Component Testing, Seismic Evaluation

and Characterization of Buckling-Restrained Braces

Cameron J. Black, M.ASCE1; Nicos Makris, M.ASCE2; and Ian D. Aiken, M.ASCE3

Abstract: This paper reports on the results from a comprehensive component testing program on a type of buckling-restrained brace
known as the Unbonded Brace™. The experimental data are used to: 共1兲 verify the results of theoretical predictions on the structural
stability of the braces; 共2兲 validate the inelasitc capacity of the braces under severe earthquake demands; and 共3兲 calibrate a macroscopic
hysteretic model that is found to predict, with fidelity, the brace force–displacement behavior. The study concludes that the unbonded
brace is a reliable and practical alternative to conventional framing systems to enhance the earthquake resistance of new and existing
structures; capable of providing both the rigidity needed to satisfy structural drift limits, while delivering a substantial and repeatable
energy absorption capability.
DOI: 10.1061/共ASCE兲0733-9445共2004兲130:6共880兲
CE Database subject headings: Seismic tests; Buckling; Bracing; Experimental data.

Introduction provides lateral support to the core and prevents global buckling.
Many of the structural failures in buildings during strong earth- A thin layer of material along the steel core/concrete interface
quake shaking have indicated that sustainable strength and stable eliminates shear transfer during the elongation and contraction of
energy dissipation capability are most desirable to maintain inter- the steel core and also accommodates its lateral expansion when
story drifts and overall structural displacements within tolerable in compression. It is the ability of the steel core to contract and
levels. The idea of using yielding metallic devices within the elongate freely within the confining steel/concrete-tube assembly
skeleton of a building frame to absorb large portions of the seis- that lead to the name unbonded brace.
mic energy is more than 30 years old 共Skinner et al. 1975; Kelly An early attempt to create a brace that dissipates energy yet
et al. 1992兲. This early work in the area of seismic protection of does not buckle is reported by Kimura et al. 共1976兲. The brace
structures considered torsional beams, flexural beams and other consisted of a conventional brace encased in a square steel pipe
structural mechanisms as the basis for energy dissipation devices.
Subsequently, a wide variety of efficient devices, that consist of
an array of mild steel plates, in triangular or hourglass shapes,
have been proposed 共Tyler 1978; Bergman and Goel 1987; Tsai
et al. 1993; Whittaker et al. 1993兲. The motivation for plate
shapes with variable triangular or hourglass sections is to achieve
uniformly distributed yielding throughout the material of the
plate, and to avoid localized deformations that will fail the de-
vices rapidly due to low cycle fatigue.
Another structural steel element that offers strength and energy
dissipation while at the same time exhibiting well-distributed
yielding is the buckling-restrained brace. One such buckling re-
strained brace is the Unbonded Brace™, manufactured by Nippon
Steel Corporation, Japan. The unbonded brace consists of a steel
core encased in a steel tube filled with concrete 共Fig. 1兲. The steel
core carries the axial load while the outer tube, via the concrete,

1
Graduate Research Assistant, Dept. of Civil and Environmental
Engineering, Univ. of California, Berkeley, CA 94720.
2
Professor, Dept. of Civil and Environmental Engineering, Univ. of
California, Berkeley, CA 94720.
3
Principal, Seismic Isolation Engineering, Inc., Oakland, CA 94611.
Note. Associate Editor: Andrei M. Reinhorn. Discussion open until
November 1, 2004. Separate discussions must be submitted for individual
papers. To extend the closing date by one month, a written request must
be filed with the ASCE Managing Editor. The manuscript for this paper
was submitted for review and possible publication on May 7, 2002; ap-
proved on December 15, 2003. This paper is part of the Journal of
Fig. 1. Schematic of unbonded brace 共a兲, core member 共b兲, and outer
Structural Engineering, Vol. 130, No. 6, June 1, 2004. ©ASCE, ISSN
tube 共c兲
0733-9445/2004/6-880– 894/$18.00.

880 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Table 1. List of U.S. Buildings Utilizing the Unbonded Brace 共as of May 2002兲
Type of construction and
Building, owner, and location building size Unbonded braces
Plant & Environmental Sciences Building New, Steel 132 braces, P y ⫽115– 550 kips
University of California, Davis, Calif. 3 stories⫹basement, 125,000 ft2 Core: JIS SM490A

Marin County Civic Center Hall of Justice Retrofit, RCa, 3– 6 stories 44 braces, P y ⫽400– 600 kips
County of Marin, Calif. 600,000 ft2 Core: JIS SN400B

Broad Center for the Biological Sciences New, Steel 84 braces, P y ⫽285– 660 kips
California Institute of Technology, Calif. 3 stories⫹basement, 118,000 gross ft2 Core: JIS SN490B

Hildebrand Hall Retrofit, RC 36 braces, P y ⫽200– 400 kips

University of California, Berkeley, Calif. 3 stories⫹basement, 138,000 ft2 Core: JIS SN400B

Wallace F. Bennett Federal Building Retrofit, RC 344 braces, P y ⫽205– 1,905 kips
Federal General Services Administration 8 stories, 300,000 ft2 Core: JIS SN490B
Salt Lake City, Utah

Building 5, HP Corvallis Campus Retrofit, Steel 60 braces, P y ⫽110– 130 kips

Hewlett-Packard, Corvallis, Ore. 2 stories, 160,000 ft2 Core: JIS LYP235

Centralized Dining & Student Services Building New, Steel 95 braces, P y ⫽210– 705 kips
University of California, Berkeley, Calif. 4 stories, 90,000 ft2 Core: JIS SN490B

King County Courthouse, Retrofit, RC 50 braces, P y ⫽200– 500 kips

King County, Seattle, Wash. 12 stories, 500,000 ft2 Core: JIS SN400B

Genome & Biomedical Sciences Building New, Steel 97 braces, P y ⫽150– 520 kips
University of California, Davis, Calif. 6 stories⫹basement, 211,000 ft2 Core: JIS SN400B

Physical Sciences Building New, Steel 74 braces, P y ⫽150– 500 kips

University of California at Santa Cruz, Calif. 5 stories, 136,500 net ft2 Core: JIS SN400B

Second Research Building 共Building 19B兲 New, Steel 132 braces, P y ⫽150– 675 kips
University of California, San Francisco, Calif. 5 stories, 171,000 ft2 Core: JIS SN400B

Kaiser Santa Clara Medical Center New, Steel 120 braces, P y ⫽265– 545 kips
Hospital Building Phase I, Kaiser Permanente 3 stories⫹basement, 266,000 ft2 Core: JIS SN400B
Santa Clara, Calif.
a
RC⫽reinforced concrete.

