Chou Et Al 2012

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS
Earthquake Engng Struct. Dyn. 2012; 41:1137–1156

Published online 20 October 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/eqe.1176
Steel buckling-restrained braced frames with single and dual corner

gusset connections: seismic tests and analyses
Chung-Che Chou1,2,*,†, Jia-Hau Liu1 and Dinh-Hai Pham1

1
Department of Civil Engineering, National Taiwan University, Taiwan
2
National Center for Research on Earthquake Engineering, Taiwan
SUMMARY
The design of a three-story buckling-restrained braced frame (BRBF) with a single-diagonal sandwiched
BRB and corner gusset was evaluated in cyclic tests of a one-story, one-bay BRBF subassembly and
dynamic analyses of the frame subjected to earthquakes. The test focused on evaluating (1) the seismic per-
formance of a sandwiched BRB installed in a frame, (2) the effects of free-edge stiffeners and dual gusset
configurations on the corner gusset behavior, (3) the frame and brace action forces in the corner gusset,
and (4) the failure mode of the BRBF under the maximum considerable earthquake level. The subassembly
frame performed well up to a drift of 2.5% with a maximum axial strain of 1.7% in the BRB. Without free-
edge stiffeners, the single corner gusset plate buckled at a significantly lower strength than that predicted by
the specificationof American Institute of Steel Construction (2005). The buckling could be eliminated by
using dual corner gusset plates similar in size to the single gusset plate. At low drifts, the frame action force
on the corner gusset was of the same magnitude as the brace force. At high drifts, however, the frame action
force significantly increased and caused weld fractures at column-to-gusset edges. Nonlinear time history
analyses were performed on the three-story BRBF to obtain seismic demands under both design and maxi-
mum considerable levels of earthquake loading. The analytical results confirmed that the BRB and corner
gusset plate achieved peak drift under cyclic loading test. Copyright © 2011 John Wiley & Sons, Ltd.
Received 6 February 2011; Revised 29 August 2011; Accepted 2 September 2011
KEY WORDS: buckling-restrained braced frame; dual gusset; frame action; cyclic test; nonlinear time
history analysis
1. INTRODUCTION
The buckling-restrained brace (BRB) is increasingly used for lateral load resistance in steel braced
structures because of its stable energy dissipation and easy construction [1–3]. Most tests were
component tests, and tests conducted on full-scale buckling-restrained braced frames (BRBFs) were
limited [4–8]. Recent studies [7–9] showed that before the BRB reaches maximum capacity, and
even when the gusset plate used to connect the BRB and framing members is designed in compliance
with codes [10, 11] of American Institute of Steel Construction (AISC), the beam-to-gusset and
column-to-gusset welds can fracture, the beam and column may exhibit local buckling near the
corner gusset, and the gusset itself may buckle. When a single-diagonal BRB is used, a corner gusset
plate is subjected to brace action and frame action forces [12, 13]; however, frame action force is not
considered in the AISC specification [11]. A methodology for determining frame action forces in the
corner gusset plate based on beam and column shears can be found elsewhere [13, 14].
To understand the interaction between the BRB and corner gusset connection, a three-story BRBF
(Figures 1(a) and (b)) was designed based on AISC seismic provisions [10], and a full-scale one-story,
*Correspondence to: Chung-Che Chou, Department of Civil Engineering, National Taiwan University, Taiwan.
†
E-mail: cechou@ntu.edu.tw
Copyright © 2011 John Wiley & Sons, Ltd.

