Anda di halaman 1dari 13

Engineering Structures 163 (2018) 38–50

Contents lists available at ScienceDirect

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

Cyclic experimental and analytical studies of buckling-restrained braces T


with various gusset connections

Jingfeng Wanga,b, , Beibei Lia, Chungche Chouc, Liang Chend
a
School of Civil Engineering, Hefei University of Technology, Anhui Province 230009, China
b
Anhui Civil Engineering Structures and Materials Laboratory, Anhui Province 230009, China
c
Department of Civil Engineering, Taiwan University, Taipei, China
d
Shanghai LANKE Building Damping Technology Co., Ltd., Shanghai 200082, China

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

Keywords: Five buckling-restrained braces (BRBs) with various gusset connections, including pinned gusset connection,
Buckling-restrained braces (BRBs) bolted gusset connection, welded gusset connection, pinned-welded gusset connection as well as pinned-bolted
Gusset plate gusset connection, were designed and tested under axial cyclic loading. Typical failure modes for all specimens
Cyclic behavior were summarized and analyzed. The experimental results showed that all specimens exhibited stable hysteretic
Equivalent stiffness
performance, good ductility and cumulative plastic deformation capacity. In addition, the relationship between
Finite element (FE) analysis
the core plate strain demand and the effective factor was adopted to assess whether a BRB deformation satisfied
the requirement of at least two times the design story drift. On the basis of Whitmore section and effective
length, a formula of equivalent stiffness for a BRB with various gusset plates was suggested and verified by the
experimental and numerical results. Furthermore, the finite element (FE) models of BRBs with various gusset
connections were established and good consistency between the numerical and test results indicated that the FE
modeling could simulate the cyclic behavior of BRB well. The elastic stiffness of gusset plate should be con-
sidered in the BRB design, and various gusset connection type and corresponding configurations provided an
alternative method for designers to adjust conveniently the equivalent stiffness of brace under the premise of
meeting the core plate strain demand.

1. Introduction frame (BRBF) and limited the BRB to fully yield to dissipate energy. Tsai
and Hsiao [11] tested a series of full-scale three-story three-bay BRBFs
Buckling-restrained brace (BRB) generally consists of a core element with bolted gusset plates under pseudo-dynamic simulation. The stif-
and a restraining element. The configuration of unbonding layer or air feners were added when out-of-plane deformation of the first-story
gap is used to eliminate force transfer and accommodate transverse gusset plates was observed during Phase 1. Three specimens of BRBFs
expansion in compression. The existence of the restraining element with welded gusset plates were conducted under cyclic loading by
ensures that the core plate can fully yield in both tension and com- López et al. [12]. The free edge of the gusset plate buckled and the
pression, providing significantly better energy dissipation capacity and gusset-to-column welds cracked during test. In order to prevent the
ductility than the conventional steel brace. Various types of BRBs were failure of gusset plates, the stocky pinned gusset plate was used in a 3/
developed and the experimental and analytical results revealed that 5-scale four-story BRBF and was implemented by Fahnestock et al. [13]
they exhibited stable hysteretic performance, high ductility and cu- using a hybrid pseudo-dynamic testing method. Moreover, Wigle et al.
mulative plastic deformation capacity [1–10]. These favorable attri- [14] studied the effect of various connection configurations, including
butes have prompted extensive application of BRBs in new and existing bolted gusset plate, welded gusset plate and pinned gusset plate, on the
frame structures in regions of high seismicity to resist lateral loads and global response based on FE analysis. Results showed that the con-
mitigate frame structure damage. nection configuration influenced local connection demands and global
The gusset plate plays a significant connection role in transferring performance.
axial force from the BRB to the surrounding frame structure. However, In order to understand the mechanical behavior of gusset plate and
undesirable buckling or failure of the bolted or welded gusset plates prevent premature buckling or failure, many scholars, such as Kaneko
occurred on some large scale tests of the buckling-restrained braced et al. [15], Tsai et al. [16], Chou et al. [17–19] and Khoo et al. [20],


Corresponding author at: School of Civil Engineering, Hefei University of Technology, Anhui Province 230009, China.
E-mail address: jfwang008@163.com (J. Wang).

https://doi.org/10.1016/j.engstruct.2018.02.008
Received 23 October 2017; Received in revised form 16 December 2017; Accepted 1 February 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
J. Wang et al. Engineering Structures 163 (2018) 38–50

focused on the opening and closing effect of the beam-column-gusset configuration.