filled with mortar. While a few stable hysteretic characteristics retrofitted buildings in the U.S. that have implemented the un-
were reported, it was found that following a compressive loading bonded brace.
cycle, the transverse deformation of the mortar resulted in a per-
manent void space sufficiently large to allow local buckling dur-
ing subsequent compressive loading. Stability Analysis
Mochizuki et al. 共1980兲 conducted tests on similar braces
which were confined by reinforced concrete, with the concrete Three distinct buckling modes are identified in the stability analy-
prevented from adhering to the internal brace by use of a shock- sis of an unbonded brace: 共1兲 global flexural buckling of the entire
absorbing material. It was found however, that under repetitive brace; 共2兲 buckling of the inner core in higher modes; and 共3兲
loading, the concrete cracked and its buckling restraining effect plastic torsional buckling of the portion of the steel core which
diminished 共Wada et al. 1989兲. extends outside the confining tube.
This concept was further refined by a team of investigators in The critical load for the case of global flexural buckling is
Japan 共Watanabe et al. 1988; Wada et al. 1989; Watanabe and merely the Euler buckling load of the outer tube 共Wada et al.
Nakamura 1992兲 and ultimately resulted in what is known today 1989; Black et al. 2002兲
as the unbonded brace. The unbonded brace has been used exten-
sively in Japan with nearly 200 buildings currently utilizing the P cr P e ␲ 2E 0I 0
brace. The simplicity of its design coupled with its reliable per- ␴ cr⫽ ⫽ ⬇ (1)
Ai A i A i 共 KL 兲 2
formance has attracted the interest of structural designers in the
U.S. over the last few years. Table 1 presents a list of new and where A i ⫽cross-sectional area of the yielding portion of the

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 881

 brace; E 0 and I 0 ⫽respectively, elastic Young’s modulus and mo-
ment of inertia of the outer tube; and KL⫽effective buckling
length of the brace.
The critical load due to buckling of the inner core in higher
modes is given by

P cr 2 冑␤E t I i
␴ cr⫽ ⫽ (2)
Ai Ai

where ␤⫽distributed spring constant with dimension 关 F 兴 / 关 L 2 兴

representing the stiffness per unit length of the encasing mortar;
E t ⫽tangent elongation modulus; and I i ⫽moment of inertia of the
inner steel core. The result of Eq. 共2兲 is well known in the litera-
ture 共Wada et al. 1989兲 and is independent of the boundary con-
Fig. 2. Brace at maximum elongation showing unsupported length, l, ditions of the brace 共Black et al. 2002兲. It can be obtained by
of inner steel core. Dark region on steel core is from unbonding layer following an energy method 共Timoshenko and Gere 1961兲 or by
after repeated large deformation cycles. direct integration 共Black et al. 2002兲.
The third possible buckling mode of an unbonded brace is the
torsional buckling of the portion of the inner core that extends
beyond the confining tube. The critical stress under elastic tor-
sional buckling is given by 共Timoshenko and Gere 1961兲

Fig. 3. 共Top兲 Compressive branch from recorded force–displacement loop; 共center兲 reduction of secant and tangent moduli as compression
progresses; 共bottom兲 torsional buckling stresses computed with total deformation 关Eq. 共4兲兴 and incremental 关Eq. 共6兲兴 theories of plasticity. Results
by intuitive formulation proposed by Bleich are also shown.

882 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Table 2. Unbonded Braces Tested at Univ. California Berkeley in Spring 1999 and Fall 2000
Steel core Outer tube
Cross section Steel grade Steel grade
type and Yield and yield and yield P cr⫽ P e
dimension Area length stressa Py Cross section Length stress 共pinned endb兲
mm mm2 mm MPa kN mm mm MPa kN
Specimen 共in.兲 共in.2兲 共in.兲 共ksi兲 共kips兲 共in.兲 共in.兲 共ksi兲 共kips兲
U.C. Davis Plant and Environmental Sciences Building 共Spring 1999兲

99-1 共⫺兲 2,907 3,090 JIS SM490A 1,217 250⫻250⫻6 3,390 JIS STKR400 5,666
19⫻153 共4.5兲 共121.7兲 418.5 共60.7兲 共273.2兲 共9.8⫻9.8⫻0.24兲 共133.5兲 317.2 共46兲 共1,273兲
共3/4⫻6兲
99-2 共⫺兲 3,876 2,990 JIS SM490A 1,622 250⫻250⫻6 3,390 JIS STKR400 5,666
19⫻204 共6.0兲 共117.7兲 418.5 共60.7兲 共364.2兲 共9.8⫻9.8⫻0.24兲 共133.5兲 317.2 共46兲 共1,273兲
共3/4⫻8兲
99-3 共⫹兲 5,149 3,450 JIS SM490A 2,155 300⫻300⫻6 3,390 JIS STKR400 9,910
19⫻145 共8.0兲 共135.8兲 418.5 共60.7兲 共485.6兲 共9.8⫻9.8⫻0.24兲 共133.5兲 317.2 共46兲 共2,228兲
共3/4⫻5.71兲

Kaiser Santa Clara Medical Center 共Fall 2000兲

00-11 共⫹兲 7,125 3,410 JIS SN400B 2,033 300⫻300⫻6 3,340 JIS STKR400 9,910
19⫻197 共11.04兲 共134.3兲 285.4 共41.1兲 共453.7兲 共9.8⫻9.8⫻0.24兲 共131.5兲 317.2 共46兲 共2,228兲
共3/4⫻7-3/4兲
00-12 共⫹兲 7,125 3,410 JIS SN400B 2,033 300⫻300⫻6 3,340 JIS STKR400 9,910
19⫻197 共11.04兲 共134.3兲 285.4 共41.1兲 共453.7兲 共9.8⫻9.8⫻0.24兲 共131.5兲 317.2 共46兲 共2,228兲
共3/4⫻7-3/4兲
a
Value obtained from coupon test of steel core material.
b
Buckling length taken as total end-to-end length of steel core—4,500 mm for all specimens.