1138 C.-C. CHOU, J.-H. LIU AND D.-H. PHAM
Figure 1. BRBF (unit: mm).
one-bay frame subassembly (Figure 1(c)) was tested in the National Center for Research on Earthquake
Engineering, Taiwan. The BRBF was designed for a high seismic zone in either Taiwan or the United
States. The bay width was 5 m and the story height was 3.92 m for the first story and 3.48 m for other
stories. A sandwiched BRB capable of sustaining stable hysteretic responses up to a core axial strain of
2.6% [15] was used in the frame subassembly. As in the typical BRBF, a single gusset configuration
with eight splice plates was used to connect the BRB and framing members. To enhance the out-of-
plane stability of the connection, a dual-gusset-plate configuration similar in size to the single-
gusset-plate configuration was also adopted in the study. Time-history analyses for the three-story
prototype frame were performed to obtain seismic demands using two sets of 10 historical
earthquake records, which were representative of design-based earthquake (DBE) and maximum-
considerable earthquake (MCE) seismic hazard levels. The objectives of this study were to evaluate
(1) the cyclic performance of the BRB installed in the frame, (2) the effects of free-edge stiffeners
and dual gusset plates on gusset behavior, (3) frame action and brace action forces acting on the
single or dual corner gussets, and (4) seismic demands and failure modes of the BRBF under both
the DBE and MCE levels.
After presenting the design of the three-story BRBF, five gusset configurations are designed by
considering brace and frame action forces in the corner gusset. Calibrating the cyclic responses of
the BRBF and the moment-resisting frame (MRF) subassemblies during the tests obtains the BRB
Copyright © 2011 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2012; 41:1137–1156
DOI: 10.1002/eqe
BUCKLING-RESTRAINED BRACED FRAME 1139
axial force and displacement, leading to performance evaluations of the BRB and corner gusset
connections. Seismic responses of the three-story frame in time-history analyses are compared to
those of the frame subassembly in the tests. Finally, the results of cyclic tests and time-history
analyses of the BRBF and gusset performances are discussed.
2. DESIGN OF A THREE-STORY BRBF
Figures 1(a) and (b) show the elevation and plan of the prototype building, which was assumed to be
located on stiff soil in Los Angeles, CA. Two one-bay BRBFs providing lateral load resistance in the
east–west direction were considered in this study; each BRBF was composed of steel columns
(H356 256 15 22), steel beams, and single-diagonal BRBs. The design dead loads were 5.28
kPa (110 psf) and 4.32 kPa (90 psf) for the floors and the roof while the live load for the floors and
the roof was 2.39 kPa (50 psf). Effective seismic weights for the floors and the roof were 2112 kN
and 1728 kN, respectively, resulting in a total seismic weight of the building equal to 5952 kN. The
design followed International Building Code (IBC) 2000 [16] with a force reduction factor R of 8,
an overstrength Ω0 of 2.5 and a displacement amplification factor Cd of 5. The mapped MCE
spectral response acceleration at a short period (SS) and 1 s (S1) was 1.5 g and 0.6 g, respectively.
For the building located at site class D, the site coefficients Fa and Fv were 1.0 and 1.5,
respectively, leading to design spectral response accelerations at a short period and 1 s of 1.0 g and
0.6 g, respectively. The structural period T and the seismic response coefficient Cs calculated by the
codified method were 0.61 and 0.122, respectively, so the seismic design base shear Vdes for one
BRBF was 364 kN.
The BRB is composed of a ductile steel core and a buckling-restraining member as shown in
Figure 2. Compared with conventional BRBs that have a steel core inserted into a restraining
member, using fully tensioned bolts to sandwich a core plate between a pair of restraining members
enables faster assembly. The advantage is the ability to disassemble the brace, which provides an
opportunity for inspection of the core after earthquakes and also enables replacement of the core
plate independently of the restraining members. A small gap between the steel core and the
buckling-restraining member is provided to minimize the transfer of axial force from the steel core
Figure 2. BRB details.
DOI: 10.1002/eqe
to the buckling-restraining member. Only the steel core is intended to provide axial stiffness to the
BRB. The BRB maximum tension force, Tmax, and maximum compression force, Cmax, are
T max ¼ Ωh ΩAy Fy (1)
C max ¼ bΩh ΩAy Fy ; (2)
where Ωh is the strain hardening factor, Ω is the material over-strength factor, Ay is the cross-sectional
area of the steel core, and Fy is the nominal yield strength of steel. According to component tests of the
BRB [15], compression strength adjustment factor b is 1.15.
The design of the restraining member and the number of bolts were determined by ratios of
Pe/AyFy ≥ 2.5 and bolt tensile capacity–demand ratio ≥1.5, where Pe is the Euler buckling load of the
restraining member such that global buckling of the restraining member can be excluded at maximum
compression load Cmax. Table I summarizes sizes of three first-floor BRBs. Because the BRBs were
manufactured according to different gusset configurations, the core yield length, Ly, was different. In
all BRBs, core plate width, bc, and thickness, tc, were 112 mm and 14 mm, respectively. The moment
of inertia of the restraining members, Ir,g, was calculated by considering the contributions of the face
plate If, channel Ic, and concrete Icon. Specifically, the concrete contributed approximately 50% of the
overall moment of inertia of the restraining member. Two restraining members were connected with
19-mm-diameter A490 bolts. The steel specified for the channel was ASTM (American Society for
Testing and Materials) A36 steel with nominal yield strength of 250 MPa, and that for the core, side,
and face plates was ASTM (American Society for Testing and Materials) A572 Gr. 50 steel. Table II
lists data obtained from tensile coupon tests of steel plates. Assuming a strain hardening factor (Ωh) of
1.3 and a compression strength adjustment factor (b) of 1.15 [15], maximum tension force (Tmax) and
maximum compression force (Cmax) of the BRB were 795 kN and 914 kN, respectively.
3. DESIGN OF CORNER GUSSET CONNECTION
The design procedure for the gusset plates, which was adopted from the AISC specification [11], is
based on the investigations of Whitmore [17] and Thornton [18]. The peak stress is expected to
occur at the location of the Whitmore effective width, be, in the gusset plate (Figure 3(a)), so the
gusset tensile yield strength is

Rn ¼ be tg Fy (3)
where tg is the gusset thickness. In a study of compressive strength in gusset plates, Thornton proposed
an approach for estimating the buckling capacity of a unit strip below the Whitmore effective width.
According to the AISC specification [11], the critical unbraced length, Lc, of the unit strip is
calculated by averaging lengths Lc1, Lc2, and Lc3; gusset buckling strength is
Pcr ¼ ð0:658Þlc Fy be tg ; lc⩽1:5

2
(4)
0:877
Pcr ¼
lc
2 Fy be tg ; lc > 1:5
Table I. BRB size.

Core plate Restraining member
Specimen bc (mm) tc (mm) Ly (mm) Channel (mm) If/Ir.g (%) Ic/Ir.g (%) Icon/Ir.g (%) Ir,g (mm4)
1 112 14 3832 112 85 4.5 4.5 4 40 56 28,839,662

2 and 3 112 14 4400 112 85 4.5 4.5 4 40 56 28,839,662
DOI: 10.1002/eqe
Table II. Material strength.

BRB core Gusset
Specimen 1 2 3 1, 2 3, 4, 5
Thickness (mm) 14 14 14 14 8
Fy (MPa) 390 375 380 385 448
Fu (MPa) 527 554 520 525 514
Figure 3. Frame and brace action in the gusset plate.

pffiffiffiffiffiffiffiffiffiffiffi
where lc ¼ kLc =pr Fy =E , E is the elastic modulus of steel, and r is the radius of gyration of the
Whitmore section. Buckling coefficient k is 0.65 when the gusset is restrained by free-edge
stiffeners. When the gusset plate is susceptible to out-of-plane movement and side-sway buckling, a
coefficient k = 2 or 1.2 on page 13–32 of the AISC specification [11] is used to calculate
compression capacity. According to the AISC seismic provisions [10], the BRBF system requires
that the gusset tensile yield strength be greater than Tmax and the gusset buckling strength be greater
than Cmax (Tables III and IV).
3.1. Frame action force

Studies [7,12, 13] show that the gusset plate, which is located in the corner of the beam-to-column
connection, is subjected to both brace and frame action forces. The beam-to-column connection
DOI: 10.1002/eqe
Table III. Tensile capacity of the gusset plate.

Specimen 1 2 3 4 5
be (mm) 272 272 353 353 448
tg (mm) 14 14 28 28 28
Rn (kN) 1466 1466 2530 2530 3211
Tmax = 795 kN.
Table IV. Compressive capacity of the gusset plate.

