Engineering Structures 153 (2017) 279–289

Contents lists available at ScienceDirect

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

Strong-axis and weak-axis buckling and local bulging of buckling-restrained T

braces with prismatic core plates
⁎
Wenjing Xu, Chris P. Pantelides
Department of Civil and Environmental Engineering, University of Utah, Salt Lake City, USA

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

Keywords: The performance of new type Buckling-Restrained Braces (BRB) is evaluated. The new type BRB is efficient since
Buckling-restrained brace the steel core is constructed with prismatic steel plates that are straight throughout its length. Connection plates
Connection plates attaching the BRB to gusset plates are welded perpendicular to the steel core plates; this configuration is dif-
Cyclic ferent from conventional BRBs in which the core plates also serve as the connection plates. Since the steel core
Steel
plates are prismatic, construction of new type BRBs saves material and reduces the cost of manufacturing. Four
Strong axis buckling
Weak axis buckling
full-scale new type BRBs were tested under quasi-static cyclic loading with either a single core plate or dual core
Bulging plates using either bolted or welded connections to gusset plates. The hysteresis curves exhibited stable behavior
with positive incremental stiffness; the ratio of maximum compression to tension capacity and the cumulative
inelastic deformation were satisfactory. There was no rupture or instability and at failure the core plates
achieved maximum tensile strains between 3.2% and 4.2%. Failure modes included core plate buckling about the
strong or the weak axis and local bulging. The strong and weak axis buckling behavior is modeled using a strut-
and-tie model. The local bulging behavior about either the strong or the weak axis is modeled using plastic
analysis. Recommendations are presented for preventing strong axis buckling of the steel core and local bulging
of the steel casing about the strong axis.

1. Introduction
Buckling Restrained Braces (BRBs) are used in buildings to dis-
sipate energy imparted to the structure from large earthquakes. BRB
implementation has expanded since the 1994 Northridge earthquake
and they are currently becoming popular in damage control designs
[1,2]. BRBs consist of a steel casing housing one or more steel core
plates, surrounded by concrete; a gap exists between the concrete
and core plates that acts as a bond-breaker. BRBs are designed to
sustain yielding in both tension and compression. A significant
amount of research has been carried out regarding the experimental
performance and design of BRBs with various cross-sections [1–19].
BRB cross-sectional configurations include: rectangular core plate
restrained by a round or rectangular steel casing filled with concrete
[10–12]; rectangular core plate restrained by two steel tubes for self-
centering [13]; cylinder-shaped core plate in a round or rectangular
steel casing [14]; H-shaped core plate restrained by a steel casing
[15]; rectangular core plate restrained by metal plates bolted to-
gether as the casing [16]; H-shaped steel sections [17]; and cruci-
form core plates restrained by concrete and steel casing [3,18,19]. As
the steel core plate starts buckling about either the strong or the
weak axis it is restrained by the concrete inside the steel casing;
under certain conditions, the contact force between the steel core
and concrete could cause bulging of the casing. Research regarding
the contact force between the steel core and casing has been per-
formed for conventional BRBs [20–22]. In addition, research has
been performed regarding the local buckling restraint condition and
local bulging of the steel casing [20,22–24].
In conventional BRBs the connection plates and steel core plate
are manufactured from one single steel plate; the core plate is shaped
by reducing the cross-section width for a certain critical length that
is intended to yield, as shown in Fig. 1(a). Manufacturing of the core
plate in conventional BRBs requires skilled labor and special ma-
chining to shape the connection plate at the ends. In the present
investigation, BRBs with a new configuration of steel core plates and
separate connection plates were developed to reduce manufacturing
and material costs. In the new type BRB, connection plates are
welded perpendicular to one or more prismatic steel core plates, as
shown in Fig. 1(b). There is no extra cutting of the core plate; con-
struction of new type BRBs is much easier than creating the radius for
the copes of conventional BRBs. An application of the new type BRBs
for a new building constructed in a high seismic region is shown in

⁎
Corresponding author.
E-mail addresses: wenjing.xu@utah.edu (W. Xu), c.pantelides@utah.edu (C.P. Pantelides).

http://dx.doi.org/10.1016/j.engstruct.2017.10.017
Received 12 July 2016; Received in revised form 28 September 2017; Accepted 9 October 2017
Available online 21 October 2017
0141-0296/ © 2017 Elsevier Ltd. All rights reserved.
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

Nomenclature Pysc axial yield strength of steel core

Ry ratio of expected yield stress to specified minimum yield
Ac cross-sectional area of a single core plate or dual core stress
plates s air gap thickness between concrete and steel core
cw concrete cover thickness of core plates in the weak axis tc thickness of core plates
cs concrete cover thickness of core plates in the strong axis ts thickness of steel casing
DCR demand-to-capacity ratio Tmax maximum tensile force in loading cycle
E cumulative energy dissipation wc width of core plates
Ej + 1 energy dissipation for each loading increment weff width of effective inner surface
fys yield strength of steel casing weff − w width and length of inner effective area for weak axis
Fp,max bulging force buckling
Fs maximum capacity of steel casing against bulging weff − s width and length of inner effective area for strong axis
i loading cycle number buckling
lw wavelength of core plates ws width and thickness of steel casing with square cross-sec-
leff length of effective inner area tion
leff − w length of inner effective area for weak axis buckling x j and x j + 1 BRB displacements at time j and (j + 1) , respectively
leff − s length of inner effective area for strong axis buckling β compression strength adjustment factor
l w−w wavelength of core plates for weak axis buckling Δby yield displacement of brace
l w−s wavelength of core plates for strong axis buckling Δmax − com maximum displacement of brace in compression
Lc-c critical length of core plates Δmax − ten maximum displacement of brace in tension
Ltot total length of BRB η cumulative inelastic deformation
Mps plastic moment capacity of equivalent beam θ angle between strut and tie
n total number of cycles μi inelastic deformation
Pj and Pj + 1 BRB forces at time j and (j + 1) , respectively ω strain hardening adjustment factor
Pmax maximum compressive force on BRB