␴ cr⫽ 冋 ␲ 2E
12共 1⫺␯ 兲 l
2
b2
2
⫹G 册 t2
b 2
(3) ␴ cr⫽ 冋
Et ␲2 b2 Es t2
⫹
3 3 l2 Et b2 册 (4)

where E and G⫽elastic Young’s and shear moduli; and l, b, and Following Ilyshin’s and Stowell’s pioneering work, Bleich
t⫽length, width, and thickness, respectively, of each of the four 共1952兲, using engineering arguments, reached an expression simi-
flanges of the cruciform section. In the stability analysis each lar to Eq. 共4兲 which, for ␯⫽0.5, gives

冋 册
flange is considered as a uniform compressed plate simply sup-
ported along three sides and free along the fourth side. Fig. 2
shows a brace under large extension deformation at the onset of ␴ cr⫽
Et ␲2 b2
3 3 l2
⫹ 冑 E t2
Et b2
(5)
compression. The protruding length of the brace at that instant is
approximately 9 cm 共3.5 in.兲. For this specimen, b/t⫽5.0. As- The plastic buckling of a cruciform has been studied extensively
suming that the flange buckles elastically, the first term in the in the literature, due in part to the fact that calculations based on
brackets of Eq. 共3兲 represents the resistance of the flange to flex- the ‘‘less respectable’’ total deformation theory of plasticity cor-
ure while the second term represents the resistance of the flange related favorably with experimental data 共Gerard and Becker
in twisting. 1957兲; whereas, the calculations from the more sophisticated in-
For unbonded braces, the protruding steel core length, l, is cremental theory of plasticity departed appreciably from experi-
relatively short so that the steel core can support an axial com- mental results 共Batdorf 1949; Onat and Drucker 1952; Bijlaard
pressive stress as large as the yield stress. In the case of imminent 1956 among others兲. Several modern monographs and books
torsional buckling, a brace that has yielded in compression should 共Hutchinson 1974; Lubliner 1990; Bazant and Cedolin 1991兲 con-
be considered. In this case, the work done by the flexural mo- clude that the increments of shear stress and shear strain at the
ments along the flange plates will involve the tangent elongation onset of plastic buckling due to a uniaxial compression are related
modulus of the material, E t . As the flange plates buckle they will to the elastic shear modulus; therefore, the resistance in twisting
also twist and significant shearing action is mobilized. The shear- is unaffected by axially induced plasticity—an untenable conclu-
ing action of the flanges under plastic torsional buckling can be sion. This is indeed the result obtained by computing the tensor of
modeled with the total deformation theory of plasticity 共Ilyshin the tangent elasto–plastic moduli of a perfectly straight column
1947; Stowell 1948; Gerard and Becker 1957兲 which yields that under axial compression 共for example, Simo and Hughes 1998兲.
the shear modulus G in Eq. 共3兲 should be replaced with the secant The apparent shortcomings of the incremental theory of plasticity
shear modulus of the material G s ⫽E s /3. According to the total were addressed rigorously in a recent paper 共Papadopoulos and
deformation theory the critical axial stress under plastic torsional Lu 1998兲 by way of finite elasto–plasticity based on the theory of
buckling 共with ␯⫽0.5兲 is Green and Nagdi 共1965兲. The Papadopoulos and Lu finite element

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 883

Fig. 4. Euler buckling curves of specimens tested at Univ. California Berkeley. Axial yield load and maximum axial load on inner core upon
strain hardening are also shown.

calculations showed that in reality the incremental theory of plas- cycle reverses and compression initiates, the secant modulus E s
ticity yields results which are in good agreement with experimen- and the tangent modulus E t are initially the same as the elastic
tal data and resolved the ‘‘troublesome paradox’’ reported by Onat Young’s modulus of the steel E⫽210 MPa 共far right of the center
and Drucker. graph兲. As compression progresses, the steel core eventually
In a recent study 共Makris 2003兲, the plastic torsional buckling yields and the tangent modulus E t reduces more rapidly than the
of the cruciform was revisited by way of simple analysis and a secant modulus E s . This leads to a rapid reduction of the critical
small-strain theory. It was shown with hand calculations that for a stress. As the steel core continues to shorten and the protruding
slightly deflected flange, the incremental theory of plasticity length, l, reduces the critical stress increases again. Fig. 3共c兲
yields that the incremental shear stress and shear strain, at the shows that the incremental theory of plasticity 关Eq. 共6兲兴, predicts
onset of plastic torsional buckling due to a uniaxial compression, a lower capacity of the cruciform to sustain plastic torsional buck-
are related via the tangent shear modulus and therefore, the resis- ling than does the total deformation theory 关Eq. 共4兲兴. Fig. 3 shows
tance of the cruciform in twisting is indeed strongly affected by that the reduced capacity according to the incremental theory is
plasticity for even an infinitesimal perturbation of the straight close to the compressive stress demand that results by dividing
geometry. The critical stress obtained is the recorded compressive force by the area of the yielding

冋 册
section.
Et ␲2 b2 ␴y t2
␴ cr⫽ ⫹1⫹3 (6)
3 3 l2 Et b2
Component Testing of Unbonded Braces
where ␴ y ⫽yield stress of the steel core. The critical plastic buck-
ling stress given by Eq. 共6兲 was derived by examining the equi-
Description of Braces
librium of the flange in its deformed configuration—in the same
manner that one computes the elastic buckling strength given by The first tests of unbonded braces in the United States were con-
Eq. 共3兲. ducted at the Univ. of California, Berkeley during the spring of
Fig. 3共a兲 plots a segment of the force–displacement loop re- 1999 and fall of 2000. The geometrical characteristics and critical
corded during the testing of an unbonded brace. Fig. 3共c兲 plots the loads of the braces tested are summarized in Table 2 and the Euler
value of the critical stress ␴ cr as a function of the protruding buckling curves along with the capacities of the inner core to
length of the steel core cruciform section l. The three capacities resist plastic torsional buckling and axial yielding are shown in
shown are computed with the total deformation theory, the equa- Fig. 4. The three unbonded braces tested in spring 1999 共Speci-
tion proposed by Bleich, and the incremental theory of plasticity mens 99-1, 99-2, and 99-3兲 are representative of the braces de-
given by Eqs. 共4兲, 共5兲, and 共6兲, respectively. When the loading signed for the Univ. of California Davis Plant and Environmental

884 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Fig. 5. Top and side views of experimental setup 共a兲 and 共b兲, location of potentiometers 共c兲, and two possible configurations of inner core 共d兲
and 共e兲