With edge stiffener Without edge stiffener
Item k = 0.65 k = 2.0 k = 1.2
Specimen 1 3 2 4 5 2
Lc (mm) 183 174 183 174 — 183
Pcr (kN) 1368 2506 759 2312 — 1157
Cmax = 914 kN.
closes when the BRB is under tension, which produces compression force, Fgm, on the gusset (Figure 3(b)).
A study [14] proposed a method of evaluating frame action force in terms of shear forces in the beam
and column. This method used an equivalent strut model (Figure 3(c)) similar to that used for the
rehabilitated moment connection [19, 20] to determine force components acting on the beam-to-
gusset and column-to-gusset interfaces. The strut axial force, Fgm, caused by beam shear, Vb, and
column shear, Vc, is [13, 14]
Fgm ¼ A′ fC1 Vb þ C2 Vc g (5)
where
Lg
A′ ¼ n o (6)
a ak1
g
þ db bþdb
2EIb 2 þ dc b
2EIc 2 þ bdc
2a
db a
C1 ¼ L′ þ (7)
2EIb 2

dc 00 b
C2 ¼ L þ (8)
2EIc 2
where Lg is the strut length, Ib is the moment of inertia of beam, Ic is the moment of inertia of column, L′
is the distance from strut end (point A) to the beam inflection point, L00 is the distance from strut end
(point B) to the column inflection point, db is the beam depth, dc is the column depth, a (= 0.6Lgb) is
the distance from the column flange face to point A, b (= 0.6Lgc) is the distance from the beam
flange face to point B, and Lgb and Lgc are the gusset dimensions. Axial stiffness, kg, of an
equivalent strut is
bs tg E tg E
kg ¼ ¼ (9)
Lg 2
where bs (= Lg/2) is the effective strut width [20].
DOI: 10.1002/eqe
From the pushover analysis of the BRBF, the BRB force and shears in the beam and column give
the force components in the corner gusset caused by brace action (PBRB) and frame action (Fgm).
Adding stresses produced by frame action and brace action gives the maximum normal stress, smax,b,
and maximum shear stress, tmax,b, at the gusset tip. The von-Mises stress at the gusset tip is
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼
s s2max;b þ 3t2max;b ⩽ F y (10)
The gusset and free-edge stiffener size can be determined by limiting von-Mises stress less than the
yield strength of steel at a design drift (i.e., 1%).
Figure 1(d) shows the pushover analysis results for the prototype frame and the one-story
subassembly frame. The BRBs on the bottom two floors started yielding at step A, and the beam
and column started yielding at steps B and D, respectively. The IBC load pattern [16] was applied
to the prototype frame with increasing amplitude; only a lateral force was applied at the top of the
one-story subassembly frame. Therefore, the analytical results revealed less lateral force in the
prototype frame than in the one-story subassembly frame, and the latter observation was considered
when designing gusset connection specimens. Table V lists forces in the BRB (PBRB) and
equivalent strut (Fgm) of the subassembly frame at 1% drift. The equivalent strut force calculated
using Equation (5) was close to that in the BRB, which indicated that the frame action force is
comparable to the brace force and should be considered when designing the gusset connection.
4. CORNER GUSSET PLATE SPECIMEN
Figure 1(c) shows the frame subassembly, which uses a single-diagonal BRB to dissipate seismic
energy. Five different corner gusset connections (Specimens 1–5) were used to connect the BRB
and framing members. Specimens 1 and 2 had a single-gusset-plate configuration and Specimens 3–5
had a dual-gusset-plate configuration. Free-edge stiffeners were used in Specimens 1 and 3. Web
stiffeners were used in the column-to-gusset and beam-to-gusset tips of all specimens to prevent
local buckling of the beam and column.
Specimen 1 used a 14-mm-thick gusset plate and two 12-mm-thick free-edge stiffeners (Figure 4
(a)). Both ends of the BRB stopped between the top and bottom corner gusset plates, and eight
splice plates with 25-mm-diameter A490 bolts were used to connect the BRB and gusset plate.
Specimen 2 was identical to Specimen 1 except that free-edge stiffeners were not used. Gusset size
was generally determined by either compression capacity or by von-Mises stress at gusset tips, so
the tensile capacity of the gusset, Rn, calculated based on the Whitmore method exceeded the tensile
capacity, Tmax, of the BRB (Table III). The ultimate compression capacity of the single gusset plate
calculated according to measured yield strength and the buckling coefficient k = 0.65 was 1368 kN
(Table IV), which exceeded that of the BRB (914 kN). In Specimen 2, which was a single gusset
plate without free-edge stiffeners, the ultimate compression capacity calculated using the buckling
coefficient of k = 2 was 759 kN, which was smaller than that of the BRB. This indicates that the
Table V. Forces and stresses in the corner gusset plate of the subassembly frame (1% drift).
(MPa)
s
Specimen Fgm (kN) PBRB (kN) Beam-to-gusset Colum-to-gusset Fy (MPa)
1 750 730 381 373 385

2 750 730 427 422 385
3 807 730 373 348 448
4 807 730 412 400 448
5 807 730 412 400 448
Vb =170 kN, Vc = 240 kN.
DOI: 10.1002/eqe
Figure 4. Gusset details.
single gusset plate without free-edge stiffeners would buckle before reaching the maximum
compression capacity of the BRB. However, if the compression capacity was calculated using a
buckling coefficient of k = 1.2 [11], compression capacity for Specimen 2 exceeded Cmax of the
BRB. Therefore, this specimen was used to test whether the AISC specification [11] for k is
appropriate.
Specimens 3 to 5 used a pair of 8-mm-thick gusset plates to connect the BRB and framing members.
Specimen 3 was identical to Specimen 4 except that 12-mm thick free-edge stiffeners were only added
in Specimen 3 (Figure 4(b)). Specimen 5 was identical to Specimen 4 except that the BRB was turned
90 degrees such that the weak axis of the core plate was transverse to the loading plane with the dual
gusset plates spaced 170 mm apart (Figure 4(c)). The BRB ends in Specimens 3 and 4 were inserted
into dual gusset plates; longitudinal fillet welds and 25-mm-diameter bolts were used to connect the
dual gusset plates and BRB. In Specimen 5, only fillet welds were used to connect the dual gusset
plates and BRB. The moment of inertia was significantly larger in the dual-gusset-plate
configuration than in a single-gusset-plate configuration of similar size, and the compression
capacity of 5-mm-thick dual gusset configuration (1566 kN and 1442 kN in Specimen 3 and 4
DOI: 10.1002/eqe
geometries, respectively) exceeded that of the BRB (914 kN). However, 8-mm-thick dual gusset plates
were used in Specimens 3–5 to limit maximum von-Mises stresses at gusset tips below yield strength
Fy (Tables II and V). Except for Specimen 2, the von-Mises stresses at the gusset tips were smaller than
gusset yield strength Fy obtained from the coupon test. The critical unbraced length of the unit strip, Lc,
was negative in Specimen 5 details because of a negative value of Lc1 (Figure 4(c)), indicating that
buckling of the dual gusset plates could be excluded. In Specimens 2, 4, and 5, which lacked free-
edge stiffeners, the effects of single or dual gusset plates on the BRB connection were evaluated.
5. BRBF SUBASSEMBLY TEST
The frame subassembly was subjected to prescribed loading according to Section T6 of the AISC
seismic provisions [10] until specimen failure or until 2% drift was reached. The MRF was also
tested to investigate its seismic performance after removing the BRB and corner gusset plates.
Comparing the cyclic responses of the BRBF and MRF subassemblies evaluated the performance of
the BRB with the corner gusset connection.
5.1. Test results