Fig. 1. Conventional and new type BRBs: (a) conventional

BRB; (b) new type BRB; (c) application of new type BRBs.

Fig. 1(c). In addition, the new type BRBs have two connection plates
at each BRB end, which makes pinned, bolted, and welded connec-
tions easier to construct.
Seismic provisions for structural steel buildings stipulate criteria
that BRBs must satisfy for acceptable performance such as compression
strength and strain hardening adjustment factors, cumulative inelastic
deformation, and cumulative energy dissipation [25]. This paper
focuses on describing the details of new type BRBs and their perfor-
mance in full-scale experiments. Steel core plate buckling and casing
bulging about the strong or the weak axis observed during the tests is
described and investigated. Recommendations are provided for pre-
venting strong axis buckling as well as local bulging of the steel casing
about the strong axis.

280
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

Table 1
Connection details and specimen geometry.

Test End connection wc (mm) Core type Ac (mm2) Lc-c (mm) Ltot (mm) Casing (mm × mm × mm)

1 Bolted 235 Dual 11935 3124 5563 305 × 305 × 6.4

2 Bolted 159 Dual 8065 3429 5309 305 × 305 × 6.4
3 Welded 254 Single 6452 3886 5563 305 × 305 × 6.4
4 Bolted 197 Single 5000 3988 5563 254 × 254 × 6.4
Raddon et al. [11] Pinned 161 Triple 12258 3662 4598 HSS 324 × 9.5

2. Experimental investigation

Four BRBs with details similar to those shown in Fig. 1(b) were
tested in full-scale experiments. In addition to determining the overall
performance of BRBs this research focuses on buckling of the core;
therefore, a wide range of sizes of core plates were used to build the
specimens. The characteristics of the specimens are summarized in
Table 1; a side view of a typical BRB is shown in Fig. 2; wc is the width
of the core plates, Ac denotes the cross-sectional area of a single core
plate or dual core plates, Lc-c is the critical length of the core plates, and
Ltot is the total length of the BRB. All BRBs had a square cross-section
with core plates ranging in width from 197 to 254 mm. Two of the BRBs
had dual core plates and two had a single core plate; the thickness of a Fig. 4. Load frame details of connection tie-down box for applying tension.
single core plate was 25 mm and that of the dual core composed of two
steel plates was 50 mm. The material for the core plates was Grade A36
steel; the properties of all core plates were the same: the yield stress was shown in Fig. 4. A steel plate was welded perpendicular to the bottom
282 MPa and the ultimate strength was 412 MPa. Three BRBs had a gusset plate and was mounted in the horizontal direction to form the
length of 5.56 m and the fourth was 5.31 m long. The outside cross- base of the bottom gusset plate. Two steel plates, parallel to the bottom
sectional dimensions of the casing were either 305 × 305 mm or gusset plate, were bolted to the east and west walls of a tie-down steel
254 × 254 mm; the steel casing was made with Grade A500; the casing box which was fixed to the ground; once the bolts were tightened, the
thickness was 6.4 mm for all specimens with a yield stress of 399 MPa bottom gusset plate was fixed and tension could be applied to the BRB.
and ultimate strength of 455 MPa. The connection plates were made The end connection condition for each of the four specimens is listed in
with A572 Grade 50 steel with a yield stress of 393 MPa and ultimate Table 1; three of the BRBs were bolted and one was welded to the gusset
strength of 483 MPa. plates. The BRBs were tested under axial quasi-static cyclic load using
Each BRB was assembled in the load frame vertically, as shown in displacement control. The loading protocol consisted of two cycles
Fig. 3. The top gusset plate, made with the same grade steel as the at ± 100% of the yield displacement, followed by a displacement cor-
connection plates, was attached to a hydraulic actuator with an inline responding to ± 0.5% interstory drift, and increasing interstory drifts
load cell. Tension was applied using a connection tie-down box as of ± 0.5% until failure.

3. Evaluation of experimental performance of new type BRBs

Typical criteria for adequate BRB performance include stability of

Fig. 2. Side sectional view of new type BRB.
the hysteresis loops, and satisfactory values of the compression strength
and strain hardening adjustment factors, cumulative inelastic de-
formation and hysteretic energy dissipation [25]. The results presented
in this section are based on the relative displacement at the two ends of
the BRB measured using Linear Variable Displacement Transformers

Fig. 3. Load frame with elevations.