Science Building. Two of the specimens had steel cores with rect- in.兲, and was displacement controlled via a linear variable dis-
angular cross sections and one specimen had a cruciform cross placement transducer mounted on the actuator. Figs. 5共a and b兲
section. All steel cores were Japanese Industrial Standard 共JIS兲 show a schematic diagram of the experimental setup used for both
Grade SM490A which is similar to American Society for Testing tests while Fig. 5共c兲 shows the location of the instrumentation
and Materials 共ASTM兲 A572/50 with a nominal yield stress of used to measure the various displacement quantities of interest.
315 MPa 共45.7 ksi兲. The tube steel was JIS grade STKR400, a Fig. 6 shows views of the experimental setup used for the fall
steel grade similar to ASTM 500. The unbonded braces tested in 2000 tests.
fall 2000 共Specimens 00-11 and 00-12兲 are representative of
braces designed for the Kaiser Santa Clara Medical Center. The
Testing Protocol
specimens tested had identical cruciform steel cores of JIS Grade
SN400B steel with 235 MPa⭐␴ y ⭐355 MPa (34.1 ksi⭐␴ y The test programs consisted of two phases. First, each brace was
⭐51.5 ksi). The tube steel was JIS STKR400. A cross section of subjected to a standard loading protocol that was agreed on by the
the test braces is included in Fig. 4. design engineers of the Univ. California Davis Plant and Environ-
The design of the bolted connection for the test specimens was mental Services Building for the spring 1999 tests, and the Calif.
identical to the braces in the actual construction. The bolts were 1 Office of Statewide Health, Planning and Development 共OSHPD兲
in. diameter, ASTM Grade A490, and the splice plates were JIS for the fall 2000 tests. These tests were followed by additional
Grade SM490A. The connections were slip-critical and thus it tests which included large-deformation, low-cycle fatigue tests
was important to ensure that correct bolt tightening force was and simulated earthquake displacement tests. Due to space limi-
achieved. tations, this paper focuses on the fall 2000 tests. Results from the
spring 1999 tests can be found in Black et al. 共2002兲.
The two specimens denoted 00-11 and 00-12 tested in the fall
Experimental Setup
of 2000 were subjected to a loading program consisting of in-
The experimental setup comprised two reaction frames anchored creasing displacement amplitude elastic and postyield cycles. This
to the laboratory strong floor. The test specimens were loaded via brace loading history, referred to as the OSHPD basic loading
a 3,100 kN 共⬃700 kip兲 hydraulic actuator with an in-line load history, was designed to subject the test specimens to maximum
cell. The actuator had a displacement capacity of ⫾15.3 cm 共⫾6 deformations and cumulative plastic deformations that exceeded

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 885

 displacement measured across the yielding portion of the brace
共the displacements recorded by potentiometers 7, 8, 9, and 10
were added and divided by 2兲. It is seen that the brace exhibited
stable hysteretic behavior for all displacement amplitudes. Fig. 7
共bottom兲 plots the recorded force versus the relative displacement
of the inner core to the outer tube measured at each end of the
tube separately. The loops from the displacement at each end are
nearly identical indicating that yielding happens uniformly
throughout the member. The maximum brace force was 3,005 kN
共675.6 kips兲 in tension and 3,400 kN 共764.3 kips兲 in compression,
a difference of 13%.
The force–displacement loops resulting from the low-cycle
fatigue test performed on Specimen 00-11 are shown in Fig. 8
共center and bottom兲. The brace exhibited stable hysteretic behav-
ior for the entire test consisting of 31 cycles. The initial intention
was to conduct the test to failure which was anticipated to occur
at roughly 20 cycles. This value was exceeded and subsequently
the test was stopped at 31 cycles 共without failure兲 in order to
avoid potential damage to the instrumentation or test setup.
Figs. 9 and 10 plot the results of the DBE and UBE displace-
ment loading tests, respectively. The hysteresis loops in the center
of each figure plot the force against the displacement measured
across the yielding section of the brace; whereas the hysteretic
loops at the bottom of the figures plot the force against the dis-
placement measured across the entire brace including the connec-
tions. During the DBE, the maximum brace displacement was
37.8 mm 共⫺1.49 in.兲 corresponding to a strain of 1.11%, while
the maximum brace displacement for the UBE was 47.2 mm
共⫺1.86 in.兲, corresponding to a maximum brace strain of 1.38%.
It is seen that the brace exhibits stable behavior throughout the
earthquake displacement histories. The flexibility of the connec-
tions contributes only slightly to the overall displacement, and its
effect can be seen by comparing the middle and bottom plots of
Figs. 9 and 10.

Fig. 6. Views of unbonded brace mounted on fall 2000 test setup at Comparison of Measured and Computed Stiffness
Richmond Field Station of the Univ. of California, Berkeley Values
In this section the experimental data is used to identify the initial
the largest demands anticipated for the earthquakes used in the elastic stiffness and the secondary stiffness of the brace, which
design. The OSHPD basic loading history is shown in Fig. 7 are of interest for the mechanical characterization of the brace.
共top兲. The parameter used to define the test program was the brace The contributions of the connection portions and the yielding por-
axial strain, defined at levels corresponding to building interstory tion of the brace to the overall stiffness is also examined.
drift ratios of 0.5, 1.0, 1.5, 2.0, and 2.25%. The acceptance crite- The measured stiffness values are extracted from the elastic
ria that had been established to evaluate the adequacy of the brace part of the recorded loops with the relation
behavior are given in Black et al. 共2002兲. 2F
In addition to the OSHPD basic loading history, one specimen K ie ⫽ (7)
U7⫹U8⫹U9⫹U10
was subjected to a low-cycle fatigue test with an amplitude cor-
responding to 1% story drift, while the other specimen was sub- where F⬍␴ y A for the yielding portion of the brace and
jected to two earthquake-induced displacement histories. The his- 2F
tories were derived from nonlinear, dynamic analyses performed e
K total⫽ (8)
U3⫹U4
as part of the project design and were representative of the most
severe brace displacement histories for the 10% in 50 year, design where F⬍␴ y A for the entire brace; while the superscript e indi-
basis earthquake 共DBE兲 and the 10% in 100 year, upper bound cates that they are experimental values. The values U3, U4, U7,
earthquake 共UBE兲 analyses. Figs. 9 共top兲 and 10 共top兲 plot the U8, U9, and U10 are displacements measured by potentiometers
brace displacement histories corresponding to the DBE and UBE 3, 4, 7, 8, 9 and 10 shown in Fig. 5, while F is the measured force
loading, respectively. applied to the brace.
Eqs. 共7兲 and 共8兲 yield that the measured elastic stiffness for the
yielding length is approximately 414.5 kN/mm 共2,367 kip/in.兲,
Test Results
whereas, over the entire length it is 362.9 kN/mm 共2,072 kip/in.兲.
Fig. 7 plots the OSHPD brace loading history imposed on Speci- Fig. 5 shows simple schematic diagrams of the steel core of
men 00-11 and the recorded force-displacement loops. Fig. 7 unbonded braces tested. The steel core shown in Fig. 5共d兲 is rep-
共center兲 plots the recorded force in the brace versus the total resentative of Specimens 99-1, 00-11, and 00-12, which consists

