Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Journal of Constructional Steel Research 159 (2019) 428–441

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Effect of beam yielding on chevron braced frames


Charles W. Roeder a, Andrew D. Sen b,⁎, Clare Terpstra c, Sara M. Ibarra c, Ruyue Liu d,
Dawn E. Lehman a, Jeffrey W. Berman a
a
Dept. of Civil and Environmental Engineering, Univ. or Washington, Seattle, WA 98195, United States of America
b
Postdoctoral Research Associate, Dept. of Civil and Environmental Engineering, Univ. of Washington, Seattle, WA, United States of America
c
Former Graduate Student, Dept. of Civil and Environmental Engineering, Univ. of Washington, Seattle, WA, United States of America
d
Doctoral Student, Xi'an University of Architecture and Technology, Xi'an, China

a r t i c l e i n f o a b s t r a c t

Article history: Special concentrically braced frames (SCBFs) are commonly used as lateral-load-resisting systems in seismic re-
Received 14 January 2019 gions. Chevron (or inverted V-braced) SCBFs are used to accommodate architectural openings, such as doors,
Received in revised form 17 April 2019 windows, mechanical openings and elevators. However, current AISC provisions for SCBFs require beams of chev-
Accepted 26 April 2019
ron SCBFs to be designed to resist large vertical and axial demands resulting from unbalanced brace forces due to
Available online 17 May 2019
post-buckling compressive strength deterioration and full tensile yielding of the braces. As a result, very large
Keywords:
beams are required to support the braces, and chevron SCBFs are seldom built today. Recent research showed
Braced frames that limiting beam response to elastic behavior may reduce the overall deformation capacity of chevron SCBFs
Steel and some beam yielding may be beneficial. A research study was initiated to investigate the advantages of and
Chevron limits for beam yielding. Six large-scale, single-story chevron SCBFs were designed, built, and tested to investigate
Seismic current SCBF requirements and the impact of reducing the beam strength below current requirements to permit
Inelastic behavior beam yielding. Nonlinear continuum analyses were performed using the Abaqus finite-element analysis program
Full-scale tests to extend the experimental research. The research shows that chevron SCBFs with moderate beam yielding may
have improved seismic performance compared to other SCBF configurations. Yielding of the beam increases its
deflection, but the increased beam deflection increases inelastic deformation capacity of the braces prior to
brace fracture. The maximum lateral resistance and overstrength of the frame is reduced, but the resulting resis-
tance remains larger than the design target and is retained through relatively large story drifts.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction • Braces are designed to resist the factored lateral-load demands,


• Braces are designed to meet global (KL/r) and local (cross-sectional
Concentrically braced frames (CBFs) have historically been a width-to-thickness ratio) slenderness limits,
preferred structural system for seismic steel design because they • Connections, beams, and columns are designed to develop the ex-
provide high strength and stiffness. CBFs with buckling braces have pected tensile (Pye = RyFyAg) and compressive (Pcre = 1.14FcreAg) ca-
paired braces, which can result in architectural conflicts unless a chev- pacities of the braces for both their buckling (Pcre = 1.14FcreAg; see
ron (inverted-V) configuration is used. Because of this, many CBFs Fig. 1b) and post-buckling (0.3Pcre = 0.342FcreAg; see Fig. 1c) states,
built prior to 1988 employed chevron bracing. In 1988, the first • Beam-to-column connections are designed either to ensure adequate
capacity-based design requirements were introduced for CBFs [1]. Fol- connection rotation or transfer the full plastic capacity of the beam,
lowing this, in 1997, the first edition of the AISC Seismic Provisions for • Gusset plates are designed to accommodate the end rotation of the
Structural Steel Buildings [2] appeared; this document expanded these brace during buckling or brace connections are able to develop the
capacity-based design requirements, which are similar to the 2016 plastic moment capacity of the brace,
AISC Special CBF (SCBF) requirements [3]. The current SCBF provisions • Demand critical welds are used in regions where yielding is
require that: expected, and
• Yielding regions (brace and gusset plate) are treated as protected
zones where other attachments and modifications are prohibited.