connection due to frame action; useful guidelines were proposed to The BRB studied in this paper was composed of a core element and a
design gusset plates to consider the frame action and axial force of the restraining element. A flat core element with a cross-section of
BRB. 25 × 41 mm was used for all specimens to provide the same force de-
The above summaries showed that many scholars focused on the mand to surrounding frame members. In general, the deformation ca-
mechanical behavior of BRBs and buckling or failure of gusset plates, pacity of a BRB was affected by the length of yielding portion. The
whereas few work studied the equivalent stiffness consisting of BRB objective of this paper was to examine the deformation capacity of a
stiffness, gusset plate stiffness and beam-to-column panel zone stiffness. BRB with different yielding length, such as 1.70 × 103 mm,
Chuang et al. [21] introduced an effective stiffness factor to consider 7.60 × 102 mm, 1.30 × 103 mm, 1.50 × 103 mm and 1.20 × 103 mm
the gusset plate stiffness and panel zone stiffness. Judd et al. [22] for five specimens and illustrated in Table 1 and Fig. 1. It was expected
provided a formula to compute the equivalent stiffness of the BRB with that the deformation capacity of specimen BRB-GP2 with the shortest
gusset plates, but it didn’t present the process of calculating the gusset yielding length of 7.60 × 102 mm among all specimens might not sa-
plate stiffness. During practical BRB design process, the BRB is usually tisfy two times the design story drift. Meanwhile, the configuration of
replaced by the truss element with equivalent cross-sectional area and abrupt variation of the cross-section in a transition part was designed to
corresponding equivalent stiffness to evaluate the cyclic behavior of the form weak part and check whether the specimen failure concentrated at
BRB. The panel zone stiffness is assumed as infinitely large and can be this position (Fig. 1).
negligible in comparison with the equivalent stiffness. However, de- The restraining steel tube infilled with fine aggregate concrete was
signers adopt only the BRB stiffness, ignoring the gusset plate stiffness, used as the restraining element. Four steel plates were welded together
which might result in overestimation of the BRB stiffness to the struc- to form a restraining steel tube with size of 120 × 120 × 8 mm based
tural lateral stiffness. On the other hand, although the core plate strains on Watanabe et al. [1]:
of most typed BRBs can reach 3% and exhibit good ductility, the length
Pe,sc/ Py ⩾ 1.5; Pe,sc = π 2EIsc/ Lsc2 (1)
of yielding portion and non-yielding portion varies with different con-
nection types and configurations at the identical length of work point- where Py is the yield strength of the core plate; Pe,sc is the elastic
to-work point. Therefore, it is necessary to check whether the core plate buckling strength of the restraining steel tube without considering the
strain capacity of the BRB meets the requirements of the AISC 341-10 contribution of concrete bending stiffness; E is the Young’s modulus; Isc
[23] that specifies the deformation of BRB shall be at least two times and Lsc are the moment of inertia and the length of the restraining steel
the design story drift. Finally, the numerical analytical results from tube (Table 1), respectively. The yield strength of the core plate is
Jiang et al. [24] showed that smaller width-to-thickness ratio of core Py = Ayfy, where Ay and fy are the cross-sectional area and yield stress
plate would reduce the material consumption and the multi-wave of the core plate, respectively, and fy can be obtained from the coupon
contact forces of core plate. Thus, a configuration that could easily test. In this study, the restraining members of all specimens were de-
reduce the width-to-thickness ratio of core plate was proposed in this signed following Eq. (1) to prevent global buckling failure. In addition,
paper and described in Section 2.1. the stopper w as welded at the mid core to prevent the restraining
In this study, a total of five BRBs with various types of gusset con- element from slipping. A 2 mm-thick silicone rubber sheet was provided
nections, including pinned gusset connection, bolted gusset connection, as a unbonding material between the core element and the restraining
welded gusset connection, pinned-welded gusset connection as well as element to accommodate transverse expansion of the core plate and
pinned-bolted gusset connection, were studied based on the above also to minimize the frictional effect. Unlike some proposed BRBs that
problems. Firstly, the failure modes of test specimens were recorded. the core plates were usually extended from the restraining elements as
Subsequently, the hysteretic behavior and the relationship between the part of the non-yielding portion, a configuration that the flexible stif-
core plate strain demand and the effective factor were also analyzed feners were welded to the end of core plate to form a non-yielding
and evaluated to check whether the core plate strain capacity of the portion of the BRB was used in five specimens (Fig. 1). There are two
BRB satisfied the demand of at least two times the design story drift. advantages for the alternative configuration. The first is that smaller
The effective factor was described in detail in Section 3.2. Additionally, width-to-thickness ratio of core plate could easily be achieved to reduce
a formula of equivalent stiffness of the BRB with various gusset plates material consumption of restraining element and the multi-wave con-
was suggested based on the Whitmore section and effective length. tact forces of core plate [24]. The second is that the thinner thickness of
Finally, the FE modeling of BRBs with various types of gusset connec- non-yielding portion of the BRB could be used by higher strength steel
tions were established and validated by the experimental results. The and conveniently match with the dimension of gusset plate.
experimental and numerical analysis results will be helpful to promote The bolted and welded gusset plates in the specimens were designed
the application of BRBs in frame structure. to satisfy Eqs. (2) and (3) to avoid failure of gusset plate as result of
yielding in tension and buckling in compression [23,25,26].
2. Test program
ϕPy,gp = ϕbe tgp fy,gp ⩾ Pmax / β (2)
2.1. Test specimens 2
ϕPcr,gp = ϕ (0.658) λc be tgp fy,gp (λ c ⩽ 1.5) ϕPcr,gp = ϕ (0.877/ λc2) be tgp fy,gp (λ c ⩽
The specimen BRB-GP3 with welded gusset connection was de- 1.5) ⩾ Pmax (3)
signed following the guidelines of the AISC 341-10 [23] as the diagonal tgp
KD w fy,gp Ll1 (Lu1) + Ll2 (Lu2) + Ll3 (Lu3)
brace and installed in a steel frame with 2.50 × 103 mm span length where λ c = πr
, Dw = r = 12 and ϕ = 0.9.
3
,
E
and 2.10 × 103 mm story height with scale of 1:2 from engineering Py,gp and Pcr,gp are the yield strength and the buckling strength of the
practice. In order to investigate the equivalent stiffness of the BRBs with gusset plate, respectively. fy,gp and tgp are the yield stress and thickness
various gusset connections in a same frame, pinned gusset connection, of the gusset plate, respectively. Pmax is the maximum strength of the
bolted gusset connection, pinned-welded gusset connection as well as BRB that is computed from Pmax = ωβPy = 501.79 kN, in which 1.5 is
pinned-bolted gusset connection were designed on the basis of spe- used for the strain hardening adjustment factor (ω) in accordance with
cimen BRB-GP3 with welded gusset plates. The dimension and detail JGJ99-2015 [28] and 1.15 is selected for the compression strength
description of the test specimens were presented in Figs. 1–3 and adjustment factor (β) based on Ref. [21]. The parameter of be is the
Table 1. The total length of all specimens was same as 3.04 × 103 mm Whitmore section [25] where the stress is distributed along a 30° from
while the dimension of BRBs and gusset plates of the five specimens the line connecting the first bolt hole to the last bolt hole (Figs. 1 and
were not identical because of various connection type and 4). The effective length (Dw) is the average length of Ll1, Ll2 and Ll3 for

39
J. Wang et al. Engineering Structures 163 (2018) 38–50

Fig. 1. Configuration of BRBs with various types of gusset connections (unit: mm).

the lower gusset plates and Lu1, Lu2 and Lu3 which in bracket for the 5069-2017 [30] were used to check whether the gusset plates met the
upper gusset plates (depicted in Fig. 1). The buckling coefficient, K of requirements in terms of tensile strength of the net cross-section of
0.65 is based on AISC-LRFD [27]. The thickness of the bolted and gusset plate hole, tensile strength of the cross-section of gusset plate end
welded gusset plates was designed as 10 mm. and shear strength of gusset plate.
For the pinned gusset plates, Eqs. (4)–(6) based on the DB34/T

40
J. Wang et al. Engineering Structures 163 (2018) 38–50

Steel tube 25 Core plate 25 Stiffner 25

61 19.5
41

10
41
Unbonding layer Concrete 37.5 37.5

1-1 2-2 3-3


10
Hinge connector 110

29 42 29

90
100

230
10

45
25 48 25
45

4-4 5-5 6-6


Fig. 2. Cross-section of BRB.

1.2Pmax / β order to avoid premature failure of gusset plates and investigate the
σh,gp = ⩽f
2tgp b1 (4) cyclic behavior of BRBs and corresponding core plate strain capacity
and equivalent stiffness of the BRBs with various typed gusset con-
1.2Pmax / β nections, the conservative design concept was employed on the gusset
σe,gp = ⩽f
2tgp (a−2d 0/3) (5) plate.

1.2Pmax / β 2.2. Material properties


τgp = ⩽ fv
2tgp Z (6)
For the steel properties, the steel coupons were cut from test spe-
where b1 = min(2tgp + 16,b−d 0/3) , Z = (a + d 0/2)2−(d 0/2)2 and other cimens to determine the tensile strength, elastic modulus as well as
parameters were illustrated in Fig. 4. It was noted that the amplified breaking elongation according to GB/T228.1-2010 [31]. Q235B grade
factor 1.2 was considered in specification JGJ99-2015 [28] to ensure structural steel in GB50017-2003 [29] with a nominal yield stress of
the gusset plate in the elastic stage under cyclic loading. The design fy = 235 MPa is specified for the core plate and restraining steel tube
value of the tensile strength (f) and shear strength (fv) for the 45 mm and Q345B grade structural steel with fy = 345 MPa is specified for the
thick pinned gusset plate are respectively 265 MPa and 155 MPa in remaining components of specimen. Table 2 showed the test results of
accordance with GB50017-2003 [29]. The computed results of the three steel material. Besides, three concrete cubes with the size of
typed gusset plates (Table 3) satisfied the specification requirements. In 150 × 150 × 150 mm for the cube compressive strength and the size of

161 71 189×200×10
180 70 59 198
23

45
70
70

70
70
87

20
106 125

11

42
70
161

42 42
10
228
228

221×200×10
428
247
230

70

20
20 22
71 186
97 125
122 128 363

(a) Pinned gusset plate-1 (b) Pinned gusset plate-2 (c) Pinned gusset plate-3 (d) Bolted gusset plate-1
406
170×90×10

445 269×200×10 403


30

252×90×10

70
202×90×10

70
45

45 70
×1
377

70
0

359
409

45
305×200×10
409
23

11

23

22
23


11
10

233×200×10 22
11

356 313×200×10
10

271×200×10
10

(e) Bolted gusset plate-2 (f) Bolted gusset plate-3 (g) Welded gusset plate-1 (h) Welded gusset plate-2
Fig. 3. Detail of gusset plate.