886 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Fig. 7. 共Top兲 Office of Statewide Health, Planning, and Development basic loading history; 共center兲 recorded force versus overall displacement
of yielding portion; and 共bottom兲 recorded force versus displacements of steel core relative to each end of tube

of a cruciform cross section throughout, with the yielding portion A conE A conE 11,225.8 mm2 •210 GPa
of the brace having a cross section smaller than the cross section K ccon⫽ ⬇ ⫽
L con 共 L total⫺L i 兲 /2 共 4,500⫺3,410兲 /2
of the end connection region. Fig. 5共e兲 shows the inner core rep-
resentative of Specimens 99-1 and 99-2. The total elastic stiffness kN
of the brace is the in-series sum of the individual stiffinesses of ⫽4,326 (11)
mm
the above described segments. For the fall 2000 specimens the
With the above values, the total stiffness, K ctotal , given by Eq. 共9兲
stiffness is given by
is 365 kN/mm. This value, and that given by Eq. 共10兲 are in close
1 agreement with the values measured from the experimental data.

冉 冊
K total⫽ (9) Table 3 lists the measured stiffness, computed stiffness, and
1 1
⫹2 the difference between these values for each test specimen. The
Ki K con measured values given in columns 2 and 3 were computed by
fitting a straight line 共using a least squares approximation兲 to the
where K i ⫽EA i /L i ⫽elastic stiffness of the yielding portion and
largest elastic level deformations on each brace. Column 2, la-
K con⫽EA con /L con⫽stiffness of the connection portion.
beled K ei , corresponds to the slope between the force and the
The computed stiffness values for the yielding and connection
displacement measured for the yielding portion of the brace 关Eq.
portions of Specimens 00-11 and 00-12 are, respectively
共7兲兴; whereas, column 3 corresponds to the computed slope of the
A i E 7,125 mm2 •210 GPa kN force–displacement relationship for the entire brace 关Eq. 共8兲兴. The
K ic ⫽ ⬇ ⫽439 (10) computed stiffness values given in columns 4, 5, and 6 were
Li 3,410 mm mm
obtained from Eqs. 共10兲, 共11兲, and 共9兲, respectively.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 887

Fig. 8. 共Bottom兲 Low cycle fatigue test; 共center兲 recorded force is plotted against displacement measured across yielding portion; and 共bottom兲
recorded force is plotted versus displacement of inner core relative to each end of tube

Table 3 indicates that the flexibility of the connection portion or alternatively, the cumulative plastic ductility 共CPD兲. The CPD,
of the brace has an appreciable affect on the overall stiffness—a which is a normalized expression of the cumulative plastic defor-
17% decrease for the fall 2000 specimens. Ideally the stiffness mation, is defined by
values used in the design should account for both portions of the
p ⫺u p 兩
兩 u max min

兺i
brace. If the stiffness is calculated assuming that the cross- i i
CPD⫽ (12)
sectional area is equal to A i , and constant over the entire length, uy
the stiffness is underestimated by 9%.
p i and u p i ⫽maximum and minimum plastic displace-
where u max min

ments, respectively, during each visit i into the inelastic range and
Secondary Stiffness
u y ⫽yield displacement of the brace.
The secondary stiffness depends in general on the loading history For example, a single monotonic displacement that reaches a
and is expressed via the postyielding ratio, ␣⫽K 2 /K i ⫽E t /E. ductility of eight results in a CPD of seven. That is, the inelastic
For an average value of the tangent elongation modulus, E t brace deformation was seven times the yield displacement 共eight
⫽4.14 GPa 共Black et al. 2002兲 and with E⫽210 GPa, ␣⫽0.02. total minus one elastic兲. Furthermore, if the loading is reversed
Using the elastic stiffness value of the inner core given by Eq. and the displacement brought back to zero the CPD would be 13.
共10兲 and ␣⫽0.02 the average value of the postyielding stiffness is This is because when loading reverses the material contracts elas-
K 2 ⫽8.36 kN/mm. tically by two yield displacements. Accordingly the CPD involves
seven units of ductility during loading and six units of ductility
during unloading—a total of 13. The calculated CPD exhibited by
Cumulative Plastic Ductility Demands
the fall 2000 specimens under the OSHPD brace loading history
A measure used in practice to describe the plastic demand on a is shown in Table 4.
buckling-restrained brace element is the cumulative plastic strain An example of the calculation of CPD is given in Fig. 11. Fig.

888 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Fig. 9. 共Top兲 Design basic earthquake displacement loading history; 共center兲 recorded force versus overall displacement of yielding length; and
共bottom兲 recorded force versus displacement of entire brace and end connection assembly

11 共top兲 shows the loading history from test phase 1 共second line attachments, a formulation that can be directly incorporated in
in Table 4兲. The heavy line segments indicate those portions of structural analysis software. The calibration of the model param-
the history during which inelastic deformation occurs. The center eters is based on the experimental data presented above.
plot of Fig. 11 shows the successive visits to the inelastic range
on an idealized force–displacement loop, while the bottom plot Macroscopic Modeling of Hysteretic Behavior
depicts the increase of cumulative plastic ductility with time.
A suitable model to approximate the nonlinear hysteretic behavior
The CPD values for all tests are summarized in Table 5. The
of an unbonded brace is
calculation of the CPD was achieved with an algorithm that de-
tects local peaks and valleys in the displacement history and ex- P 共 t 兲 ⫽␣Ku 共 t 兲 ⫹ 共 1⫺␣ 兲 Ku y z 共 t 兲 (13)
amines whether or not the brace has experienced plastic straining
where u(t)⫽axial deformation of the brace; K⫽preyielding stiff-
共i.e., an inelastic visit兲. Both the CPD for each individual test and
ness; ␣⫽ratio of the post- to preyielding stiffness; u y ⫽yield dis-
the total CPD that a given brace experienced throughout the test-
placement; and z(t)⫽hysteretic dimensionless quantity governed
ing protocol are given. It is noted that the CPD of the brace
by the differential equation
during the testing protocol exceeds, by far, the CPD demand of
the UBE. u y ż 共 t 兲 ⫹␥ 兩 u̇ 共 t 兲 兩 z 共 t 兲 兩 z 共 t 兲 兩 n⫺1 ⫹␤u̇ 共 t 兲 兩 z 共 t 兲 兩 n ⫺u̇ 共 t 兲 ⫽0 (14)
In the above equation ␤, ␥, and n⫽dimensionless quantities that
Characterization of Mechanical Behavior control the shape of the hysteretic loop. The hysteretic model,
expressed by Eqs. 共13兲 and 共14兲, was originally proposed by Bouc
In this section the force–displacement relation of an unbonded 共1971兲 for n⫽1, subsequently extended by Wen 共1975, 1976兲,
brace is characterized at the macroscopic level via the Bouc–Wen and used in random vibration studies of inelastic systems. The
model. This approach is selected since the force, P(t), of the physical meaning of the parameters of the Bouc–Wen model is
brace is given as a function of the displacement and velocity of its shown in Fig. 12 together with the data from the first inelastic