⁎ Corresponding author.
E-mail addresses: croeder@uw.edu (C.W. Roeder), adsen@uw.edu (A.D. Sen),
For chevron-configured SCBFs, this results in the requirement of
delehman@uw.edu (D.E. Lehman), jwberman@uw.edu (J.W. Berman). nearly elastic response of the beams, resulting in very large beam

https://doi.org/10.1016/j.jcsr.2019.04.044
0143-974X/© 2019 Elsevier Ltd. All rights reserved.
C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 429

Fig. 1. Chevron braced frames.

sections required in design to resist the brace demands. Braces in chev- • Fully welded brace-beam-column connections which restricted
ron SCBFs are oriented in pairs, as shown in Fig. 1a, such that one brace out-of-plane movement of the braces, resulting in in-plane (IP) buck-
is in compression while the other is in tension. Using the seismic design ling and a corresponding increase in the expected brace buckling load.
provisions provided above, the beam supporting the braces (referred to
as the “chevron beam” herein) must be designed to support both the full More recent research suggests that the beam strength requirement
expected resistances, Pye and Pcre, of the braces (see Fig. 1b) as well as may be excessive. For example, a recent single-story chevron braced
the reduced post-buckling resistances, Pye and 0.3Pcre (see Fig. 1c). frame test was conducted using the earthquake simulator at the E-
This latter requirement induces a significant axial load and a large bend- Defense 3-D Full-Scale Earthquake Test Facility in Miki, Japan [9]. A pre-
ing moment in the chevron beam. As a result, chevron SCBFs require a liminary pushover analysis suggested that the chevron beam would
deep, heavy beam. Costanzo et al. [4] note that similar strength require- yield at its midspan under the unbalanced brace forces, but beam yield-
ments for chevron beams also exist in Canadian and European design ing was only observed in the test near the beam-to-column connections,
provisions, although beam stiffness relative to brace stiffness can vary which were fully restrained. The beam deformed elastically, resulting in
substantially. While chevron CBFs were common prior to about 1988, limited tensile elongation of the braces and hence an unbalanced force
they are rare today, because this chevron beam strength requirement which was lower than required in current design. Nonetheless, the
reduced the economy and efficiency of the system. frame provided good seismic performance even under severe seismic
The discussions leading to these provisions often focus on physical- shaking.
damage and serviceability issues associated with large floor deflections, Another recent study addressed the seismic performance of chevron
rather than the life-safety and collapse-prevention performance states. CBFs with yielding beams designed to pre-1988 requirements [10]. This
Yet, these latter performance states are the focus of current seismic de- two-story frame was tested under inelastic cyclic loading at the National
sign specifications. A recent review of literature suggests that the cur- Center for Research on Earthquake Engineering Laboratory in Taipei,
rent chevron beam strength requirement was instituted because of Taiwan. This frame had numerous additional deficiencies including
damage observed during a large-scale US-Japan CBF test performed at large local slenderness of the tubular braces, understrength beam-
the Building Research Institute in Tsukuba, Japan in 1983 [5,6]. This column-brace connections, and welds not meeting strength or current
US-Japan test appears to be the motivation of the current chevron demand critical toughness requirements. This frame was tested multi-
beam strength requirement and other current SCBF capacity-based de- ple times with different failures and retrofits (which were designed to
sign requirements. This test was performed on a three-dimensional, emulate the SCBF requirements), but the same weak beam was retained
full scale, six-story chevron CBF with joint funding by the US and in all tests. Significant beam yielding occurred, in particular for the
Japan [7,8]. The CBF had chevron beams with resistance well below retrofitted test where very significant inelastic deformation occurred
the current SCBF requirements and sustained early brace fracture, sig- prior to failure. Although this test did not meet all of the SCBF require-
nificant beam deflection, and slab damage. However, this CBF had nu- ments, the results suggest that yielding beams can result in good seismic
merous other deficiencies that likely contributed and/or caused this performance of chevron SCBFs.
unwanted response, including: Although compelling, not all research supports this conclusion.
Tremblay and Robert [11] used nonlinear response-history analysis to
• An unusual beam splice at the midspan brace-to-beam connections, investigate two-dimensional building models of chevron CBFs with
• Rectangular tubular braces with very high local slenderness (not in yielding beams. The researchers used the DRAIN-2D computer program
compliance with the current specification), and with all beam-to-column connections and column splices modeled as
430 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