41
J. Wang et al. Engineering Structures 163 (2018) 38–50

Table 1
Information of BRBs with various gusset connections.

Specimens Ly (mm) Lb (mm) Lsc (mm) Lbg (mm) Core plate bc × tc (mm) Gusset connection type

BRB-GP1 1.70 × 103 2.50 × 103 2.30 × 103 3.04 × 103 25 × 41 Pinned gusset connection
BRB-GP2 7.60 × 102 2.20 × 103 1.36 × 103 3.04 × 103 25 × 41 Bolted gusset connection
BRB-GP3 1.30 × 103 2.10 × 103 1.90 × 103 3.04 × 103 25 × 41 Welded gusset connection
BRB-GP4 1.50 × 103 2.30 × 103 2.10 × 103 3,04 × 103 25 × 41 Pinned-welded gusset connection
BRB-GP5 1.20 × 103 2.30 × 103 1.80 × 103 3.04 × 103 25 × 41 Pinned-bolted gusset connection

Note: Ly, Lb, Lsc and Lbg are the length of yielding portion, BRB, restraining steel tube and BRB with gusset plates, respectively.

100 × 100 × 300 mm for the elastic modulus were conducted. The is the peak displacement at each displacement amplitude described
cube compression strength of fine aggregate concrete was determined above, Pt (Fig. 6) is the measured peak axial force at each displacement
by the standard cubic compression test. The average cube compressive amplitude, and knyb is the non-yielding portion of the BRB and can be
strength of concrete is 48.9 N/mm2 at 28 days and the elastic modulus computed using Eqs. (13) and (14). The two equations were explained
is 33657 N/mm2. in Section 3.3

2.3. Test setup and loading protocol 3. Test results and analysis

The test was conducted under axial cyclic loading on the electro- 3.1. Failure mode
hydraulic testing machine with a force capacity of 2000 kN. The top
view of the test setup was shown in Fig. 5. The auxiliary device was Table 5 presented the failure modes of all specimens, which were
bolted to rigid reaction frame in advance, then the BRB with gusset not completely consistent with the expected failure mode focused on
connection was assembled to auxiliary device in a horizontal config- the transition part. Three types of failure modes were summarized as
uration to simulate the boundary conditions of beam-to-column joints. follows: tension rupture of the core plate between yielding and stif-
The actuator body was also bolted to the rigid reaction frame to execute fening portion (specimen BRB-GP1 and BRB-GP5), large compression
axial load to test specimens. deformation of the core plate and local bulging failure of the restraining
One string potentiometer labeled SP1 was set between the two end steel tube accompanied by cracking of welding seam (specimen BRB-
of the gusset plates to measure the total deformation of the specimen GP3 and BRB-GP4), compression deformation of the core plate and
(Δ1). The BRB axial deformation (Δ2) was obtained by another string partially rupture of the core plate (specimen BRB-GP2). For the spe-
potentiometer labeled SP2 that fixed at the end of brace. Two linear cimen BRB with pinned gusset connection, the outward deformation
variable displacement transducers labeled LVDT3 and LVDT4 were occurred on the cover plates at the end of the restraining steel tube
mounted on a independent supporter to monitor the out-of-plane dis- when the axial displacement reached 19.47 mm for specimen BRB-GP1
placement of restraining steel tube. The axial force of each specimen and 14.61 mm for specimen BRB-GP4 (Table 5). The cover plates were
was measured by the load cell of the actuator. pulled out because the air gap between the cover plate and the end
The specification JGJ 99-2015 [28] describes a loading protocol plate was only 20 mm. This phenomenon can be used to explain sud-
that three cycles shall be imposed on test specimen at each displace- denly rise of axial peak load for specimen BRB-GP1 and BRB-GP4 de-
ment amplitude of Lb/300, Lb/200, Lb/150, Lb/100 to obtain the hys- picted in Fig. 6. Additionally, no global buckling for the BRB or local
teretic performance of each BRB, where Lb is the length of the BRB. deformation for the gusset plate occurred during the test.
Additional displacement amplitudes of Lb/2000, Lb/1000, Lb/600 were The core plates of specimens BRB-GP1 and BRB-GP2 ruptured at the
applied to each specimen to measure the axial initial stiffness of the cross-section transition between yielding and stiffening portion at the
BRBs with various typed gusset connections and slowed down the core plate strain of −2.33% and −3.62%, respectively (Table 5). It is
process of elastic stage to elasto-plastic stage. Besides, the large dis- indicated that the welding has an adverse effect on the fatigue life and
placement amplitudes of Lb/75, Lb/50, Lb/30 were also introduced to ductility of the BRB core due to welding residual stress and stress
test specimen to study the ultimate state of BRBs with various typed concentration. Therefore, an improved configuration of easing the
gusset connections. The displacement amplitude was listed in Table 4 transition from the stiffening cruciform portion to the flat yielding
and taken as the controlled displacement by string potentiometer la- portion, such as toe-finished method [32], is necessary for avoiding
beled SP2. The data in brackets was the actual core plate strain during unexpectedly premature failure of the core plate.
test process and can be obtained from ε y,t = (Δ2−Pt / k nyb)/ L y , in which Δ2 Specimen BRB-GP2, BRB-GP3 and BRB-GP4 failed during the

P Z
be

30°
L1

a
d0
L2

b
L3

(a) Whitmore’s design methodology (b) Pinned gusset plate’s design methodology
Fig. 4. Design methodology of gusset plate.

42
J. Wang et al. Engineering Structures 163 (2018) 38–50

Table 2
Material properties of steel.

Category Material grade Thickness (mm) Yield strength (N/ Ultimate strength (N/ Elastic modulus (N/ Elongation (%)
mm2) mm2) mm2)

Core plate Q235B 25 283.8 447.0 2.02 × 105 31.2


External steel tube Q235B 8 328.6 452.5 2.01 × 105 27.4
Bolted or welded gusset plate, splice plate and Q345B 10 368.4 511.7 2.04 × 105 25.2
stiffener
Pinned gusset plate Q345B 45 340.8 416.78 2.08 × 105 20.8