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 889

Fig. 10. 共Top兲 Upper bound earthquake displacement loading history; 共center兲 recorded force is plotted versus displacement of yielding portion;
and 共bottom兲 recorded force versus overall displacement of entire brace and end connection assembly

cycle of testing for Specimen 00-12. When parameter n assumes When one is interested in modeling the hysteretic behavior of
large values 共say n⬎10) the transition from the elastic to the the yielding portion of the brace, the stiffness in the Bouc–Wen
postyielding regime is sharp and the Bouc–Wen model reason- model is taken as
ably models bilinear behavior. A recent study that elaborates on
the Bouc–Wen model and shows how it can be generalized for K⫽K i ⫽A i E/L i (15)
degrading structures has been presented by Sivaselvan and Rein- whereas in the event that one is interested in also incorporating
horn 共2000兲. the flexibility of the connections K⫽K total , given by Eq. 共9兲. The
The parameters K, ␣, and u y are uniquely determined from the yield displacement u y is estimated by
geometrical and material properties of the brace. The values of ␤
and ␥ are constrained to each other through the relation ␤⫹␥⫽1 ␴y
u y ⫽⑀ y L i ⫽ L (16)
which leaves two fitting parameters, ␤ and n 共or ␥ and n兲. E i

Table 3. Comparison of Measured and Computed Elastic Stiffness Values

Measured stiffness Computed stiffness
kN/mm 共kip/in.兲 kN/mm 共kip/in.兲 Difference
e c c
K ie K total K ci K con K total
Specimen Eq. 共7) Eq. 共8兲 Eq. 共10兲 Eq. 共11兲 Eq. 共9兲 Ki K total
00-11 415 共2,370兲 363 共2,072兲 439 共2,507兲 4,326 共24,702兲 365 共2,084兲 ⫺5.5% 0.5%
00-12 424 共2,421兲 363 共2,072兲 439 共2,507兲 4,326 共24,702兲 365 共2,084兲 ⫺3.6% 0.5%

890 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Table 4. Cumulative Plastic Ductility Calculations for Office of Table 5. Cumulative Plastic Ductility of Tests Performed at Univ.
Statewide Health, Planning, and Development Brace Loading History California Berkeley
Brace Cumulative Maximum Cumulative
No. of displacement Plastic plastic ductility plastic
cycles ␦ br Ductility ductility ductility Specimen Test description ␮ max ductility Total
Test phase n mm 共in.兲 ␮ ␮p Total 兺 n ␮ p
U.C. Davis Plant and Environmental Sciences Building 共Spring 1999兲
d 2 4.83 1.0 0.0 0.0
共0.19兲 99-1 SACa basic loading history 10.0 243.5 323.8
1 2 15.50 3.3 2.3 17.4 99-1 SAC near-field protocol 20.0 80.3
共0.61兲 99-2 SAC basic loading history 10.0 243.5 879
2 2 31.50 6.7 5.7 63.0 99.2 Low-cycle fatigue test 10.0 636.5
共1.24兲 99-3 SAC basic loading history 10.0 243.5
3 2 45.00 10.0 9.0 135.0 99-3 Displacement derived 2.35 8.6 278.7
共1.85兲 from Sylmar motion
4 2 62.74 13.3 12.3 233.4 99.3 Displacement derived 8.30 26.6
共2.47兲 from El Centro motion
5 2 70.61 15.0 14.0 345.4
共2.78兲 Kaiser Santa Clara Medical Center 共Fall 2000兲

00-11 Brace loading history 15.0 345.4 1,045

00-11 Low cycle fatigue test 6.7 699.6
00-12 Brace loading history 15.0 345.4
00-12 Displacement derived 3.49 54.3 537.7
from Design Basis
Earthquake
00-12 Displacement derived 7.26 138.0
from Upper Bound
Earthquake
a
SAC Joint Venture 共a partnership of: Structural Engineers Association of
California, Applied Technology Council, California Universities for Re-
search in Earthquake Engineering兲, ‘‘A Program to Reduce the Earth-
quake Hazards of Steel Moment Frame Structures,’’ funded by the Fed-
eral Emergency Management Agency.

and the postyield to preyield stiffness is given by

␣⫽K 2 /K i ⫽E t /E (17)
Figs. 9 and 10 共center兲 compare the prediction of the Bouc–Wen
model 共shown as a dotted line兲 with the recorded force–
displacement loops recorded during the DBE and the UBE, re-
spectively. Prior to each of these tests, the steel core had experi-
enced appreciable inelastic deformations and has therefore been
strain hardened. To account for this, the yield displacement used
in the Bouc–Wen model has been modified to
␴y
u y ⬇1.25 L (18)
E i
The 1.25 correction factor in Eq. 共18兲 can also be computed geo-
metrically using the bilinear idealization.
The parameters used in Figs. 9 and 10 are given in Table 6.
Fig. 12 compares the prediction of the Bouc–Wen model against
the data recorded from the first inelastic cycle. In this case, the
brace has not yet experienced any strain hardening and the 1.25
correction factor in Eq. 共18兲 is not needed in estimating the u y
parameter.

Fig. 11. 共Top兲 Simple loading history normalized with respect to Estimation of Structural Response
yield displacement 共two cycles at ␮⫽3.33兲. Thick segments indicate With the macroscopic constitutive equation given by Eq. 共13兲 one
inelastic brace deformation. 共Center兲 Force–displacement loop.
can proceed with the dynamic analysis of a structure containing
Successive excursions in inelastic range are counted with index i.
unbonded braces. Fig. 13共a兲 portrays a frame with elastic lateral
共Bottom兲 Increase of cumulative plastic ductility with time.
stiffness K 0 and damping constant C 0 , which supports a mass m.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 891

Fig. 12. 共Solid line兲 Recorded data from first yielding cycle for specimen 00-12; 共dotted line兲 prediction of Bouc–Wen model. Physical meaning
of parameters appearing in Bouc–Wen model is also shown.