281
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

Table 2
Failure mode of four specimens.
Test Description of failure
1 Casing bulging/core buckling/tensile failure of core at
two ends
2 Casing bulging/core buckling/compression buckling
of core
3 Casing bulging/core buckling/tensile failure of single
core plate
4 Core buckling/tensile failure at bottom connection
plate
Note: w-a = weak-axis; s-a = strong-axis.
(LVDT) and the force from the actuator load cell. In addition, strain
gauges were attached to the casing of the BRBs to detect local buckling
and bulging.
The four full-scale BRBs tested in this research demonstrated a
variety of failures under quasi-static cyclic loading, as summarized in
Table 2. The steel casing of the BRB in Test 1 experienced bulging due
to strong-axis buckling of the core plates, as shown in Fig. 5(a); the dual
core plates experienced strong axis buckling, as shown in Fig. 5(b);
ultimately the dual core plates failed in tension and fractured at both
ends near the connection plates; the normalized hysteresis loops were
stable, as shown in Fig. 5(c). The normalization values of yield force
and yield displacement are given in Table 3. The steel casing of the BRB
in Test 2 experienced bulging due to weak axis buckling as shown in
Fig. 6(a); the dual core plates buckled about the weak-axis as shown in
Fig. 6(b); the hysteresis loops were stable, as shown in Fig. 6(c). The
BRB in Test 3 failed in tension after the single core plate buckled about
the strong axis and is shown in Fig. 7(a); the steel casing did not show
visible bulging, however, strain readings from gauges applied at mid-
height of the casing in the hoop direction measured a tensile strain two
times the yield strain, as shown in Fig. 7(b), implying imminent bul-
Buckling direction
ging; the hysteresis loops were stable, as shown in Fig. 7(c). The BRB in
s-a Test 4 failed in tension after the single core plate buckled about the
strong-axis, as shown in Fig. 8(a); no bulging of the casing was ob-
w-a
served; the hysteresis loops were stable, as shown in Fig. 8(b). The
s-a maximum tensile strain achieved in the core plates of the BRBs is given
in Table 3. The maximum tensile strain achieved in the core plates of
s-a the four tests, expressed as the ratio of the maximum displacement over
the core length, ranged from 3.21% to 4.16%.
3.1. Compression strength adjustment factor
This factor is the ratio of maximum compressive to maximum tensile
force, measured in the same loading cycle. AISC 341 provisions require
this factor to be within certain limits, expressed as [25]:
Pmax
1.00 < β = < 1.30
Tmax (1)
where β is the compression strength adjustment factor; Pmax is the
maximum compressive force; and Tmax is the maximum tensile force in
the corresponding loading cycle. Table 3 lists the β values for the four
tests; all values vary from 1.06 to 1.19 and are below 1.30.
3.2. Strain hardening adjustment factor
Strain hardening is the strengthening of a metal under plastic de-
formation for tensile stresses larger than the yield strength. This factor,
denoted as ω is defined as the ratio of the maximum measured tension
force to the measured yield force:

Fig. 5. Test 1: (a) strong axis steel casing bulging;

(b) strong axis core buckling; (c) hysteresis.
(Triangle shows when the casing started to bulge).

Table 3
Structural performance parameters.

Test Maximum displacement Final Maximum strain β ω η E (kN-m) Yield Force1 Yield Displacement2 Yield Energy3 (kN-
(mm) cycle (%) (kN) (mm) m)

1 130 16 4.16 1.06 1.57 617 12882 3357 5.3 14.9

2 112 14 3.28 1.11 1.46 471 6500 2268 5.3 12.1
3 125 13 3.21 1.25 1.61 521 5977 1815 5.3 9.7
4 132 13 3.30 1.19 1.67 472 5560 1406 5.3 7.5
Raddon et al. 104 39 2.83 1.26 1.63 486 13110 3480 6.0 21
[11]

1
Yield force = (yield stress of core plate) × (cross-sectional area of core plates).
2
Yield displacement = (yield force)/(equivalent stiffness of the core plate and connection plates).
3
Yield energy = (yield force) × (yield displacement).

282
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

Fig. 6. Test 2: (a) weak axis steel casing bulging;

(b) weak axis core buckling; (c) hysteresis.
(Triangle shows when the casing started to bulge).

Fig. 7. (a) Strong axis core buckling; (b) strains

on steel casing; (c) hysteresis (Triangle shows
when the casing started to bulge).
3
Cumluative Displacement, x10 mm
0 1000 2000 3000
mm/mm

4000
3000 'SG #18'
'SG #19'
-6

'SG #20'
2000
Strain, x10

'SG #21'

1000
0
0 40 80 120
3
Cumulative Displacement, x 10 in.