compression excursion into next larger strain amplitude. Large com- pinned gusset plates on account of that the existence of clearance in
pression deformation occurred on the core plates of specimen BRB-GP3 pinned gusset connection (Fig. 6). The phenomenon of approximately
and BRB-GP4 and the restraining tube bulged out locally accompanied 1.2 mm bolt slippage was also observed in specimen BRB-GP2 when the
by cracking of welding seam. The failure modes of specimen BRB-GP3 axial force reached 181–199 kN. Then, larger bolt pretension force was
and BRB-GP4 were inconsistent with the expected design failures. A applied on subsequent specimen BRB-GP5 to avoid the bolt slippage
demand-to-capacity ratio (DCR) was adopted to evaluate the capacity of during the loading process.
the restraining steel tube in resisting local bulging failure according to It was noted that the actual core plate strain of all specimens, except
Ref. [33]: specimen BRB-GP1, exceeded 3% and exhibited excellent ductility
(Table 4). However, all specimens, except specimen BRB-GP2, suc-
Pb 4Pmax,t (2s + vεmax tc )(bs−bc )
DCR = = < 1.0 cessfully reached the deformation requirement of two times the design
Ps L w ts2 fy,s (2bs−bc ) (7) story drift specified by AISC 341-10 [23] (the design story drift is de-
4π 2 (EI )
fined as 1% of the story height and equals to 21 mm). These results
where the wavelength L w = eff
and the effective flexural stiff- induced a discussion about the relationship between the core plate
Py
ness (EI)eff is about 0.055EI. s is the thickness of unbonding layer and v strain demand (εy) and the effective factor (α). As shown in Fig. 7, the
denotes the Poisson ratio and equals to 0.5 for steel in inelastic stage. ts, average axial strain of the equivalent brace can be computed, ne-
bs, and fy,s represent the thickness, depth and yield stress of the re- glecting axial deformation of beam and column, as follows:
straining steel tube, respectively. The maximum axial force (Pmax,t) for
ΔL δ cosφ 1
specimen BRB-GP3 and BRB-GP4 were 517.32 kN and 554.38 kN, re- ε= = = θsin2φ
L L 2 (8)
spectively. So that the tested DCR using Eq. (7) was 0.64 and 0.68 for
the two specimens, indicating sufficient tube capacity. It demonstrated where ΔL and L are the deformation and the length between work point-
that the unqualified welding seam between the restraining steel plates to-work point, respectively. θ and φ are the story drift and the in-
mainly led to failure of specimen BRB-GP3 and BRB-GP4 rather than clination angle of the BRB to the horizontal beam, respectively. On the
insufficient capacity of the restraining steel tube in resisting local bul- other hand, ΔL can be divided into yielding portion deformation (ΔLy)
ging failure. Hence, the satisfaction of Eq. (7) and the guarantee of and non-yielding portion deformation (ΔLny) and supposed that the
quality of full penetration weld between the restraining steel plates lower non-yielding portion is same as the upper non-yielding portion.
during welding work should be implemented to resist the outward force Thus, the average axial strain of the equivalent brace can also be ex-
from local buckling deformation of the core plates under compression pressed as:
to prevent unexpected failure.
ΔL = ΔL y + 2ΔL ny (9)

3.2. Hysteretic behavior ΔL y + 2ΔL ny ΔL y L y ΔL ny L−L y


ε= = · + · = ε y α + εny (1−α )
L Ly L L ny L (10)
The axial force-displacement hysteretic curves of all specimens are
presented in Fig. 6, in which the shaded triangles indicated the decline where the effective factor α = L y / L . On the basis of Eqs. (8) and (10),
of load bearing capacity (tension being positive and compression being the core plate strain demand (εy) can be given by:
negative). Each plot showed the axial force of the BRB versus the total 1 1
axial displacement recorded by string potentiometer SP1. The gusset εy = θsin2φ−εny ⎛ −1⎞
2α ⎝α ⎠ (11)
plate end-to-gusset plate end deformation (Δ1) was larger than the BRB
end-to-BRB end deformation (Δ2) due to the elastic deformation in Given that the non-yielding portion deformation is very small and
gusset plate and the clearance in pinned gusset connection. All speci- can be ignored compared to the yielding portion deformation in sub-
mens exhibited stable and repeatable cyclic behavior without stiffness sequent discussion. The story drift (θ) of 1/100 [23] and more severe
and strength degradation. Suddenly increased tension force appeared story drift of 1/70 [9] were selected as the design story drift for the
on specimen BRB-GP1 and BRB-GP4 because of cover plates at both steel structures. The effective factor (α) of the five test specimens was
ends of the restraining steel tube. In addition, there were respectively 0.23, 0.37, 0.40, 0.46, 0.52 and other α of 0.2, 0.3, 0.5, 0.6, 0.7, 0.8
about 3.0 mm, 1.3 mm and 1.1 mm slippage occurring on BRBs with were chosen to discuss using Eq. (11) and depicted in Fig. 8 under the φ

Table 3
Design strength of gusset plate.

Gusset plate ϕ Py,gp (kN) ϕ Pcr,gp (kN) σh,gp σe,gp (N/mm2) gp (N/mm2)

Bolted gusset plate-1 1.01 × 103 9.45 × 102 –


Bolted gusset plate-2 1.01 × 103 9.35 × 102
Bolted gusset plate-3 1.01 × 103 8.78 × 102
Welded gusset plate-1 1.48 × 103 1.48 × 103
Welded gusset plate-2 1.48 × 103 1.48 × 103
Pinned gusset plate-1, 2, 3 – 1.76 × 102 2.43 × 102 0.832 × 102

43
J. Wang et al. Engineering Structures 163 (2018) 38–50

Reaction frame Auxiliary device


Actuator Bolted gusset plate
SP1( 1)
SP2( 2)

LVDT4( 4)
BRB LVDT3( 3)
+ - Pinned gusset plate
Tension Compression

(a) Test setup

Reaction frame

Auxiliary device LVDT4( 4)

SP2( 2)
SP1( 1)

BRB
Bolted gusset plate Pinned gusset plate

LVDT3( 3)

(b) Photograph of test setup


Fig. 5. Test setup and instrumentations.

of 40°. For the specimen BRB-GP2 with bolted gusset plates, it was portion (Ly) would be variable when using various types of gusset
easily found that the actual core plate strain of 3.67% were smaller than connections and corresponding configurations in a same frame in en-
the computed strain demand of 4.23% and 6.04% and not satisfied the gineering practice. Therefore, the Eq. (11) can be conveniently applied
requirement of two times the 1/100 and 1/70 of the story height. The during the design of the BRB to check whether the axial deformation
other four specimens basically met the requirement of two times the 1/ capacity of the BRB meets the demand of two times the design story
100 and 1/70 of the story height. drift.
Fig. 8 also showed that the core plate strain demand (εy) dropped
with increasing of the effective factor (α) because the core plate strain 3.3. Equivalent stiffness of BRB with gusset plates
of the BRB with shorter length of yielding portion (Ly) was higher than
that of the BRB with longer length of yielding portion (Ly). Therefore, During the practical BRB design process, the BRB is usually replaced
the length of yielding portion (Ly) played a decisive role in contribution by the truss element with equivalent cross-sectional area and corre-
to the core plate strain demand (εy). Whereas, the length of yielding sponding equivalent stiffness to evaluate the cyclic behavior of the BRB.

Table 4
Loading protocol.