The frame includes an unbonded brace at an angle ␪. The analysis

presented here investigates the effect of the yielding brace on a
linear structure. The analysis of a real structure should include the
inelastic behavior of the frame together with the effect of the end
moments on the brace. Fig. 13共b兲 plots the time history of array
No. 5 recorded during the 1979 El Centro earthquake. When the
elastic frame is subjected to a ground excitation, ü g (t), dynamic
equilibrium gives
mü 共 t 兲 ⫹C 0 u̇ 共 t 兲 ⫹K 0 u 共 t 兲 ⫹ P 共 t 兲 cos共 ␪ 兲 ⫽⫺mü g 共 t 兲 (19)
where u(t)⫽lateral displacement of the frame and P(t)⫽axial
force resulting from the inclined brace given by
P 共 t 兲 ⫽␣Ku 共 t 兲 cos共 ␪ 兲 ⫹ 共 1⫺␣ 兲 Ku y z (20)
In Eq. 共20兲, K is either K i or K total , and is given by Eqs. 共7兲 or 共8兲;
whereas u y , ␣, and z are given by Eqs. 共18兲, 共17兲, and 共14兲,
respectively.
With the introduction of the normalized yield force

Table 6. Parameters Used in Bouc–Wen Model

Parameter Symbol Value
Initial stiffness K⫽K i 414.5 kN/mm
共2,367 kips/in.兲
Ratio of secondary to initial ␣ 0.025
stiffness
␴y 6.04 mm
Yield displacement uy⫽1.25 L 共0.238 in.兲
E i
Bouc–Wen parameter ␤ 0.55 Fig. 13. 共Top兲 Schematic of single degree of freedom structure with
Bouc–Wen parameter 共1⫺␤兲 ␥ 0.45 an unbonded brace and 共bottom兲 accelertion time history of 1979 El
Bouc–Wen parameter n 1 Centro earthquake, array No. 5 record

892 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004

Fig. 14. Exact drift and base shear spectra for linear single degree of freedom structure with unbonded brace subjected to 1979 El Centro
earthquake, array No. 5 record: 共solid line兲 Bouc–Wen model with n⫽1; and 共dashed line兲 bilinear approximation n⫽20

F y Ku y The solid lines in the graphs of Fig. 14 are computed by set-

f⫽ ⫽ (21) ting the parameter n in the Bouc–Wen model equal to one. The
mg mg
dashed lines are computed by setting n⫽20, which gives a bilin-
Eq. 共20兲 becomes ear model 共i.e., sharp transition from the elastic to the postyield-
u ing regime兲. The good agreement between the solid and the
P 共 t 兲 ⫽␣ f mg cos共 ␪ 兲 ⫹ 共 1⫺␣ 兲 f mgz (22) dashed lines indicate that engineers can use with confidence the
uy
bilinear element, which is included in commercially available
Substitution of Eq. 共22兲 into Eq. 共19兲 gives structural software in order to represent the behavior of unbonded
braces.
u共 t 兲
ü 共 t 兲 ⫹2␨␻ 0 u̇ 共 t 兲 ⫹␻ 20 u 共 t 兲 ⫹␣ f g cos2 共 ␪ 兲
uy

⫹ 共 1⫺␣ 兲 f g cos共 ␪ 兲 z⫽⫺ü g 共 t 兲 (23) Conclusions

where ␻ 0 ⫽ 冑K/m, ␨⫽C 0 /(2m␻ 0 ), and z is described by Eq. This paper summarizes results from a two-phase experimental
共14兲. program on the axial behavior of unbonded braces. The test re-
Fig. 14 plots nonlinear displacement and base shear spectra for sults demonstrate the good performance of the brace under vari-
the response of the one-degree-of-freedom structure shown in Fig. ous loading histories specified by the SAC and OSHPD protocols.
13 as a function of the structural period (T 0 ⫽2␲/␻ 0 The experimental data were used to: 共1兲 verify the results of
⫽2␲ 冑m/K 0 ). The response was computed with a state-space theoretical predictions on the structural stability of the braces; 共2兲
formulation, where Eqs. 共14兲 and 共23兲 are integrated simulta- validate the inelastic capacity of the braces under severe earth-
neously 共Makris and Chang 2000兲. quake demands; and 共3兲 calibrate a macroscopic hysteretic model
Fig. 14 indicates that the brace ductility demand increases rap- that was found to predict, with fidelity, the brace force–
idly with the period of the unbraced structure, T 0 . Given that displacement behavior.
unbonded braces can achieve ductilities on the order of 15 or For the configuration of the tested braces, this study found that
higher, Fig. 14 indicates that an unbraced structure with period plastic torsional buckling of the inner core is the most critical
T 0 ⫽1.5 s needs a brace strong enough to deliver a yield force stability mode. It is concluded that if the yielding portion extends
F y ⫽ f mg that exceeds 20% of the weight of the structure. When outside of the confining tube, the flanges of the yielding portion
unbonded braces are used in building structures the satisfaction of should have a width to thickness ratio b/t⬍5. The study indicates
the drift requirements should be met by considering the simulta- that the incremental theory of plasticity predicts the correct criti-
neous yielding of the frame of the structure which will most likely cal load for plastic torsional buckling, provided that the flanges of
require an even stronger brace. the cruciform are assumed to be slightly bent.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004 / 893