(a) (b) (c)

3.3. Cumulative inelastic deformation

2
Inelastic deformation, μi , is the ratio of the permanent or plastic
portion of BRB axial displacement divided by the length of the yielding
1
portion of the brace; this factor indicates the ductility of the brace. The
Force/ Yield Force

inelastic deformation for each cycle, μi, is expressed as:

0
2(Δmax − com + Δmax − ten)
μi = −4
Δby (3)
-1
where Δmax − com is the maximum displacement of the brace in com-
pression; Δmax − ten is the maximum displacement of the brace in tension;
-2 and Δby is the yield displacement of the brace. The cumulative inelastic
deformation, η, is obtained from Eq. (4) as:
-30 -20 -10 0 10 20 30 n
Displacement/ Yield Displacement η= ∑ μi
(a) (b) i=1 (4)

Fig. 8. Test 4: (a) strong axis core buckling; (b) hysteresis.
Tmax
ω= ⩾ 1.00
Ry Pysc
where Tmax is the maximum tensile force; Ry is the ratio of the expected
yield stress to the specified minimum yield stress, fys; factor Ry need not
be applied if the axial yield strength of the steel core, Pysc, is established
using the yield stress determined from coupon tests, which is the case in
this research. The four BRBs eventually attained strain hardening and
Table 3 shows that the maximum values of ω ranged from 1.46 to 1.67.
where i is the loading cycle number, and η is the total number of cycles.
From Table 3, the cumulative inelastic deformation ranges from 471 for
Test 2 to 617 for Test 1. The results for the four specimens exceeded the
AISC 341 cumulative inelastic deformation requirement of 200 times
(2)
the yield displacement by a large margin [25].
3.4. Cumulative hysteretic energy dissipation
Earthquake input energy is dissipated as kinetic, damping, elastic
and hysteretic energy. Cumulative hysteretic energy dissipation is used
to measure the energy dissipated by a BRB. Hysteretic energy dissipa-
tion for each loading increment is obtained as:

283
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

800 and x j + 1 are BRB displacements at time j and (j + 1) , respectively.

Cumulative hysteretic energy dissipation is the summation of energy for
each loading increment, expressed as:
Cumulative Energy / Yield Energy

n−1
600
E= ∑ Ej + 1
j=1 (6)

where Ej + 1 is the hysteretic energy dissipation for each loading incre-

400 ment and E is the cumulative hysteretic energy dissipation.
Test1 The normalized cumulative hysteretic energy dissipation for the four
Test2 new BRBs is plotted in Fig. 9 and final values are given in Table 3. The
Test3 normalization values for yield energy are shown in Table 3. From
200 Test4 Table 3, it is clear that the cumulative hysteretic energy dissipation for
Tests 1–3 is higher than that for Test 4. BRBs in Tests 1–3 have the same
cross-sectional steel casing dimensions of 305 mm × 305 mm × 6.4 mm
and a steel core cross-sectional area equal to or larger than 6452 mm2, as
0 shown in Table 1. However, the BRB in Test 4 has cross-sectional steel
0 200 400 600 800
casing dimensions of 254 mm × 254 mm × 6.4 mm and a steel core
Cumulative Displacement/ Yield Displacement cross-sectional area of only 5000 mm2. As expected, BRBs with larger
Fig. 9. Cumulative hysteretic energy dissipation of new type BRBs. core cross-sectional area dissipate higher hysteretic energy.
Hysteretic energy is the area under the force–displacement hyster-
esis curves; Test 2 has a smaller steel core and thus the core force is
800 smaller; moreover the cumulative displacement for Test 2 is smaller as
Cumulative Energy / Yield Energy

shown in Fig. 9; thus the hysteretic energy is also smaller.

600 3.5. Comparison with conventional BRBs

The new type BRBs were compared to conventional BRBs previously

400 tested in the same load frame under a similar loading protocol. Raddon
et al. [11] reported cyclic test results for a conventional BRB with
pinned connections; this BRB had a steel core consisting of three plates
with a total cross-sectional area of 12258 mm2. The specimen geometry
200 Test 1
Raddon et al. (2009) and structural performance for the conventional BRB is shown in Tables
1 and 3, respectively. The conventional BRB had an HSS 324 × 9.5 mm
steel circular pipe as the casing. The new BRB of Test 1 in the present
0 investigation with bolted connections has a cross-sectional area of
0 200 400 600 800 11935 mm2 which is 97.4% that of the conventional BRB tested by
Cumulative Displacement / Yield Displacement Raddon et al. [11]. The new BRB in Test 1 reached a maximum strain of
(a) cumulative hysteretic energy dissipation 4.16% while the convectional BRB reached a maximum strain of 2.83%.
Table 3 shows that the conventional BRB dissipated overall 2% higher
hysteretic energy than the BBR in Test 1. The normalized cumulative
Raddon. et al, 2009 hysteretic energy dissipation (cumulative energy/(yield dis-
2 Test 1 placement × yield force)) vs. normalized displacement (displacement/
yield displacement) for these two tests is plotted in Fig. 10; the new
type BRB in Test 1 dissipated a similar or higher amount of normalized
Force/ Yield Force,

1 hysteretic energy as the conventional BRB, and moreover the normal-

ized displacement reached was higher than that of the conventional
BRB. Therefore, the new type BRB is more efficient than the conven-
0 tional BRB regarding normalized cumulative hysteretic energy dis-
sipation. In addition, the new type BRB had a smaller compression
strength adjustment factor and strain hardening adjustment factor, and
-1 a higher cumulative inelastic deformation, as shown in Table 3.