pinned (i.e. no flexural restraint or resistance). The chevron beams were results to verify the numerical model. Finally, the simulations were
designed to develop 60%, 80%, or 100% of the tensile yield force of the used to extend the results beyond the experiments and to aid in
brace. The analytical results suggest that limited chevron beam yielding interpreting the experimental results. The second phase of the research
is acceptable for short buildings, but taller buildings showed deteriora- (described in a separate paper) evaluates the response of multi-story,
tion of performance, concentration of damage into a single story, and ductile chevron CBF systems with yielding beams and provides design
potential collapse or stability failure. The researchers proposed the use expressions for use in moderate to high seismic regions.
of weaker chevron beams but only for shorter building heights. A
more recent study [12] analyzed existing (i.e., not special) chevron 2. Experimental program
CBFs by the ASCE 41-13 [13] linear-dynamic and nonlinear-dynamic
procedures using OpenSees [14]. Beam-to-column connections were Six single-story specimens were built and tested in the University of
again modeled as pinned. While there are clear limitations in the ASCE Washington Structural Research Laboratory to study the effect of beam
41 procedures for evaluating CBFs, the general conclusions of this strength on the seismic performance of chevron CBFs. This paper sum-
study were similar in that significant seismic upgrade was typically re- marizes those results; more information about the specific tests can be
quired. These analytical studies suggest a less desirable picture of found in the reference theses and reports, including Terpstra [19], Ibarra
beam yielding in chevron CBFs, but it must be emphasized that they [20], and Roeder et al. [21]. Fig. 2a shows the specimen geometry and
were based on models with pinned connections, whereas research components. For each specimen, a pair of HSS4 × 4 × 5/16 A1085 braces
shows that braced frames have significant connection rotational re- were used, which meet the seismic AISC SCBF compactness criteria for
straint due to the gusset-plate connections used in practice. As such, highly ductile members. In addition, all specimens used W12 × 50 col-
these analytical studies are not conclusive. umns, which meet the requirements for columns in SCBFs. The primary
Other numerical studies have identified the impacts of chevron study parameter was the beam strength. In addition, beam depth and
beam stiffness, particularly where beams are still relatively strong brace steel specification was varied for two of the six specimens. The
[15–17]. Notably, D'Aniello et al. [16] found that if beams are too flexible full set of study parameters are given in Table 1 and described below.
(e.g., if the ratio of the beam flexural stiffness to brace vertical stiffness, The first four specimens are called Chevrons 1 through 4 and inves-
KF, is less than about 0.1), the braces do not yield in tension due to elastic tigated the effects of different beam strengths (all four specimens used
deflection of the beam. This behavior was generally associated with W14 beams because a large variety of W14 sections are readily avail-
poor seismic behavior, since energy dissipated by the braces is reduced. able). Chevron 1 was designed to have a beam with a section that satis-
However, in these analyses, the chevron beam's role in dissipating seis- fied current SCBF requirements (or the full resistance required in AISC
mic energy was likely limited, and seismic performance could be en- 341 [3]), but Chevrons 2, 3 and 4 were designed to have approximately
hanced by engaging the beam as a yielding element. 50%, 33%, and 25% of the currently required resistance (under combined
Current seismic provisions are developed to have an acceptable col- loading). The relative strength of the beam listed in Table 1 is deter-
lapse risk during relatively severe earthquakes. However, since some of mined by the P-M interaction curve resulting from application of the
the concerns leading to the chevron beam strength provisions focused AISC SCBF unbalanced force requirement. Table 1 also lists the beam-
on observed damage and serviceability, it is possible that yielding to-brace stiffness ratio, KF, defined by D'Aniello et al. [16]. Here, KF
beams will not increase the probability of collapse. In addition, yielding ranges from 0.17 in Chevron 1 to 0.03 in Chevron 4.
of chevron beams provides a secondary yield mechanism, and prior The test setup was designed to vary beam strength and stiffness of
work on the balanced design procedure shows significant benefit of sec- the specimens while minimizing other variations between individual
ondary yield mechanisms to improve system ductility and resist col- tests, since this more clearly isolated the effect of the beam on the in-
lapse [18]. elastic performance from other parameters. Ordinarily, large varia-
Founded on those results and to improve constructability, AISC tion in beam depth would be expected with this variation in
funded a research study to investigate decreasing the required chevron beam resistance. Changes in beam depth also result in changes in
beam strength. This paper describes the integrated experimental and brace length, angle, and story height, which complicate data
analytical tasks which form the first phase of that research focusing on interpretation.
single-story CBFs. Six single-story chevron CBFs with various beam The two remaining specimens studied the impact of tube specifica-
strengths were tested. Nonlinear analyses were performed in support tions (Chevron 5) and beam depth (Chevron 6). Chevron 5 was identical
of the analytical research study, where initial analyses were conducted to Chevron 3 except that it used A500 Grade C rather than A1085 braces.
simultaneously with the experiments to aid in selecting and placing in- In contrast to the A500 specification, the A1085 specification introduces
strumentation, and later analyses were compared to experimental a larger nominal yield stress, stricter wall-thickness tolerance, and