 Results from a comprehensive experimental program Ilyushin, A. A. 共1947兲. ‘‘The elasto-plastic stability of plates.’’ Technical
demonstrate that unbonded braces deliver ductile, stable, and Note 1188, National Advisory Commission on Aeronautics.
repeatable hysteretic behavior. The plastic deformation capacity Kelly, J. M., Skinner, R. I., and Heine, A. J. 共1992兲. ‘‘Mechanism of
of the braces exceeds the specified requirements both in energy absorption in special devices for use in earthquake resistant
terms of ultimate deformation and in terms of cumulative plastic structures.’’ Bull. N.Z. Nat. Soc. Earthquake Eng., 5共3兲, 63– 88.
Kimura, K., Takeda, Y., Yoshioka, K., Furuya, N., and Takemoto, Y.
strain.
共1976兲. ‘‘An experimental study on braces encased in steel tube and
A Bouc–Wen model was proposed to approximate the macro-
mortar.’’ Proc., Annual Meeting of the Architectural Institute of Japan,
scopic behavior of the brace. All but two of the model parameters Japan 共in Japanese兲.
are determined from the geometrical and physical properties of Lubliner, J. 共1990兲. Plasticity theory, Macmillan, New York.
the brace. A single set of calibrated model parameters was found Makris, N. 共2003兲. ‘‘Plastic torsional buckling of cruciform compression
to satisfactorily predict the cyclic and transient behavior of the members.’’ J. Eng. Mech., 129共6兲, 689– 696.
brace when subjected to a variety of loading histories. The pro- Makris, N., and Chang, S. P. 共2000兲. ‘‘Effect of viscous, viscoplastic and
posed macroscopic model can be easily implemented to compute friction damping on the response of seismic isolated structures.’’
the dynamic response of structures equipped with unbonded Earthquake Eng. Struct. Dyn., 29, 85–107.
braces. Additional parametric studies indicated that a simple bi- Mochizuki, N., Murata, Y., Andou, N., and Takahashi, S. 共1980兲. ‘‘An
linear model captures satisfactorily the brace participation for experimental study on buckling of unbonded braces under centrally
structural design purposes. applied loads.’’ Proc., Annual Meeting of the Architectural Institute of
It is concluded that unbonded braces represent a reliable and Japan, Japan 共in Japanese兲.
practical alternative to conventional framing systems to enhance Onat, E. T., and Drucker, D. C. 共1952兲. ‘‘Inelastic instability and incre-
the earthquake resistance of existing and new structures; capable mental theories of plasticity.’’ J. Aeronaut. Sci., 20共3兲, 181–186.
Papadopoulos, P., and Lu, J. 共1998兲. ‘‘A general framework for the nu-
of providing both the rigidity needed to satisfy structural drift
merical solution of problems in finite elasto-plasticity.’’ Comput.
limits, while delivering a stable and substantial energy absorption
Methods Appl. Mech. Eng., 159, 1–18.
capability. Sivaselvan, M. V., and Reinhorn, A. M. 共2000兲. ‘‘Hysteretic models for
deteriorating inelastic structures.’’ J. Eng. Mech., 126共6兲, 633– 640.
Simo, J. C. and Hughes, T. J. R. 共1998兲. Computational Inelasticity, in-
Acknowledgments terdisciplinary applied mathematics, Vol. 7, Springer, New York.
Skinner, R. I., Kelly, M. J., and Heine, A. J. 共1975兲. ‘‘Hysteretic dampers
The brace test specimens and the financial support for their test- for earthquake-resistant structures.’’ Earthquake Eng. Struct. Dyn., 3,
ing was provided by Nippon Steel Corporation, Japan. The tech- 287–296.
nical assistance during testing from Mr. Don Clyde, Mr. Wes Stowell, E. Z. 共1948兲. ‘‘Unified theory of plastic buckling of columns and
Neighbor, and Mr. David MacLam is appreciated. plates.’’ Technical Note 1681, National Advisory Committee on Aero-
nautics.
Timoshenko, S. P., and Gere, J. M. 共1961兲. Theory of elastic stability,
References McGraw-Hill, New York.
Tsai, K. C., Chen, H. W., Hong, C. P., and Su, Y. F. 共1993兲. ‘‘Design of
steel triangular plate energy absorbers for seismic-resistant construc-
Batdorf, S. B. 共1949兲. ‘‘Theories of plastic buckling.’’ J. Aeronaut. Sci.,
tion.’’ Earthquake Spectra, 9共3兲, 505–528.
405– 408.
Bazant, Z. P., and Cedolin, L. 共1991兲. Stability of structures, Oxford Tyler, R. G. 共1978兲. ‘‘Tapered steel cantilever energy absorbers.’’ Bull.
University Press, Oxford, U.K. N.Z. Nat. Soc. Earthquake Eng., 11共4兲, 282–294.
Bergman, D. M., and Goel, S. C. 共1987兲. ‘‘Evaluation of cyclic testing of Wada, A., Saeki, E., Takeuch, T., and Watanabe, A. 共1989兲. ‘‘Develop-
steel-plate devices for added damping and stiffness.’’ Technical Rep. ment of unbonded brace.’’ Column Technical Publication No. 115
No. UMCE 87-10, Univ. of Michigan, Ann Arbor, Mich. 1989.12, Nippon Steel, Japan.
Bijlaard, P. P. 共1956兲. ‘‘Theory of plastic buckling of plates and applica- Watanabe, A., Hitomoi, Y., Saeki, E., Wada, A., and Fujimoto, M. 共1988兲.
tion to simply supported plates subjected to bending or eccentric com- ‘‘Properties of braces encased in buckling-restraining concrete and
pression in their plane.’’ J. Appl. Mech., 23共1兲, 27–34. steel tube.’’ Proc., 9th World Conf. on Earthquake Engineering, Vol.
Black, C. J., Makris, N., and Aiken, I. D. 共2002兲. ‘‘Component testing, IV, Tokyo-Kyoto, Japan, 719–724.
stability analysis and characterization of buckling restrained braces.’’ Watanabe, A., and Nakamura, H. 共1992兲. ‘‘Study on the behavior of
Rep. No. PEER 2002/08, Univ. of California, Berkeley, Calif. buildings using steel with low yield point.’’ Proc., 10th World Conf.
Bleich, F. 共1952兲. Buckling strength of metal structures, McGraw–Hill, on Earthquake Engineering, Balkema, Rotterdam, The Netherlands,
New York. 4465– 4468.
Bouc, R. 共1971兲. ‘‘Modèl mathématique d’hysteresis.’’ Acustica, 24, 16 – Wen, Y.-K. 共1975兲. ‘‘Approximate method for nonlinear random vibra-
25. tion.’’ J. Eng. Mech., 101共4兲, 389– 401.
Gerard, G., and Becker, H. 共1957兲. ‘‘Handbook on structural stability: Wen, Y.-K. 共1976兲. ‘‘Method for random vibration of hysteretic systems.’’
Part I—Buckling of flat plates.’’ Technical Note 3781, National Advi- J. Eng. Mech., 102共2兲, 249–263.
sory Committee on Aeronautics. Whittaker, A. S., Aiken, I. D., Bergman, D., Clark, P. W., Cohen, J., Kelly,
Green, A. E., and Naghdi, P. M. 共1965兲. ‘‘A general theory of an elastic- J. M., and Scholl, R. 共1993兲. ‘‘Code requirements for the design and
plastic continuum.’’ Arch. Ration. Mech. Anal., 18, 251–281. implementation of passive energy dissipation systems.’’ Proc., ATC
Hutchinson, J. W. 共1974兲. ‘‘Plastic buckling.’’ Advances in applied me- 17-1 Seminar on Seismic Isolation, Passive Energy Dissipation, and
chanics, C.-S. Yih, ed., Academic, New York. Active Control, Vol. 2, San Francisco, 497–508.

894 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JUNE 2004