The BRBF subassembly had stable energy dissipation throughout the test except for Specimen 2
(Figure 5). The BRB generally yielded at an interstory drift of 0.5%; the column and beam yielded
at an interstory drift of 1%. Recall that the design intent, informed by nonlinear static pushover
analysis of the three-story BRBF (Figure 1(d)), was that the first-story BRB yielded at a
considerably low drift of 0.55% and there would be no yielding of the first-story frame members up
to an interstory drift of 1.1%. The results indicated that this requirement was satisfied as the one-
Figure 5. Lateral force versus story drift responses
DOI: 10.1002/eqe
story BRBF subassembly test. In Specimen 1, the beam-to-gusset and column-to-gusset edge tips
yielded at an interstory drift of 1.5%. Meanwhile, the restraining members slid down to the stiffened
core segment, which was not observed in the component tests [15]. At an interstory drift of 2%,
neither the gusset nor the BRB buckled (Figure 6(a)); however, in the moment-resisting beam-to-
column connection near the actuator, the beam web top access hole fractured in negative loading
direction (Figure 6(b)). A top cover plate and two side plates were utilized to rehabilitate the
moment-resisting connection so that the frame subassembly could be reused in further tests. The
moment-resisting connections rehabilitated with side plates within beam top and bottom flanges
were verified to sustain cyclic loading tests up to 4% drift without strength degradation [21, 22].
Figure 6. Observed performance in BRBF subassembly test.
DOI: 10.1002/eqe
Free-edge stiffeners were removed from the gusset plate after completing the Specimen 1 test. The
BRBF subassembly was then tested with the same BRB and corner gusset plate. The top gusset plate
buckled laterally at an interstory drift of 0.63% when the BRB was in compression. The bottom
gusset plate did not buckle, rotating the BRB about the bottom gusset plate. The out-of-plane
deformation for the BRB and gusset plates resembled the first-buckling mode shape of the BRBF
obtained from finite element analysis [9]. This out-of-plane deformation shape was not the same as
that for BRBs with a pinned gusset connection, which allows BRB end rotation [23]. The BRBF
maintained peak strength when the BRBF was displaced in the reverse direction to an interstory
drift of 1% (Figure 5(b)). The out-of-plane deformation of the single gusset plate was significant at
an interstory drift of 1% (Figure 6(c)). Because the peak lateral force at this drift level (1%)
decreased from 1220 kN (Specimen 1) to 980 kN (Specimen 2), the test was stopped. Two
restraining members were removed from the BRB to inspect the core; the core revealed no fractures,
and only one end of the BRB revealed flexural bending. Two stoppers located in the middle of the
core unexpectedly fell off because of false fabrication by the manufacturer, which caused the
restraining members to slide during the test. After refabricating the other two BRB cores to correct
this defect, the BRBs were reassembled with original restraining members.
A new BRB with a dual-gusset-plate configuration and free-edge stiffeners (Specimen 3) was
installed in the frame subassembly. As in the Specimen 1 test, energy dissipation in the frame
subassembly was stable throughout the test (Figure 5(c)). Significant yielding was observed in dual
gusset plates near the beam-to-gusset and column-to-gusset edge tips (Figure 6(d)). After removing
free-edge stiffeners from the dual gusset plates, the frame was retested with the same BRB and dual
gusset plates (Specimen 4). Unlike Specimen 2, however, the dual gusset plates sustained maximum
lateral force of 1720 kN at interstory drift of 2% without buckling (Figures 6(e) and (f)), indicating
that, without free-edge stiffeners, the out-of-plane stability is much better in the dual-gusset-plate
configuration than in a single-gusset-plate configuration of similar size.
After removing the BRB and dual gusset plates, the MRF was tested to investigate its hysteretic
behavior. The rehabilitated moment connection performed well up to an interstory drift of 2%
(Figure 5(f)). The frame was then installed with a new BRB and Specimen 5 gusset details. The
frame subassembly also had stable energy dissipation up to an interstory drift of 2% (Figure 5(e)).
However, a crack occurred in the beam bottom flange near the junction between the gusset tip and
web stiffener (Figure 6(g)) while the frame subassembly was moved to an interstory drift of 2%
(frame opening). The crack was attributed to beam flange out-of-plane bending cased by frame
action. However, for Specimen 4 with dual gusset plates welded below the beam flange–web
junction, the beam flange out-of-plane bending was minimized so no crack was initiated at the same
drift level. Significant yielding was also observed near the bottom gusset tips, but no fractures were
noted (Figure 6(h)). Two additional cycles with 2.5% drift were conducted on this specimen frame.
Weld fractures were observed in the top and bottom column-to-gusset interfaces at an interstory drift
of 2.5% (Figures 6(i) and (j)). These cracks were initiated at the column-to-gusset tips and
propagated along the gusset weld when the frame subassembly opened as illustrated in Figure 3(b).
These cracks closed when the frame closed at an interstory drift of 2.5%. Strength was slightly
reduced in the second cycle because of beam buckling near the rehabilitated moment connection.
Figure 7 shows the axial force versus axial displacement of the BRB. Axial force in the BRB was
obtained by subtracting the lateral force of the MRF (Figure 5(f)) from that of the BRBF and then
multiplied by the inclined angle of the BRB. The axial displacement of the BRB was obtained by
measuring the relative displacement at both ends of the BRB. In Specimen 2, the gusset plate
buckled when the axial force in the BRB was 693 kN. This compression force was much smaller
than the compression capacity of the gusset plate (1157 kN) calculated based on the buckling
coefficient k = 1.2 and the true yield strength of the steel. However, the compression capacity of the
gusset plate calculated based on the buckling coefficient k = 2 was 759 kN, which was slightly larger
than that observed in the test. This indicates that k = 2 is better than k = 1.2 when using column strip
method to estimate the compression capacity of a gusset without free-edge stiffeners. Table VI
shows the ratios of maximum compressive force to maximum tensile force for the BRBF and BRB,
respectively; except for Specimens 1 and 2, the ratios were around 1.1 and 1.2, which were lower
than 1.3 specified by the AISC seismic provisions [10]. The ratio of Specimen 1 was larger than
DOI: 10.1002/eqe
Figure 7. BRB axial force versus axial displacement response.
Table VI. Force ratios in BRB and BRBF (2% drift).
Specimen 1 3 4 5
BRB 1.33 1.21 1.19 1.23
BRBF 1.17 1.11 1.11 1.12
Figure 8. Lateral force ratio between MRF and BRB.
DOI: 10.1002/eqe
others because of sliding of the restraining member at high drifts. Figure 8 shows the lateral force ratios
between the MRF and BRB. Although the BRB sustained stable energy dissipation up to an interstory
drift of 2%, the lateral force ratio of the BRB decreased with drift. The BRB provided approximately
70% resistance at low drift levels and approximately 50% resistance at high drift levels because of its
yielding.
5.2. Frame action