Fig. 2. (a) Experimental specimen and (b) Test setup


C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 431

Table 1
Test specimen member sizes and material designations.

Specimen Chevron 1 Chevron 2 Chevron 3 Chevron 4 Chevron 5 Chevron 6


Study parameter Beam size Brace type Beam depth
Beam size W14×120 W14×61 W14×38 W14×26 W14×38 W21×44
Brace size HSS4×4×5/16
Colunn size W12×50
Brace ASTM A1085 A1085 A1085 A1085 A500 Gr. C A1085
Bm./col. ASTM A992
Stiffness ratio, KF 0.17 0.08 0.05 0.03 0.05 0.10
Beam interaction 0.87 1.81 2.83 4.32 3.02 1.77
Note: Highlighted cells show study parameter. Color indicates comparison specimens.

minimum Charpy V-notch toughness. Chevron 6 was identical to Chev- • The welds joining the gusset plate to the beams, columns, and base
ron 2 except that it used a deeper beam section with comparable plates are strong enough to develop the plastic capacity of the gusset;
strength and moderate changes to the brace angle and length; the pri- • The rotational clearances for end rotation of the buckling braces are
mary purpose of this specimen was to investigate the impact of stiffness defined by the 6tp vertical clearance for the midspan gusset plates
(since the W21 beam was stiffer than its W14 counterpart) and axial re- [23] and the 8tp elliptical clearance for corner gusset plates [24],
sistance on the response. since these methods result in more compact gussets which facilitate
The members were designed to the requirements of AISC 341–16 end rotation;
[3] and AISC 360-16 [22], with the exception of the beam resistance • The Whitmore width for evaluating tensile yielding and compressive
of some specimens. The column transfers very little axial load from buckling of the gusset plate employs a 37-degree angle in each direc-
the brace, and so it was primarily designed to support an axial com- tion to more accurately estimate the resistance of the gusset plate
pressive load typical of the factored live and dead loads (LL and DL, while promoting a thinner, more compact plate which permits limited
respectively) for seismic loading (i.e., 1.0DL + 0.25LL) in steel build- yielding after buckling and yielding of the brace; and
ing construction as well as the maximum shear force transferred • Block shear requirements are slightly reduced to further permit lim-
from the beam to the column. The beams and columns were all of ited yielding of the gusset plate after bucking and yielding of the brace.
A992 steel. The gusset-plate connections were designed to develop
the expected capacity of the brace as required for SCBFs in AISC The corner gusset-plate connections were identical for all test spec-
341–16 [3], with enhanced drift capacity using the balanced design imens, but slight geometric changes were required for midspan gussets
procedure (BDP) [18]. In addition to the normal SCBF requirements, to accommodate differences in beam depth. Typical gusset connection
the BDP requires that: details are shown in Figs. 2a, 3a, and c.

