low-cycle fatigue tests.
Earthq Eng Struct Dyn 2012;41:64361.
[18] Usami T. Developing high-performance damage control seismic dampers. In:
10th symposium on ductile design method for bridges (Special Lecture): JSCE;
2007. p. 1122 [in Japanese].
[19] Kaufman JG. Properties of aluminum alloys: fatigue data and the effects of
temperature, product form, and processing: ASM Intl 2008.
[20] Usami T, Wang C-L, Funayama J. Improving low-cycle fatigue performance of
high-performance buckling-restrained braces by toe-nished method. J Earthq
Eng. http://dx.doi.org/10.1080/13632469.2012.703385.
[21] Stephens RI, Ali F, Stephens RR, Fuchs HO. Metal fatigue in engineering. 2nd
ed. New York: John Wiley & Sons, Inc.; 2001.
[22] Tateishi K, Hanji T, Minami K. A prediction model for extremely low cycle
fatigue strength of structural steel. Int J Fatigue 2007;29:88796.

You might also like