4. Strong and weak axis buckling

-2
-30 -20 -10 0 10 20 30 In the four full-scale new type BRB tests, core plates buckled about
Displacement/ Yield Displacement the weak or the strong axis, as summarized in Section 3 and Table 2.
(b) hysteresis curves Strong-axis buckling was observed in Tests 1, 3 and 4 as shown in
Figs. 5(b), 7(a) and 8(a). Weak-axis buckling was observed in Test 2, as
Fig. 10. Comparison of new type and conventional BRBs: (a) cumulative hysteretic en- shown in Fig. 6(b). The occurrence of buckling in different directions is
ergy dissipation; (b) hysteresis curves.
investigated herein using the strut-and-tie model (STM) [26]. The STM
model is a truss idealization of composite members consisting of steel
(Pj + 1 + Pj )(x j + 1−x j ) ties, concrete struts and concrete nodes, commonly used for idealization
Ej + 1 =
2 (5) of reinforced concrete and composite steel/concrete structures [26,27].
The BRB cross-sections in Tests 1 and 2 are shown in
where Pj and Pj + 1 are BRB forces at time j and (j + 1) , respectively; x j Fig. 11(a) and (b), respectively; Fig. 11(c) shows an idealized truss

284
W. Xu, C.P. Pantelides
model denoted as ABC, also known as a strut-and-tie model (STM), of a
portion of the BRB cross-section. The steel casing is modeled as tie AB,
the diagonal concrete struts are AC and BC, and the concrete nodes are
A, B, and C. The ACI 318 building code specifies the minimum angle, θ,
between a strut and tie as 25 degrees; angle θ needs to be large enough
to mitigate cracking and avoid incompatibilities due to shortening of
the strut and lengthening of the tie [26]. The European Code Eurocode
2 [28] limits the minimum angle between a strut and tie to a value of
21.8 degrees instead of 25 degrees. If this angle is reduced any further,
this results in larger values of principal tensile strain in the cracked
concrete and, hence, lower values of the crushing strength of the
cracked concrete [29]. When angle θ between a strut and a tie is less
than the minimum value, the concrete nodes and concrete struts are
likely to crush under significant axial compression. Once concrete
crushes, the core plates will buckle first about the strong axis as shown
in Fig. 5(b) before they have the opportunity to buckle about the weak
axis. When the angle between the strut and tie is larger than 25 degrees,
the concrete nodes and struts are stable and the steel core plates are
expected to buckle first about the weak-axis, as shown in Fig. 6(b).
STM models were constructed using the steel casing as the tie and
concrete inside the casing as struts and nodes for Test 1 as shown in
Fig. 11(a) (A1B1C1), and for Test 2 as shown in Fig. 11(b) (A2B2C2).
Angle θ between the strut and tie for Test 1 was 13.3 degrees, which is
less than both 21.8 and 25 degrees, so struts A1C1 and B1C1 and node C1
were expected to crush. Similarly, Test 3 had an angle θ of 8.1 degrees
and Test 4 had an angle θ of 11.6 degrees and the steel core plates in
both tests buckled about the strong axis. Angle θ between the strut and
tie for Test 2 was 28.9 degrees, which exceeds both 21.8 and 25 de-
grees, so the zone of concrete on the side of the core plates (A2B2C2)
shown in Fig. 11(b) was not expected to crush and this allowed the steel
core plates to buckle about the weak-axis, as shown in Fig. 6(b).
285
Engineering Structures 153 (2017) 279–289
Fig. 11. Strut-and-tie models for two BRB cross-sections: (a) Test 1;
(b) Test 2; (c) angle θ.
For a given steel casing dimension, ws, assumed as a square HSS
tube, one could design a new type BRB by limiting the steel core width,
wc, so that angle θ exceeds 25 degrees. For typical HSS square tube
sections, the STM model of Fig. 11(c) shows that angle θ exceeds 25
degrees when the ratio of the core width to casing dimension wc/ws is
less than 0.57. In the present tests, weak axis buckling occurred for Test
2 with wc/ws equal to 0.52, whereas strong axis buckling occurred for
Tests 1, 3, and 4 with wc/ws values ranging from 0.77 to 0.83.
5. Strong and weak axis local bulging
After a significant length of the steel core plate develops large in-
elastic deformations the wavelength stops changing. Once this occurs
the steel core compresses the concrete against the steel casing which
starts to deform and bulge locally. Bulging of the steel casing in Tests 1,
3 and 4 was about the strong axis, similar to Fig. 5(a). Bulging of the
steel casing in Test 2 was about the weak axis, as shown in Fig. 6(a).
The contact or local bulging force, applied on the concrete and steel
casing by the steel core plates is critical for determining whether the
steel casing will bulge. This bulging force is shown as Fp,max in
Fig. 12(b) for weak axis bulging and Fig. 13(b) for strong axis bulging.
For simplicity, only one wave is drawn in both Figs. 12(b) and 13(b).
The maximum transverse bulging force and the resistance of the steel
casing must be estimated to prevent this type of local bulging failure.
Lin et al. [20] showed that the maximum bulging force is proportional
to the maximum compression capacity of the BRB as well as the air gap
thickness, and is inversely proportional to the wavelength, as shown in
Eq. (7), with reference to Fig. 12:
2s
Fp,max = Pmax
(l w /2) (7)
W. Xu, C.P. Pantelides
where Pmax is the maximum compressive force applied on the BRB; s is
the air gap thickness between the concrete and the steel core and l w is
the wavelength. The capacity of the steel casing to resist local bulging is
modeled by considering an equivalent beam with a span equal to the
steel tube casing dimension, ws , and a width equal to the bulging force
region, leff , as shown in Fig. 12 for weak axis buckling, and Fig. 13 for
strong axis buckling.
For bulging of the casing due to weak axis buckling, Lin et al. [20]
proposed a 45 degree angle for spreading the bulging force Fp,max in
both the transverse and longitudinal directions of the BRB as shown in
Fig. 12, which is adopted in this research. Therefore, the width and
length of the effective inner face can be expressed as:
weff − w = wc + 2c w ⩽ ws (8a)
leff − w = 2c w ⩽ ws (8b)
where weff − w and leff − w are the width and length of the inner effective
area for weak-axis buckling of the core plates; wc is the width of the core
plate; c w is the concrete cover thickness in the weak-axis of the core
plates.
For bulging of the casing due to strong axis buckling, since the steel
core width is significantly larger than the core thickness, the contact
between the steel core and concrete is assumed as a line contact. The
fact that a line contact rather than a point contact can occur in the
buckling of a steel core has also been shown analytically [24]. ACI 318
[26] recommends that when the supporting area is wider than the
loaded area on all sides, the surrounding concrete confines the bearing
286
Engineering Structures 153 (2017) 279–289
Fig. 12. Effective inner surface for bulging due to buckling about the
weak-axis: (a) width, (b) length of a single wave; (c) equilibrium
beam model.
area, resulting in an increase in bearing strength. The supporting area
can be obtained by spreading the loaded area with a ratio of 2 to 1
(width to depth), as shown in Fig. 13. The width of the effective inner
surface is assumed equal to the thickness of the steel core plates plus
four times the concrete cover. For strong axis buckling, the experiments
and numerical simulations indicate that a contact surface equal to one-
third of the core width, wc/3, is appropriate with a spreading slope of
2:1 in both the transverse and longitudinal directions to determine the
width and length of the effective area, as shown in Fig. 13. The width
and length of the effective inner face are expressed in Eq. (9) as:
weff − s = tc + 4cs ⩽ ws (9a)
leff − s = wc /3 + 4cs ⩽ l w − s (9b)
where weff − s and leff − s are the width and length of the inner effective
area for strong-axis buckling of the core plates; tc is the thickness of the
core plates; cs is the concrete cover thickness in the strong-axis of the
core plates; based on previous research the wavelength is obtained as
l w − s= 5.9 wc in the case of strong axis buckling [30].
Both ends of the equivalent beam, representing the corners of the
steel casing tube, in Figs. 12(c) and 13(c) are assumed fixed. The de-
velopment of three plastic hinges, one at midspan and two at the fixed
ends of the steel casing tube is considered to be the flexural limit state
of the equivalent beam. The maximum bulging force, Fs, required to
develop the flexural limit state is the capacity of the steel casing given
as [20]:
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