Loading protocol Specimen Cycle number

BRB-GP1 BRB-GP2 BRB-GP3 BRB-GP4 BRB-GP5

Lb/2000 ± 1.25 (0.06%) ± 1.10 (0.06%) ± 1.05 (0.05%) ± 1.15 (0.06%) ± 1.15 (0.06%) 3
Lb/1000 ± 2.50 (0.12%) ± 2.20 (0.18%) ± 2.10 (0.12%) ± 2.30 (0.12%) ± 2.30 (0.14%) 3
Lb/600 ± 4.17 (0.22%) ± 3.67 (0.36%) ± 3.50 (0.22%) ± 3.83 (0.22%) ± 3.83 (0.27%) 3
Lb/300 ± 8.33 (0.46%) ± 7.33 (0.83%) ± 7.00 (0.49%) ± 7.67 (0.47%) ± 7.67 (0.59%) 3
Lb/200 ± 12.5 (0.71%) ± 11.00 (1.30%) ± 10.50 (0.75%) ± 11.50 (0.72%) ± 11.50 (0.92%) 3
Lb/150 ± 16.67 (0.96%) ± 14.67 (1.77%) ± 14.00 (1.02%) ± 15.33 (0.98%) ± 15.33 (1.22%) 3
Lb/100 ± 25.00 (1.44%) ± 22.00 (2.71%) ± 21.00 (1.55%) ± 23.00 (1.49%) ± 23.00 (1.85%) 3
Lb/75 ± 33.33 (1.92%) ± 29.33 (3.67%) ± 28.00 (2.09%) ± 30.67 (1.99%) ± 30.67 (2.48%) 3
Lb/50 ± 50.00 (2.90%) – ± 42.00 (3.16%) ± 46.00 (3.01%) ± 46.00 (3.75%) 3
Lb/30 – – ± 70.00 (5.30%) ± 76.67 (5.06%) – 3

Note: The data in bracket was the actual core plate strain during test process.

44
J. Wang et al. Engineering Structures 163 (2018) 38–50

600 600
BRB-GP1 BRB-GP2
400 400

Axial force (kN)


Axial force (kN) 200 200

0 0

-200 -200

-400 -400
Test Test
-600 Predicted -600 Predicted

-60 -40 -20 0 20 40 60 -40 -30 -20 -10 0 10 20 30 40


Axial displacement (mm) Axial displacement (mm)
(a) Specimen BRB-GP1 (b) Specimen BRB-GP2

600 600
BRB-GP3 BRB-GP4
400 400
Axial force (kN)

Axial force (kN)


200 200

0 0

-200 -200

-400 -400
Test Test
-600 Predicted -600 Predicted

-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80


Axial displacement (mm) Axial displacement (mm)
(c) Specimen BRB-GP3 (d) Specimen BRB-GP4

600
BRB-GP5
400
Axial force (kN)

200

-200

-400
Test
-600 Predicted

-60 -40 -20 0 20 40 60


Axial displacement (mm)
(e) Specimen BRB-GP5
Fig. 6. Hysteretic curves of specimens.

The equivalent stiffness consists of BRB stiffness, gusset plate stiffness EA


k con,c =
and beam-to-column panel zone stiffness. The panel zone stiffness is too Lcon (13)
large to neglect it in computing the equivalent elastic stiffness.
However, many designers either adopted a empirical factor or without E (A1 −A3 )
k var,c =
considering the contribution of gusset plate stiffness to the equivalent L var (lnA1 −lnA3 ) (14)
stiffness due to the complexity of the gusset plate configuration and the −1 −1
1 1 1 ⎤ 1 1 1 1 ⎤
force action. In order to obtain the gusset plate stiffness, the following klgp,c = ⎡ + + , k ugp,c = ⎡ + + +
Eqs. (12)-(15) were proposed to predict the equivalent stiffness of the ⎢ kl1
⎣ kl2 ⎥
klw ⎦ ⎢
⎣ k u1 k u2 k u3 ⎥
k uw ⎦
BRB with gusset plates based on Whitmore section [25] and effective (15)
length. where kb,c, klgp,c and kugp,c are the computed elastic stiffness of the BRB,
−1 lower gusset plate and upper gusset plate, respectively. Eq. (13) is used
1 1 1 ⎤
k e,c = ⎡ to calculate the elastic stiffness of the segment with constant cross-
⎢ kb,c + klgp,c + k ugp,c ⎥ sectional area, while Eq. (14) aims to compute elastic stiffness of the
⎣ ⎦ (12)
segment with variable cross-sectional area. kb,c can be calculated using
Eq. (13) or (14) for the elastic stiffness of yielding portion, stiffening

45
J. Wang et al. Engineering Structures 163 (2018) 38–50

Table 5
Failure modes of specimens.

Decline of load
Specimen Failure modes
bearing capacity

At -2.33% in the 3rd


BRB-GP1 2.90% tension
excursion
Tension rupture of core plate Tension rupture of cover plate

At -3.62% in the 1st


BRB-GP2 3.67% compression
excursion
Compression deformation and
High order buckling of core plate
partially rupture of core plate

At -2.03% in the 1st


BRB-GP3 5.30% compression
excursion
Large compression deformation of core Local bulging failure of steel tube
plate and welding seam fracture

At -4.04% in the 1st


BRB-GP4 5.06% compression
excursion
Local bulging failure of Tension
Large compression deformation of
steel tube and welding rupture of
core plate
seam fracture cover plate

At -1.77% in the 2nd


BRB-GP5 3.75% tension
excursion
Tension rupture of core plate

8
Work point
Core plate strain demand ( c-%)

7
y
Ln

6 BRB-GP2

5
L

BRB-GP3
Ly

4
BRB-GP1
3

BRB-GP5
y
Ln

BRB-GP4
1

Work point
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Fig. 7. The diagonal configuration of a BRBF.
Effective factor ( )
portion, transition portion or connection portion according to different Fig. 8. Core plate strain demand.
configurations of the BRB, as shown in Fig. 1. The elastic stiffness of
lower or upper gusset plates can be obtained by Eq. (15), in which kl1, respectively (Fig. 1).
kl2, klw, ku1, ku2, ku3, and kuw are the stiffness of corresponding segments The theoretically computed elastic stiffness of the lower and upper
of Dl1, Dl2, Dlw, Du1, Du2, Du3, and Duw and can be computed using Eq. gusset plates (klgp,c, kugp,c), the BRB (kb,c) and the BRB with gusset
(14) based on partitions in gusset plate (Fig. 1). For the corner partition plates (ke,c) was listed in Tables 6 and 7. The tested equivalent stiffness
of gusset plate, the effective width be is the Whitmore section [25] and (ke,t) of the BRB with gusset plates was also presented in Table 7 by
the effective length Dlw and Duw are the average length of Ll1, Ll2 and Ll3 linear regression of the hysteretic curves within the elastic stage.
for the lower gusset plate and Lu1, Lu2 and Lu3 for the upper gusset plate, Moreover, the numerical equivalent stiffness (ke,n) of corresponding

46
J. Wang et al. Engineering Structures 163 (2018) 38–50

Table 6
Elastic stiffness of gusset plates.

Specimens Lower gusset plate stiffness Upper gusset plate stiffness

kl1 (kN/mm) kl2 (kN/mm) klw (kN/mm) klgp,c (kN/mm) ku1 (kN/mm) ku2 (kN/mm) ku3 (kN/mm) kuw (kN/mm) kugp,c (kN/mm)

BRB-GP1 1.15 × 104 – 2.95 × 104 8.26 × 103 1.71 × 104 – – 1.37 × 104 7.62 × 103
BRB-GP2 1.81 × 104 2.89 × 104 5.03 × 103 3.46 × 103 1.97 × 104 2.33 × 104 – 4.60 × 103 3.21 × 103
BRB-GP3 1.05 × 104 9.37 × 103 1.50 × 105 4.80 × 103 1.02 × 104 2.07 × 104 1.52 × 104 4.57 × 104 4.27 × 103
BRB-GP4 1.32 × 104 – 2.29 × 104 8.36 × 103 1.02 × 104 2.07 × 104 1.52 × 104 4.57 × 104 4.27 × 103
BRB-GP5 1.29 × 104 – 2.38 × 104 8.38 × 103 1.68 × 104 3.23 × 104 – 3.34 × 103 2.56 × 103

Table 7
Elastic stiffness of BRB with gusset plates.