The equivalent strut force in the gusset was compared to the axial force in the BRB. In each test, shears
in the beam and column were determined based on rosette readings, the shear modulus of steel, and the
cross-sectional area of the beam and column, respectively. The force in the BRB, PBRB, was obtained
from Figure 7. The equivalent strut force, Fgm, computed using Equation (5) were of the same order of
the brace force before an interstory drift of 1% (Figure 9). Afterwards, the incremental change in force
on the BRB was smaller than that in the equivalent strut because most beam and column sections
remained in the elastic range, whereas the BRB remained within the yield plateau range. Likewise,
Figure 8 confirms that the lateral force ratio of the BRB decreased with drift. Compared to the BRB,
the frame exerted a larger force on the gusset at high drift levels, which caused the gusset edges to
fracture.
Figure 10 shows the strut deformation in the gusset caused by frame action. Displacement
transducers M1 and M8 were positioned diagonally in the top and bottom gussets (Figures 1(c) and
10(a)). As shown, the centerline of the BRB was aligned to intersect the centerline of the column at
the top of the column’s base plate, and a bottom gusset was welded to the column flange and base
plate. No flexural deformation was introduced in the base plate during the test. In general, the
bottom gusset strut deformation was similar in Specimens 1–4 (Figure 10(b)) and was smaller than
the top gusset strut deformation in Specimens 1–2 (Figure 10(c)). Beam flange bending caused by
frame action reduced the magnitude of strut deformation, especially in the top dual-gusset-plate
connection. In the positive loading direction, the top gusset plate, which was subjected to frame
close (Figure 3(b)), was shortened in the strut direction. Strut deformation in the top gusset plate of
Specimen 2 was larger than that of Specimen 1 (Figure 10(c)) because the BRBF in Specimen 2
lacked free-edge stiffeners. Moreover, strut deformation was greater in the single gusset plate
configuration (Specimen 1) than in the dual-gusset-plate configuration (Specimen 3) because of the
beam flange deformation in the dual-gusset-plate configuration. When the single gusset plate was
positioned right below the beam web, the strut force caused by frame action was transferred directly
to the beam web. Only gusset deformation was measured along the strut. However, when dual
Figure 9. Frame action force versus brace action force.
DOI: 10.1002/eqe
Figure 10. Strut and beam flange deformation.
gusset plates were placed away from the beam web, the beam flange was subjected to strut
compression force under the frame close, which bent the beam flange (Figure 10(d)) and reduced the
magnitude of compression deformation measured by the displacement transducer M8. The reduction
caused by beam flange deformation was significant in Specimens 4 and 5, which lacked free-edge
stiffeners. Because this did not occur in Specimens 1 and 2, the axial deformation in the strut
increased with drift.
5.3. Gusset configuration

Figure 11 shows normal and shear strain profiles along the beam-to-gusset interface. In the positive
loading direction, the gusset plate is subjected to tension force from the BRB and compression force
from frame close (Figure 3(b)). The gusset tip, which is mainly affected by frame action, shows
compression strain; the gusset center, which is mainly affected by brace action, shows tension strain.
The normal strain in the gusset tip and center was larger in Specimen 1 than in Specimen 3. Shear
strain was also larger in Specimen 1 than in Specimen 3. The single gusset configuration has a
relatively more direct load path because the beam web, gusset, and column web plates are all in the
coplane, and also because this configuration is unaffected by beam flange bending, which is
unavoidable in the dual gusset configuration. However, the dual-gusset-plate configuration provides
better out-of-plane stability than the single-gusset-plate configuration. Without utilizing free-edge
stiffeners, the single gusset plate (Specimen 2) buckled when the BRB was subjected to an axial
force of 693 kN. Dual gusset plates (Specimen 4) similar in size to a plate in single gusset sustained
an axial load exceeding 1000 kN without buckling (Figure 7(d)).
Figure 11. Strain profiles along gusset-beam interface (positive loading).
DOI: 10.1002/eqe
Figure 12. Out-of-plane deformation along the BRB length.
Figure 12 shows out-of-plane deformation at three points along the BRB axis: the bottom gusset plate
(point A), the brace midlength (point B), and the top beam–column-gusset connection (point C). As
expected, the out-of-plane deformation of the bottom gusset plate was small in Specimens 3–5. The
data were not measured in the Specimen 1 test, but the out-of-plane deformation should be small
because the bottom gusset plate was welded to the column’s base plate. At 2% interstory drift, where
gusset plate distortion and brace instability usually occurred in previous studies [5,7], large out-of-
plane deformations at points C and B were also observed in Specimen 1 using a single-gusset-plate
connection (Figure 12(a)). However, dual-gusset-plate connections (Specimens 3 and 5) significantly
reduced out-of-plane deformation of the top gusset plate and BRB at the same drift level (Figures 12
(b) and (d)), indicating larger drift to be achieved. Specimen 4 showed large out-of-plane deformation
at point B (Figure 12(c)) as Specimen 1 because the residual out-of-plane deformation of the BRB
after completing the first test (Specimen 3) expedited lateral deformation of the BRB in the second
test (Specimen 4).
5.4. Free-edge stiffeners