Fig. 13. Effective inner surface for bulging due to buckling about the
strong-axis: (a) width, (b) length of a single wave; (c) equilibrium
beam model.

16Mps Fp,max
Fs = DCR =
(2ws−weff ) (10a) Fs (11)

When the DCR ratio exceeds 1.0 local bulging failure is expected. The
leff ts2 DCR ratio can be expressed in terms of the important BRB design
Mps = f ys
4 (10b) parameters as:

where Mps is the plastic moment capacity of the equivalent beam, weff is
the width of the effective inner surface, as shown in Fig. 12(a) for weak
axis buckling (weff − w ) and Fig. 13(a) for strong axis buckling (weff − s ),
and ts and fys are the thickness and yield strength of the steel casing. Eq.
(10) is adopted here for strong axis bulging. Lin et al. [20] proposed a
demand-to-capacity ratio (DCR) to evaluate the local bulging failure of
the steel casing, which is the ratio of maximum bulging force applied on
the casing given by Eq. (7) to the capacity of the casing given in Eq. (10)
as follows:
Pmax (2ws−weff ) s
DCR =
l w leff ts2 f ys (12)
which shows that the DCR ratio depends on the maximum compressive
force applied on the BRB, the air gap thickness, the cross-sectional di-
mensions of the steel casing, the width and length of the effective inner
surface, the wavelength, and the thickness and yield strength of the
steel casing. The wavelength for strong axis was given in Eq. (9b); for
weak axis buckling, the wavelength is l w − w= 11.9 tc, where tc is the

Table 4
DCR for bulging failure of steel casing and minimum steel casing thickness.

Test Pmax (kN) Pb (kN) Pb tc (mm) wc (mm) ws (mm) weff (mm) lw (mm) l eff (mm) ts (mm) DCR Observed failure Minimum thickness* (mm)
Pmax

1 5671 5577 0.98 51 235 305 165 1386 193 6.4 1.19 Bulging 7.9
2 3874 3243 0.84 51 159 305 305 605 241 6.4 1.02 Bulging 7.9
3 3834 3260 0.85 25 254 305 102 1499 161 6.4 1.02 Bulging 7.9
4 2900 N/A N/A 25 197 254 114 1161 155 6.4 0.81 None 6.4

* Note: Minimum theoretical casing thickness to prevent bulging; P b is the load when the casing started to bulge.