Specimens Computed stiffness Test stiffness Numerical stiffness Error

kb,c (kN/mm) ke,c (kN/mm) ke,t (kN/mm) ke,n (kN/mm) ke,c/ke,t ke,n/ke,t ke,c/ke,n

BRB-GP1 105.26 102.54 88.38 95.42 1.16 1.08 1.07


BRB-GP2 160.39 146.37 141.23 148.87 1.04 1.05 0.98
BRB-GP3 122.59 116.29 114.20 108.47 1.01 0.93 1.07
BRB-GP4 113.27 108.90 99.76 102.51 1.09 1.03 1.06
BRB-GP5 132.73 124.70 115.61 123.18 1.08 1.07 1.01

Table 8 3.4. Evaluation of cyclic behavior


Summary of test results of specimens.
Through the analysis of the hysteresis curves of the test specimens,
Specimens β ω μt μc CPD
performance indices of BRBs with various types of gusset connections
BRB-GP1 1.27 1.74 18.95 21.01 582.43 can be obtained, such as the compression strength adjustment factor
BRB-GP2 1.17 1.83 15.34 16.52 373.32 (β), the strain hardening adjustment factor (ω), the ductility (μ) and the
BRB-GP3 1.08 1.92 27.79 16.93 522.69 cumulative plastic deformation (CPD).
BRB-GP4 1.09 1.83 29.66 23.71 600.80
BRB-GP5 1.16 1.86 21.50 21.31 533.25
(1) Compression strength adjustment factor β

specimens was obtained by the FE model to check out the accuracy of The experimental results presented in Table 8 showed that the
the proposed formulas and the FE model will be discussed in detail in maximum β (=Pc,max/Pt,max) was 1.27 at the core plate strain amplitude
proceeding Section 4. As can be seen from Table 7, the computed of 2.90% for specimen BRB-GP1 and the minimum β was 1.08 at the
stiffness (ke,c) of the BRB with bolted gusset plates and the BRB with core plate strain amplitude of 3.16% for specimen BRB-GP3. The AISC
welded gusset plates agreed well with the tested stiffness (ke,t) within 341-10 [23] specifies that the compression strength adjustment factor
5% difference, while there were differences of 16%, 9% and 8% be- (β) shall be less than 1.3 to prevent large unbalanced force acting on
tween the computed stiffness (ke,c) and the tested stiffness (ke,t) for the beams and columns, and in this case all specimens conformed to this
BRB with pinned gusset plates, the BRB with pinned-welded gusset requirement.
plates and the BRB with pinned-bolted gusset plates, respectively. The
results of larger difference possibly because that the discreteness of (2) Strain hardening adjustment factor ω
experimental data and the clearance in pinned gusset connection. On
the other hand, the average difference of all specimens between nu- The strain hardening adjustment factor (ω) is the ratio of the max-
merical stiffness (ke,n) and tested stiffness (ke,t) was 6% showed the imum axial tension force (Pt,max) to the yield strength of the BRB (Py).
accuracy of the FE model. Meanwhile, results showed the computed The experimental value ω of five specimens ranges from 1.74 to
stiffness (ke,c) can fit the numerical stiffness (ke,n) well for the five 1.92. The average value ω of 1.84 was larger than the recommended
specimens with an average difference of 4.4%. The analysis results value of 1.5 specified in JGI 99-2015 [28] for the Q235B grade struc-
demonstrated that the proposed Eqs. (12)-(15) can be used to predict tural steel.
the equivalent stiffness of BRBs with various typed gusset connections.
Furthermore, it should be noted that the values of computed stiffness (3) Ductility μ and cumulative plastic deformation CPD
(kb,c) of BRBs were all larger than the computed stiffness (ke,c) of BRBs
with gusset plates from Table 7, which indicated that it would over- Table 8 summarized the ductility in tension (μt), ductility in com-
estimate the contribution of the BRB stiffness to the structural lateral pression (μc) and cumulative plastic deformation (CPD).
stiffness if ignoring the elastic stiffness of gusset plate. Although the core plate strain of specimen BRB-GP2 with bolted
It also found that the tested stiffness of specimen BRB with pinned gusset plates reached 3.67%, the ductility of μt, μc and CPD were 15, 17
gusset plates was 88.38 kN/mm smallest among others. The tested and 373, respectively and they were the smallest among the five spe-
stiffness of specimen BRB-GP2, BRB-GP3, BRB-GP4 and BRB-GP5 was cimens due to that its core plate length was only 760 mm. However, the
increased by 59.8%, 29.2%, 12.9% and 30.8%, respectively, in com- five specimens exhibited good ductility capacity and the value of CPD
parison with specimen BRB-GP1. This inferred that the gusset connec- remarkably exceeded the minimum required value of 200 specified in
tion type and corresponding configuration could affect the equivalent the AISC 341-10 [23].
stiffness of brace and provide an alternative method to adjust the
equivalent stiffness of brace to achieve accurately expected structural
lateral stiffness.

47
J. Wang et al. Engineering Structures 163 (2018) 38–50

Fig. 9. FE analysis modeling.

n
4. Numerical analysis Ck pl
αk = ∑ γk
(1−e−γk ε )
k= 1 (17)
4.1. Finite element modeling
where the constants Ck and γk can be determined based on the coupon
Numerical study using the FE program ABAQUS [34] was carried test; the rate of Ck/γk expresses the maximum change value in backs-
out for the five specimens to predict the cyclic behavior of BRBs with tress; γk is the rate of change of the backstress as the plastic strain in-
various typed gusset connections and to achieve higher accuracy. The creases. For the kinematic hardening, the critical parameters including
specimen BRB-GP4 with pinned-bolted gusset plates was selected as n = 3, C1 = 8.0 × 103 MPa, γ1 = 100, C2 = 9.0 × 104 MPa,
typical FE model and illustrated in Figs. 9 and 10. All FE models de- γ2 = 2.40 × 10 , C3 = 1.20 × 10 MPa, γ3 = 0 were calibrated by
3 3

veloped in this study were same as the test specimens. The solid ele- curve fitting from stress-strain relationship of the coupon test.
ment C3D8R was adopted to model the components of the BRB with Since the remaining steel components including the non-yielding
gusset plates except the restraining steel tube using shell element S4R. portion of BRB and gusset plate remained elastic state during cyclic
In order to simulate accurately the large plastic deformation and higher loading. The simple bilinear behavior was used for the remaining of
mode buckling of core plate, a refined mesh of 30 mm was applied to steel components except the core plate. The post-yielding modulus was
the core element, the gusset plate, the pin, the bolt and the splice plate, taken as 1% of the elastic modulus in the plastic stage. The elastic
while a coarser mesh of 60 mm was used for the restraining element to modulus, yield and maximum strength of the remaining steel compo-
improve the computational efficiency. nents were obtained from the coupon test, as listed in Table 2. A
The nonlinear combined isotropic and kinematic hardening rule practical and accurate model proposed by Han et al. [36] was used to
[35] were employed for the core plate to describe the strength devel- simulate the infilled concrete mechanical performance in this paper.
opment of core plate accurately under axial cyclic loading. On the one The contact properties with hard contact and tangential Coulomb
hand, the isotropic hardening model defines the yield surface size (σ0) frictional behavior were established between the core element and the
which is a function of the equivalent plastic strain (εpl) by Eq. (16): restraining element for all BRBs and also employed between the pin and
pl the hole in BRB clevises. The coefficient of Coulomb friction between
σ 0 = σ |0 + Q∞ (1−e−bε ) (16)
the core plate and the restraining element was set as 0.05 and value of
where σ∣0 represents the yield stress when equivalent plastic strain 0.45 [29] was applied for the pinned gusset connection. The tie con-
reaches zero; Q∞ denotes the maximum change in yield surface and b is straint was employed for the bolted and welded gusset connections. The
the rate factor at which the yield surface changes as the plastic strain damping factor for contact control was set to be 2 × 10−4 to achieve a
increases. For the isotropic hardening, the parameters Q∞ = 30 MPa better convergence in the simulation process. In addition, the dis-
and b = 1.2 were adopted in core plate based on the coupon test. placement of the center of core plate and restraining element in the
On the other hand, the kinematic hardening model gives a definition longitudinal direction were coupled to prevent the axial rigid body
of backstress (αk), as described in Eq. (17): movement of the restraining element.