By comparing test results of Specimens 1 and 2, utilizing free-edge stiffeners significantly improved
stability and compression capacity of a single-gusset-plate connection. The improved behavior was
not observed in dual-gusset-plate connection tests (Specimens 3 and 4) because gusset instability
was already improved by using two instead of one plate to transfer BRB forces to the frame.
Moreover, Figure 13 shows strain profiles along the beam-to-gusset interface for Specimens 3 and 4.
Figure 13. Strain profiles along gusset-beam interface (negative loading).
DOI: 10.1002/eqe
Normal strains at the gusset tip did not vary remarkably in Specimen 3 but did increase with drift in
Specimen 4 (Figure 13(a)). Using strain gages that were placed in the free-edge stiffener near the
beam, tensile strains increased with drift, which is evident for force transfer through the gusset tip.
Shear strains in the gusset plate were similar in Specimens 3 and 4 (Figure 13(b)). This indicated
that free-edge stiffeners were effective in reducing tensile strains near the gusset tips and ineffective
in transferring shear forces from the brace to the beam.
6. SEISMIC RESPONSE OF THE THREE-STORY BRBF UNDER EARTHQUAKES
Nonlinear time-history analyses of the three-story BRBF were performed to determine seismic
demands and to verify the capacity of the BRB and its connection under tests. A cyclic pushover
analysis of the one-story frame subassembly (Figure 14(a)) was first performed using the PISA
computer program [24] to correlate test results. A bilinear inelastic model with a strain hardening
ratio of 5% was used to model the rotational hysteretic hinge at the end of the beam and column
sections. The isotropic plastic hardening model developed by Chou and Tsai [25] was used to model
the inelastic behavior of the BRB under uniaxial loading. The P-Delta effects because of gravity
loading were not considered in the model. The strength and stiffness degradation of flexural hinges
in the beam, column, and BRB were also not considered in this numerical model. Rigid elements
were used to represent the gussets and connect the BRB end to the centerlines of the beam and
column. The pushover curve as shown in Figure 1(d) indicates that the BRB was designed to yield
at a roof drift of 0.55%. Figures 14(b) and (c) show the lateral force versus drift for the BRBF and
the axial force versus axial displacement for the BRB. Because the BRB was modeled as a beam–
column element, the confining effects of restraining members could not be simulated. Therefore,
estimations of lateral force when the BRB was in compression had a larger discrepancy than that
when the BRB was in tension. The peak lateral force of the BRB was reasonably predicted by the
model, but the analytical error in predicting the hysteretic energy dissipation of the BRBF during
the test was large at high drift levels (Figure 14(d)). Although the analytical errors in predicting the
hysteretic energy dissipation of the BRBF under pseudodynamic tests were as high as 100%,
Figure 14. Comparison between the test and modeling of the subassembly frame.
DOI: 10.1002/eqe
the prediction on peak story shear and lateral displacement responses was within reasonable accuracy
in severe earthquakes [8].
The same modeling technique was adopted for the three-story prototype BRBF. A Rayleigh type
damping of 5% of critical was assigned in the first mode and the second mode. Nonlinear time-
history analyses of the prototype BRBF were performed using the PISA computer program. A time
step of 0.001 s was used for the analyses to eliminate numerical convergence problems. An
ensemble of 10 strong ground motions was chosen from earthquake records that were recorded in
California or Taiwan. These records were free of any forward directivity effects (near-field effects)
and were recorded for soil types C or D, generated by earthquakes of moment magnitude ranging
from 6.7 to 7.3. A 5% damped design earthquake response spectrum for buildings built on soil class
D was constructed and was used as the target spectrum [16]. Each record was scaled to minimize
the square of error between its 5% damped response spectrum and the target spectrum, which
represents the design spectrum for DBE level. Table VII lists the ground motion data and the scaled
PGA for each record to match the IBC response spectrum. A good match existed between mean
spectra values and the target spectrum although small deviations could be observed in the medium
period range (periods from 1 to 1.7 s), which was beyond the fundamental period of the frame
(about 0.69 s). The DBE ground motions were also scaled by a factor of 1.5 to investigate the
seismic demands of the frame under the MCE level.
6.1. Demands of the BRBF under DBE and MCE levels

Figures 15(a) and (b) present the distribution along the building height of the mean (m) and the mean
plus one standard deviation (m + s) values of the interstory drifts and the BRB strains under DBE and
MCE records. The means (m) for the maximum interstory drift in the BRBF under DBE and MCE
records were 1.1% and 1.8%, respectively, which satisfied the 2% design requirement. The largest
m + s value of the interstory drift for the first floor was about 2.5% under the MCE level (Figure 15(a)),
which corresponded with the seismic hazard level when the gusset edge tip fractured during the
frame test. Although the largest m + s value of the maximum strain in the BRB approached 2%
under the MCE level (Figure 15(b)), which exceeded the 1.7% obtained in the frame test, this value
was still within the 2.6% strain capacity of the sandwiched BRB obtained in component tests [15].
Table VII lists maximum roof drift for the BRBF under each ground motion. The displacement
amplification factor is defined as a maximum roof drift divided by a roof drift (0.55%) when the first
plastic hinge forms in the BRBF. The means (m) for the maximum displacement amplification factor
in the BRBF under DBE and MCE records were 1.9 and 3.2, respectively, which were smaller than
Cd value of 5 for moment-resisting beam-to-column connections in AISC seismic provisions [10].
The largest m + s value of the displacement amplification factor was about 4.2 under the MCE level,
indicating that the design intent for lateral displacement of the BRBF was satisfied.
Table VII. Earthquake records.