287
W. Xu, C.P. Pantelides
thickness of the steel core [30].
Table 4 lists the DCR for all specimens; Pmax is the maximum com-
pressive force obtained in the experiments. The compressive force ob-
tained in the experiments when the casing started to bulge is denoted as
Pb . According to Eq. (12), BRBs in Tests 1, 2 and 3 exhibited bulging of
the steel casing, whereas the BRB in Test 4 did not bulge; the DCR
values in Table 4 agree with the test observations. The casing thickness
required to prevent local bulging can be obtained from Eq. (12) by
limiting the value of the DCR to be equal to or less than 1.0 as follows:
Pmax (2ws−weff ) s
ts =
l w leff f ys (13)
Using Eq. (13), the minimum thickness of the steel casing, for the
BRBs tested in this research, is shown in the last column of Table 4
using available thicknesses from the Steel Construction Manual [31]. A
casing thickness of 7.9 mm would have been required in Tests 1–3 to
avoid local bulging. Eq. (13) can also be used in design by estimating
the value of Pmax as given in Eq. (14) obtained from AISC 341 re-
commendations [23]:
Pmax = βω (Ry f ys ) Ac (14)
where the values of β, ω, and Ry must be assumed. For the new type
BRBs tested in this research the following values are recommended: β =
1.15, ω = 1.60, and Ry = 1.00.
The ratio of casing bulging force to maximum compressive force, Pb
/ Pmax , is shown in the fourth column of Table 4. This ratio is 0.98, 0.84,
and 0.85 for Tests 1, 2, and 3, respectively. A conclusion can be drawn
that the steel casing starts to bulge when the compressive force reaches
84% of the maximum compressive force. However, this conclusion is
based on three tests; further tests need to be carried out to support this
conclusion.
6. Conclusions
New type buckling-restrained braces (BRBs) were tested which
differ from conventional BRBs in that they utilize prismatic steel plates
for the steel core. Connection plates oriented perpendicular to the steel
core were used to attach the BRBs to gusset plates. The following
conclusions can be drawn from this research:
(1) The new type BRBs use steel more efficiently and reduce the cost of
manufacturing. Based on the fact that there are two connection
plates perpendicular to the steel core at each BRB end, it is easy to
construct welded, pinned, or bolted connections.
(2) The hysteretic loops of the four full-scale BRBs tested in this re-
search exhibited stable behavior with positive incremental stiffness
by dissipating substantial amounts of hysteretic energy. There was
no rupture, brace instability, or brace end-connection failure up to
displacements corresponding to 2% story drift; the maximum ten-
sile strain achieved in the core plates ranged from 3.2% to 4.2%.
The ratio of maximum compression to maximum tension force was
less than 1.30 and the cumulative inelastic deformation ranged
from 471 to 617.
(3) Compared to conventional BRBs, the new type BRBs with similar
total steel core area performed in a satisfactory manner dissipating
similar amounts of hysteretic energy.
(4) Occurrence of either strong or weak axis buckling of the core plates
was observed. The geometry of the casing and width of the core
plates determine whether buckling will occur about the strong or
weak axis and this could be analyzed using the strut-and-tie
method.
(5) It is equally acceptable for the core plates to buckle either about the
strong axis or the weak axis; in either case the BRBs dissipate ap-
proximately equal amounts of energy based on the comparisons
presented in this paper.
288
Engineering Structures 153 (2017) 279–289
(6) Local bulging of the steel casing was observed about either the
strong or the weak axis in three of the four tests. A ratio of the
applied maximum bulging force compared to the force capacity of
the steel casing, using a plastic hinge model, predicted the occur-
rence of local bulging failure about either the strong or the weak
axis.
(7) The research shows that by limiting the steel core width for a given
dimension of steel casing, strong axis buckling of the steel core can
be prevented; recommendations are given for the ratio of maximum
core width to casing dimension to prevent strong axis buckling.
Local bulging of the casing about the strong or the weak axis can be
prevented by providing adequate thickness of the steel casing; re-
commendations are given for the minimum casing thickness.
Acknowledgements
The authors acknowledge the financial support of StarSeimic LLC
and the University of Utah. The input and assistance of Steve Powell
and Kimberley Robinson of StarSeismic LLC is acknowledged. The au-
thors also acknowledge helpful discussions with Professor Emeritus
Lawrence D. Reaveley of the University of Utah. The assistance of Mark
Bryant of the University of Utah in the experiments is acknowledged.
References
[1] Uang CM, Nakashima M, Tsai KC. Research and application of buckling-restrained
braced frames. Int J Steel Struct 2004;4(4):301–13.
[2] Usami T, Wang CL, Funayama J. Developing high performance aluminum alloy
buckling-restrained braces based on series of low-cycle fatigue tests. Earthq Eng
Struct Dyn 2012;41(4):643–61.
[3] Black C, Makris N, Aiken I. Component testing, seismic evaluation and character-
ization of buckling-restrained braces. J Struct Eng 2004;130(6):880–94.
[4] El-Tawil S, Ekiz E. Inhibiting steel brace buckling using carbon fiber-reinforced
polymers: large-scale tests. J Struct Eng 2009;135(5):30–538.
[5] Rahaei AR, Alinia MM, Salehi SMF. Cyclic performance of buckling restrained
composite braces composed of selected materials. Int J Civil Eng 2009;7(1):1–8.
[6] Genna F, Gelfi P. Analysis of the lateral thrust in bolted steel buckling-restrained
braces. I: Experimental and numerical results. J Struct Eng 2012;138(10):1231–43.
[7] Dusicka P, Tinker J. Global restraint in ultra-lightweight buckling-restrained braces.
J Compos Constr 2013;17(1):139–50.
[8] Chou CC, Chen YC, Phan DH, Truong VM. Steel braced frames with dual-core SCBs
and sandwiched BRBs: mechanics, modeling and seismic demands. Eng Struct
2014;72(1):26–40.
[9] Wang C, Li T, Chen Q, Wu J, Ge H. Experimental and theoretical studies on plastic
torsional buckling of steel buckling-restrained braces. Adv Struct Eng
2014;17(6):871–80.
[10] Tremblay R, Bolduc P, Neville R, Devall R. Seismic testing and performance of
buckling-restrained bracing systems. Can J Civ Eng 2006;33:183–98.
[11] Raddon BJ, Pantelides CP, Reaveley LD. Full scale subassemblage testing of pow-
ercat series buckling-restrained brace. Report No. CVEEN-09/1, Dept. Civil and Env.
Eng., Univ. of Utah, Salt Lake City, Utah, USA; 2009.
[12] Mirtaheri M, Gheidi A, Zandi AP, Alanjari P, Samani HR. Experimental optimization
studies on steel core lengths in buckling restrained braces. J Constr Steel Res
2011;67(8):1244–53.
[13] Miller DJ, Fahnestock LA, Eatherton MR. Development and experimental validation
of nickel-titanium shape memory alloy self-centering buckling-restrained brace. Eng
Struct 2012;40:288–98.
[14] Palazzo G, Lopez-Almansa F, Cahís X, Crisafulli F. A low-tech dissipative buckling
restrained brace: design, analysis, production and testing. Eng Struct
2009;31(9):2152–61.
[15] Ju YK, Kim MH, Kim J, Kim SD. Component tests of buckling-restrained braces with
unconstrained length. Eng Struct 2009;31(2):507–16.
[16] Wang CL, Usami T, Funayama J, Imase F. Low-cycle fatigue testing of extruded
aluminium alloy buckling-restrained braces. Eng Struct 2013;46:294–301.
[17] Mazzolani FM, Corte GD, D’Aniello M. Experimental analysis of steel dissipative
bracing systems for seismic upgrading. J Civil Eng Mgmt 2009;15(1):7–19.
[18] Sun FF, Li GQ, Guo XK, Hu DZ, Hu BL. Development of new-type buckling-re-
strained braces and their application in aseismic steel frameworks. Adv Struct Eng
2011;14(4):717–30.
[19] Zhao J, Wu B, Li W, Ou J. Local buckling behavior of steel angle core members in
buckling-restrained braces: cyclic tests, theoretical analysis, and design re-
commendations. Eng Struct 2014;66:129–45.
[20] Lin PC, Tsai KC, Wang KJ, Yu YJ, Wei CY, Wu AC, et al. Seismic design and hybrid
tests of a full-scale three-story buckling-restrained braced frame using welded end
connections and thin profile. Earthq Eng Struct Dyn 2012;41:1001–20.
[21] Chou CC, Chen SY. Subassemblage tests and finite element analyses of sandwiched
buckling-restrained braces. Eng Struct 2010;32:2108–21.
[22] Takeuchi T, Hajjar JF, Matsui R, Nishimoto K, Aiken ID. Local buckling restraint
W. Xu, C.P. Pantelides Engineering Structures 153 (2017) 279–289