(a) Pinned (c) Bolted gusset


(b) Core element
gusset plate plate

(f) Splice plate and


(d) Pin (e) Restraining element
bolt
Fig. 10. 3-D idealization of the specimen components.

48
J. Wang et al. Engineering Structures 163 (2018) 38–50

(a) Specimen BRB-GP2 (b) Specimen BRB-GP5


Fig. 11. Core plate deformation when specimen failures (amplification factor = 5).

In view of the manufacture error of braces, initial geometrical im- maximum tension and compression forces were also observed at initial
perfection was assumed as its first order buckling shape with central phase due to the clearance in pinned gusset connection resulting in
amplitude of 1/1000 of the BRB length in the FE model. Finally, one inaccurate measurement of axial displacement. The comparison of
end of gusset plate was fixed and the other end was for displacement equivalent stiffness between experimental and numerical results were
loading only in the axial direction of core plate without any other also discussed in Section 3.3. The average deviation of ke,n/ke,t and kec/
transition or rotation. The displacement loading protocol recorded by ken was respectively 6% and 4.4% for the five specimens and showed
string potentiometer SP1 were imposed on the FE models. that the FE model could predict the elastic stiffness of BRBs with var-
ious types of gusset connections well.
Another important parameter of predicting BRB performance is the
4.2. Numerical analysis results
total energy-dissipating capacity and the numerical and experimental
results of total energy enclosed by all hysteretic cycles were illustrated
The compressive deformations of the core plates of specimen BRB-
in Fig. 13. The average difference of the total dissipated energy was
GP2 and BRB-GP5 were depicted in Fig. 11 when each specimen failed.
4.9% which demonstrated that they agreed each other very well. Fig. 13
The multiwave buckling deformation was appeared on the core plate
also showed specimen BRB-GP2 with the shortest length of core plate
with increasing axial displacement. It was concluded that the com-
dissipated the least energy and specimen BRB-GP4 with the longer
pressive buckling shape of core plate could be simulated by FE analysis.
length of core plate dissipated the most energy. The energy consump-
The numerical hysteresis results of five specimens were plotted in Fig. 6
tion of specimen BRB-GP1 with the longest length of core plate justly
in comparison with the experimental results. Fig. 6 showed that the
ranked second among five specimens due to premature failure. It was
hysteretic curves kept in good agreement with the experimental data.
indicated that the length of the core plate played a controlling role in
Meanwhile, the FE model can predict the asymmetric behavior well at
energy dissipation. In summary, the FE model provided a reasonable
large axial strain amplitude. Furthermore, the prediction error of the
prediction for the compressive buckling shape, equivalent stiffness,
maximum tension force (Pt,max) and compression force (Pc,max) at each
peak strength and dissipated energy of the BRB with gusset plates.
loading amplitude for all specimens was depicted in Fig. 12. There were
two main reasons for the larger error of maximum tension force and fell
within 15%: one reason was that the cover plates were not considered 5. Conclusions
in FE model and cannot simulate suddenly increased tension force oc-
curring on specimen BRB-GP1 and BRB-GP4; another reason was that Cyclic loading test and numerical analysis of the BRBs with various
higher strain hardening was still emerged on experimental specimen types of gusset connections were carried out to investigate the effect of
BRB-GP2, BRB-GP3 and BRB-GP5 before specimen failure while lower gusset connection type on the cyclic behavior in terms of the hysteretic
strain hardening occurred on FE models at the later loading amplitudes. performance, core plate strain capacity, equivalent stiffness, ductility
The prediction error of the compression force at each loading cycle was and energy dissipation capacity. The following conclusions can be
mostly less than 8% and even within 5% showed that the prediction drawn with the limitation of the experimental research reported in this
capability of the developed FE model. The higher deviations for the paper.

BRB-GP1 BRB-GP2 BRB-GP3 BRB-GP4 BRB-GP5

15 15

10 8% error margin
10

5% error margin
5 5
Error (%)

Error (%)

0 0

-5 -5
5% error margin

-10 -10 8% error margin

-15 -15
0 1 2 3 4 5 6 0 1 2 3 4 5 6
(%) (%)
(a) Maximum tensile force (b) Maximum compressive force
Fig. 12. Percentage error of predicted maximum tension or compression force for specimens.

49
J. Wang et al. Engineering Structures 163 (2018) 38–50

700 acknowledge the assistance of Dr. Bo Wang, Quan Deng, Huijie Zhang
Test and Jinzhou Zuo of Hefei University of Technology.
600 Predicted 4.4%
Energy dissipation (kN m)