Scaled PGA (g) Maximum roof drift (%)
Distance Original
Earthquake Record Site (km) PGA (g) DBE MCE DBE MCE
Chi-Chi TCU 074 D 18 0.60 0.38 0.57 1.12 1.71

TCU 039 D 17 0.21 0.41 0.62 1.28 2.10
Landers JOSHUA 90 C 11 0.28 0.41 0.62 0.93 1.31
IND 090 D 56 0.11 0.40 0.60 1.18 2.96
Loma AND 360 C 21 0.24 0.50 0.75 0.93 1.31
Prieta STG 090 C 13 0.32 0.56 0.84 0.68 1.17
Northridge CAST 360 C 21 0.51 0.4 0.60 1.02 2.41
CNP 196 D 16 0.42 0.47 0.71 1.36 1.95
Superstition ICC 090 D 14 0.25 0.46 0.69 0.85 1.33
Hills WSM 180 D 13 0.21 0.38 0.57 1.16 1.17
DOI: 10.1002/eqe
Figure 15. Seismic demands in the three-story BRBF under DBE and MCE motions.
7. CONCLUSIONS
A three-story BRBF with a sandwiched BRB was designed according to AISC seismic provisions [10].
Two corner gusset schemes for connecting a BRB and frame members were designed and tested in this
study. The first scheme used a single-gusset-plate connection as seen in current practice. The second
scheme used a dual-gusset-plate connection so the BRB could be placed as close as possible to the
beam-to-column connection to minimize gusset dimension. A pushover analysis was performed to
determine frame and brace action forces in the corner gusset plate. A one-story full-scale BRBF
subassembly was tested to evaluate (1) the seismic performance of a sandwiched BRB installed in a
frame, (2) the effects of free-edge stiffeners and a dual-gusset-plate configuration on the BRB-
connection behavior, and (3) the frame and brace action forces in the corner gusset plate. Time-
history analyses of the three-story BRBF were performed to verify if the sandwiched BRB and
corner gusset configuration can achieve the desired performance under DBE and MCE levels. The
following conclusions are based on the experimental results and nonlinear-time history analyses:
1. A BRB connection designed based on the frame and brace action forces exhibits stable energy
dissipation up to an interstory drift of 2.5% with a maximum axial strain of 1.7% in the BRB.
2. Compared to a dual-gusset-plate configuration, a single-gusset-plate configuration provides more
direct transfer of forces because the beam web, gusset, and column web plates all exist in the
coplane. Additionally, a single-gusset-plate configuration does not accompany beam flange bend-
ing, which is unavoidable in dual gusset plates placed far from the beam flange–web junction.
However, a dual-gusset-plate configuration provides much better out-of-plane stability compared
to a single-gusset-plate configuration. In a single-gusset-plate configuration without free-edge
stiffeners (Specimen 2), the gusset plate buckles at a low interstory drift. This does not occur
in a dual-gusset-plate configuration (Specimen 4) of similar size. Thus, for the weak axis of
the BRB core within the loading plane, a dual-gusset-plate connection is preferred because gusset
instability and weld fractures do not occur at high drift levels as seen in a single-gusset-plate
configuration [7, 8]. For the weak axis of the BRB core transverse to the loading plane, a dual-
gusset-plate connection sustains interstory drift up to 2.5% without strength degradation and
experiences gusset weld fractures because of frame action. The gusset weld fractures might be
delayed to even higher drift levels if free-edge stiffeners were provided in the dual-gusset-plate
connection. Moreover, these gusset weld cracks close after cyclic tests and can be easily repaired
for future use.
3. When free-edge stiffeners are not adopted in gusset connections, a side-sway mode of gusset
buckling can occur and the coefficient k when estimating compression capacity should be 2 rather
DOI: 10.1002/eqe
than 1.2. In addition to stabilizing gusset connections, free-edge stiffeners can reduce normal
stresses at gusset tips and gusset strut deformation through frame action.
4. Although the corner gusset plate in the BRBF is subjected to both brace and frame action forces,
the latter is not considered in AISC specification [11]. In the corner gusset, forces caused by
frame action and brace action are comparable at low interstory drifts. At high interstory drifts,
however, the frame action force in excess of brace action force causes weld fractures at gusset
tips under frame opening.
5. Beam–column elements can be used in the model to predict the peak lateral force of the BRBF
subassembly and axial force of the BRB under cyclic loading tests. However, the analytical error
on the hysteretic energy dissipation of the BRBF in the test is large, especially at high drift levels.
Under the MCE level, the largest m + s value of the interstory drift for the first floor of the BRBF
is 2.5%, which corresponds to the seismic hazard level observed when the gusset edge welds
fractured in the test. The largest m + s value of the displacement amplification factor is about
4.2, less than 5 specified in AISC seismic provisions [10], indicating that the design intent for
lateral displacement of the BRBF is satisfied.
ACKNOWLEDGEMENTS
The authors would like to thank the National Science Council, Taiwan for financially supporting this
research under Contract No. NSC 98-2625-M-002-017. The authors would also like to thank Prof. K. C.
Tsai in providing valuable advice in the program.
REFERENCES
1. Uang C-M, Nakashima M. Steel buckling-restrained braced frames. In Earthquake Engineering from Engineering
Seismology to Performance-based Engineering, Chapter 16, Bozorgnia Y, Bertero VV (eds.). CRC Press LLC: Boca
Raton, FL, 2003; 16-1–16-37.
2. Merritt S, Uang C-M, Benzoni G. Subassemblage testing of CoreBrace buckling-restrained braces. Report No.
TR-2003/01, University of California, San Diego, 2003.
3. Prinz GS, Richards PW, Fremming S. Seismic response of buckling-restrained frames with beam splices. 14th world
Conference on Earthquake Engineering, Beijing, China, 2008.
4. Lopez WA, Gwie DS, Saunders M, Lauck TW. Lessons learned from large-scale tests of unbonded braced frame
subassemblage. Proceedings of the 71st Annual Convention, SEAOC, Sacramento, CA, 2002; 171–183.
5. Roeder CW, Lehman DE, Christopulos A. Seismic performance of special concentrically braced frames with
buckling restrained braces. Proceedings of the 8th U.S. National Conf. on Earthquake Engineering, Earthquake
Engineering Research Institute, Oakland CA, Paper No. 1503, 2006.
6. Kasai K, Ooki Y, Ito H, Motoyui S, Hikino T, Sato E. Full-scale tests of passively-controlled 5-story steel
building using E-defense shake table. International Conference for Behavior of Steel Structures in Seismic Area,
Pennsylvania, USA, 2009.
7. Aiken ID, Mahin SA, Uriz P. Large-scale testing of buckling-restrained braced frames. Proceedings of Japan Passive
Control Symposium, Tokyo Institute of Technology, Japan, 2002; 35–44.
8. Tsai K-C, Hsiao P-C, Wang K-J, Weng Y-T, Lin M-L, Lin K-C, Chen C-H, Lai J-W, Lin S-L. Pseudo-dynamic tests
of a full scale CFT/BRB frame-Part I: Specimen design, experiment and analysis. Earthquake Engineering and
Structural Dynamics 2008; 37:1081–1098.
9. Chou C-C, Chen P-J. Compressive behavior of central gusset plate connections for a buckling-restrained braced
frame. Journal of Constructional Steel Research 2009; 65(5):1138–1148.
10. AISC. Seismic provisions for structural steel buildings. American Institute of Steel Construction: Chicago, IL, 2005.
11. AISC. Specification for structural steel buildings. ANSI/AISC 360–05. American Institute of Steel Construction:
Chicago, IL, 2002.
12. Kaneko K, Kasai K, Motoyui S, Sueoka T, Azuma Y, Ooki Y. Analysis of beam-column-gusset components in
5-story value-added frame. 14th World Conference on Earthquake Engineering, Beijing, China, 2008.
13. Chou C-C, Liu J-H. Frame and brace actions in corner gusset plate connections of steel buckling-restrained braced
frames. 12th Taiwan-Korea-Japan Joint Seminar on Earthquake Engineering for Building Structures, Kaohsiung,
Taiwan, 2010.
14. Chou C-C, Liu J-H. Frame and brace action forces on steel corner gusset plate connections in buckling-restrained
braced frames. Earthquake Spectra 2011 (112410EQS197M, accepted for publication).
15. Chou C-C, Chen S-Y. Subassemblage tests and finite element analyses of sandwiched buckling-restrained braces.
Engineering Structures 2010; 32:2108–2121.
16. IBC. International Building Code. International Code Council: Falls Church, VA, 2000.
17. Whitmore RE. Experimental investigation of stresses in gusset plate. Bulletin No. 16, Engineering Experiment
Station, University of Tennessee, 1952.
18. Thornton WA. Bracing connections for heavy construction. Engineering Journal (AISC) 1984; 21(3):139–148.
DOI: 10.1002/eqe
19. Yu QS, Uang C-M, Gross J. Seismic rehabilitation design of steel moment connection with welded haunch. Journal
of Structural Engineering ASCE 2000; 126(1):69–78.
20. Lee CH. Seismic design of rib-reinforced steel moment connections based on equivalent strut model. Journal of
Structural Engineering ASCE 2002; 128(9):1121–1129.
21. Chou C-C, Tsai K-C, Wang Y-Y, Jao C-K. Seismic rehabilitation performance of steel side plate moment
connections. Earthquake Engineering and Structural Dynamics 2010; 39:23–44.
22. Chou C-C, Jao C-K. Seismic rehabilitation of welded steel beam-to-box column connections utilizing internal flange
stiffeners. Earthquake Spectra 2010; 26(4):927–950.
23. Zhao JX, Wu B, Ou JP. A novel type of angle steel buckling-restrained brace: cyclic behavior and failure mechanism.
Earthquake Engineering and Structural Dynamics 2010. DOI: 10.1002/eqe.1071.
24. Tsai K-C, Lin B-Z. Development of an object-oriented nonlinear static and dynamic 3D structural analysis program.
CEER/R92-04, Center for Earthquake Engineering Research, National Taiwan University, Taipei, Taiwan 2003.
25. Chou C-C, Tsai K-C. Plasticity-fiber model for steel triangular plate energy dissipating devices. Earthquake
Engineering and Structural Dynamics 2002; 31(9):1643–1655.
DOI: 10.1002/eqe