condition for core plates in buckling restrained braces. J Constr Steel Res
2010;66(2):139–49.
[23] Takeuchi T, Hajjar JF, Matsui R, Nishimoto K, Aiken ID. Effect of local buckling core
plate restraint in buckling restrained braces. Eng Struct 2012;44:304–11.
[24] Wu B, Mei Y. Buckling mechanism of steel core of buckling-restrained braces. J
Construct Steel Res 2015;107:61–9.
[25] American Institute of Steel Construction. Seismic provisions for structural steel
buildings. AISC 341–10, Chicago, Illinois, USA; 2010.
[26] American Concrete Institute. Building code requirements for structural concrete and
commentary. ACI 318–14, Farmington Hills, Michigan, USA; 2014.
[27] Xiao Y, Zhang Z, Hu J, Kunnath S, Guo P. Seismic behavior of CFT column and steel
pile footings. J Bridge Eng 2011;16(5):575–86.
[28] EN 1992-1-1 (2004) (English): Eurocode 2: Design of concrete structures - Part 1–1:
General rules and rules for buildings [Authority: The European Union Per
Regulation 305/2011, Directive 98/34/EC, Directive 2004/18/EC].
[29] Collins MP, Mitchell D. Rational approach to shear design—the 1984 Canadian code
provisions. ACI J 1986;83(6):925–33.
[30] Wu AC, Lin PC, Tsai KC. High-mode buckling responses of buckling-restrained brace
core plates. Earthq Eng Struct Dyn 2014;43:375–93.
[31] AISC (American Institute of Steel Construction). “Steel Construction Manual”. AISC
325–11, Chicago, IL; 2011.

289