3.6%
Diff. 5.0% Diff. References
500 Diff. 6.3%
Diff. [1] Watanabe A, Hitomi Y, Saeki E, Wada A, Fujimoto M. Properties of brace encased in
400 buckling-restraining concrete and steel tube. In: Proceedings of the 9th world conference
on earthquake engineering, Tokyo-Kyoto, Japan; 1988. p. 719–24.
5.3% [2] Iwata M, Murai M. Buckling-restrained brace using steel mortar planks: performance
300 Diff. evaluation as a hysteretic damper. Earthq Eng Struct Dyn 2006;35:1807–26.
[3] Tremblay R, Bolduc P, Neville R, DeVall R. Seismic testing and performance of buckling-
restrained bracing systems. Can J Civ Eng 2006;33:183–98.
200 [4] Palazzo G, López-Almansa F, Cahís X, Crisafulli F. A low-tech dissipative buckling re-
strained brace. Design, analysis, production and testing. Eng Struct 2009;31(9):2152–61.
100 [5] Ju YK, Kim MH, Kim J, Kim SD. Component tests of buckling-restrained braces with
unconstrained length. Eng Struct 2009;31(2):507–16.
[6] Chou CC, Chen SY. Subassemblage tests and finite element analyses of sandwiched
0 buckling-restrained braces. Eng Struct 2010;32(8):2108–21.
BRB BRB BRB BRB BRB [7] Zhao JX, Wu B, Ou JP. A novel type of angle steel buckling-restrained brace: Cyclic be-
-GP1 -GP2 -GP3 -GP4 -GP5 havior and failure mechanism. Earthq Eng Struct Dyn 2011;40(10):1083–102.
[8] Wang CL, Usami T, Funayama J. Low-cycle fatigue testing of extruded aluminium alloy
Fig. 13. Total energy enclosed by all hysteretic cycles for specimens. buckling-restrained braces. Eng Struct 2013;46:294–301.
[9] Piedrafita D, Cahis X, Simon E, Comas J. A new perforated core buckling restrained brace.
Eng Struct 2015;85:118–26.
(1) The main failure modes of the five BRBs with various gusset con- [10] Guo YL, Zhang BH, Zhu BL, Zhou P, Zhang YH, Tong JZ. Theoretical and experimental
studies of battened buckling-restrained braces. Eng Struct 2017;136:312–28.
nections were summarized: tension rupture of the core plate be- [11] Tsai KC, Hsiao PC. Pseudo-dynamic test of a full-scale CFT/BRB frame-Part II: Seismic
tween yielding and stiffening portion; large compression deforma- performance of buckling-restrained braces and connections. Earthq Eng Struct Dyn
tion of the core plate and local bulging failure of the restraining 2008;37(7):1099–115.
[12] López WA, Gwie DS, Lauck TW, Saunders CM. Structural design and experimental ver-
steel tube accompanied by cracking of welding seam; compression ification of a buckling-restrained braced frame system. Eng J 2004;41(4):177–86.
deformation of the core plate and partially rupture of the core plate. [13] Fahnestock LA, Ricles JM, Sause R. Experimental evaluation of a large-scale buckling-
restraining braced frame. J Struct Eng 2007;133(9):1205–14.
(2) All specimens exhibited stable hysteretic performance and high
[14] Wigle VR, Fahnestock LA. Buckling-restrained braced frame connection performance. J
ductility. The cumulative plastic deformation remarkably exceeded Constr Steel Res 2010;66(1):65–74.
the minimum required value of 200 specified in the AISC 341-10. It [15] Kaneko K, Kasai K, Motoyui S, Sueoko T, Azuma Y, Ooki Y. Analysis of beam-column-
gusset components in 5-story value-added frame. 14th WCEE, Beijing; 2008.
found that the gusset connection type affected the ductility, cu- [16] Tsai KC, Lin PC, Wu AC, Chuang MC, Li CH, Wang KJ. Seismic design and experiment of
mulative plastic deformation capacity, dissipated energy and single and coupled corner gusset connections in a full-scale two-story buckling-restrained
equivalent stiffness of brace at the same cross-section dimension of braced frame. Earthq Eng Struct Dyn 2015;44(13):2177–98.
[17] Chou CC, Chen PJ. Compressive behavior of central gusset plate connections for a
core plate and the total length of the BRB with gusset plates. buckling-restrained braced frame. J Constr Steel Res 2009;65(5):1138–48.
(3) The relationship between the core plate strain demand and the ef- [18] Chou CC, Liu JH, Pham DH. Steel buckling-restrained braced frames with single and dual
corner gusset connections: seismic tests and analyses. Earthq Eng Struct Dyn
fective factor showed that the core plate strain demand descended 2012;41(7):1137–56.
with increase of the effective factor. The proposed formula could be [19] Chou CC, Liu JH. Frame and brace action forces on steel corner gusset plate connections
conveniently applied to check whether the axial deformation ca- in buckling-restrained braced frames. Earthq Spectra 2012;28(2):531–51.
[20] Khoo HH, Tsai KC, Tsai C, Tsai CY, Tsai CY, Wang KJ. Bidirectional substructure pseudo-
pacity of the BRB meets the demand of two times the design story dynamic tests and analysis of a full-scale two-story buckling-restrained braced frame.
drift. Earthq Eng Struct Dyn 2016;45(7):1085–107.
[21] Chuang MC, Tsai KC, Lin PC, Wu AC. Critical limit states in seismic buckling-restrained
(4) On the basis of Whitmore section and effective length, a formula of
brace and connection designs. Earthq Eng Struct Dyn 2015;44(10):1559–79.
the equivalent stiffness of the BRB with various gusset plates was [22] Judd JP, Marinovic I, Eatherton MR, Hyder C, Phillips AR, Tola AT, et al. Cyclic tests of
proposed and validated experimentally and numerically. It was all-steel web-restrained buckling-restrained brace subassemblages. J Constr Steel Res
2016;125:164–72.
noted that the ignorance of the elastic stiffness of gusset plate would [23] American Institute of Steel Construction (AISC). Specification for Structural Steel
overestimate the contribution of the BRB stiffness to the structural Buildings. American Institute of Steel Construction: Chicago, Illinois; 2010.
lateral stiffness. Meanwhile, the influence of different types of [24] Jiang ZQ, Guo YL, Zhang BH, Zhang XQ. Influence of design parameters of buckling-
restrained brace on its performance. J Constr Steel Res 2015;105:139–50.
gusset connections on the equivalent stiffness of brace should be [25] Whitmore RE. Experimental investigation of stresses in gusset plates. Bulletin No.16,
considered in the design of the BRBF structures, especially for the Engineering Experiment Station, University of Tennessee; 1952.
[26] Thornton WA. Bracing connections for heavy construction. Eng J (AISC)
bolted gusset connection.
1984;21(3):139–48.
(5) The FE analytical models for the all specimens were developed and [27] Manual of steel construction load and resistance factor design (AISC). Chicago (IL):
verified by the experimental results in terms of peak axial force at American Institute of Steel Construction; 2005.
[28] JGJ99-2015. Technical specification for steel structure of tall buildings. Beijing: China
each loading amplitude, equivalent stiffness and total dissipated Architecture and Building Press; 2015 [in Chinese].
energy. It was demonstrated that the FE modeling can predict the [29] GB50017-2003 Code for design of steel structures. Beijing: China Planning Press; 2003 [in
cyclic behavior of the BRBs with various gusset connections well. Chinese].
[30] DB34/T 5069–2017 Technical specification for buckling-restrained brace structures.
(6) The experimental and numerical work presented in this paper will Hefei: Anhui Engineering Construction Standard Design Press; 2010 [in Chinese].
provide the reference for design and application of BRBs with [31] GB/T228.1-2010. Metallic materials-Tensile testing-Part 1: method of test at room tem-
perature. Beijing: Standards Press of China; 2010 [in Chinese].
various gusset connections in structure buildings. Further numer- [32] Wang CL, Usami T, Funayama J. Improving low-cycle fatigue performance of high-per-
ical study is needed to study the mechanical behavior of the gusset formance buckling-restrained braces by toe-finished method. J Earthq Eng
plate and the seismic performance of the frame structure with BRBs 2012;16(8):1248–68.
[33] Lin PC, Tsai KC, Chang CA, Hsiao YY, Wu AC. Seismic design and testing of buckling-
using various gusset connections. restrained braces with a thin profile. Earthq Eng Struct Dyn 2016;45(3):339–58.
[34] ABAQUS. Standard user's manual version 6.10. Pawtucket, RI: Hibbitt, Karlsson &
Sorensen, Inc.; 2010.
Acknowledgements
[35] Chaboche JL. Time-independent constitutive theories for cyclic plasticity. Int J Plast
1986;2(2):149–88.
This work described in paper is supported by the National Natural [36] Han LH, Yao GH, Zhao XL. Tests and calculations of hollow structural steel (HSS) stub
columns filled with self-consolidating concrete (SCC). J Constr Steel Res
Science Foundation of China (Project 51478158 and Project 51508143) 2005;61(9):1241–69.
and the New Century Excellent Talents in University (Project NCET-12-
0838), which is greatly appreciated. The authors would also like to

50

Anda mungkin juga menyukai