Chou Et Al 2012

Uploaded by

Copyright:

Available Formats

You might also like

Chou Et Al 2012

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chou Et Al 2012

Uploaded by

Copyright:

Available Formats

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2012; 41:1137–1156

Steel buckling-restrained braced frames with single and dual corner

Chung-Che Chou1,2,*,†, Jia-Hau Liu1 and Dinh-Hai Pham1

Received 6 February 2011; Revised 29 August 2011; Accepted 2 September 2011

Copyright © 2011 John Wiley & Sons, Ltd.

Figure 1. BRBF (unit: mm).

2. DESIGN OF A THREE-STORY BRBF

Figure 2. BRB details.

C max ¼ bΩh ΩAy Fy ; (2)

3. DESIGN OF CORNER GUSSET CONNECTION

Pcr ¼ ð0:658Þlc Fy be tg ; lc⩽1:5

Table I. BRB size.

1 112 14 3832 112 85 4.5 4.5 4 40 56 28,839,662

Table II. Material strength.

Figure 3. Frame and brace action in the gusset plate.

3.1. Frame action force

Table III. Tensile capacity of the gusset plate.

Table IV. Compressive capacity of the gusset plate.

Fgm ¼ A′ fC1 Vb þ C2 Vc g (5)

where bs (= Lg/2) is the effective strut width [20].

4. CORNER GUSSET PLATE SPECIMEN

1 750 730 381 373 385

Figure 4. Gusset details.

5. BRBF SUBASSEMBLY TEST

5.1. Test results

Figure 5. Lateral force versus story drift responses

Figure 6. Observed performance in BRBF subassembly test.

Figure 7. BRB axial force versus axial displacement response.

Table VI. Force ratios in BRB and BRBF (2% drift).

Figure 8. Lateral force ratio between MRF and BRB.

5.2. Frame action

Figure 9. Frame action force versus brace action force.

Figure 10. Strut and beam ﬂange deformation.

5.3. Gusset conﬁguration

Figure 11. Strain proﬁles along gusset-beam interface (positive loading).

Figure 12. Out-of-plane deformation along the BRB length.

5.4. Free-edge stiffeners

Figure 13. Strain proﬁles along gusset-beam interface (negative loading).

6. SEISMIC RESPONSE OF THE THREE-STORY BRBF UNDER EARTHQUAKES

6.1. Demands of the BRBF under DBE and MCE levels

Table VII. Earthquake records.

Chi-Chi TCU 074 D 18 0.60 0.38 0.57 1.12 1.71

You might also like