4-25 mm A490-X BOLTS


DOUBLER PL* 64 mm SPACING
STIFFENER PL 9.5 mm 279×152×6 19 STD HOLES
305

W14 W14 13
GUSSET PL 9.5 mm
6 13
57 960 DOUBLER PL
406×254×13
SHEAR PL
313 8 254×121×16 W12×50
229
4 SIDES *CHEVRONS 2-6 ONLY
9.5 TYP
HSS4×4×5/16
TYP (a) Midspan gusset plate (b) Shear plate
4 SIDES
9.5 HSS4×4×5/16
(a) (b)
W12×50 356

GUSSET PL 9.5 mm
229
533 (c)
CJP
FLANGE
8

BASE PL 38 mm
9.5

(c) Corner gusset plate (d) Connection locations

Fig. 3. Typical specimen connection details.


432 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

A bolted shear-plate connection was used at the beam-column joint, The columns were loaded with a compressive load simulating the
as shown in Figs. 2a and 3b. Shear-plate connections can accommodate approximate dead and live loads for the earthquake load combination;
the required rotation and are likely to be the preferred connection for the column axial load was applied prior to the lateral load. The base
beam-to-column connections in an SCBF without a corner gusset plate of the south column base was post-tensioned to the strong wall
plate. The bolts, shear plate, and plate-to-column welds were designed such that it was integral with the strong wall, and the north column
to resist the shear and axial forces estimated from the brace forces that base was post-tensioned to a reinforced-concrete block, which was
corresponded to the expected plastic moment capacity of the beam at post-tensioned to the strong floor. Analysis and experimental results
the midspan gusset. As a result, this connection changed modestly showed that both column base connections were essentially fully re-
with different beam sizes. In all specimens, A490-X bearing bolts were strained connections.
used, and the steel plates for the shear plates and gussets were A572 Lateral support was applied to the beam and column as illustrated in
Grade 50 steel. Bearing of the bolts on the beam web or the shear Figs. 2b and Fig. 4. The locations of these lateral supports for the columns
plate often controlled the bolt resistance, and beam-web doubler plates were the same for all tests, but the lateral support of the beam was in-
were used for Chevrons 2 through 6 to prevent block shear. Web dou- creased or decreased for each test to ensure that each beam had ade-
blers were also added in the columns in the vicinity of the beam- quate support to achieve its full plastic moment capacity in all cases.
column joint because the actuator load was distributed to the frame at The test was conducted using displacement control through the ac-
these locations (such accommodations would likely be unnecessary in tuator linear variable displacement transducer (LVDT), but the actual
practice). measurement used to determine the applied displacement was the
The gusset- and shear-plate welds were demand critical welds, since story drift measured using separate instrumentation attached indepen-
significant yield and deformation was expected. dently to the strong floor. The cyclic displacement protocol shown in
The specimens were tested horizontally, as shown in Figs. 2b and 4; Fig. 5 was derived to achieve story drifts necessary to initiate brace
this setup facilitated control of the out-of-plane response. The framing buckling and tensile yielding of the brace as well as multiples of the
direction was parallel to the strong floor, but the center plane of the yield drift. The protocol was expanded by the ATC-24 method [25].
frame was 686 mm (27 in.) above the level of the strong floor to permit Instrumentation of the frame measured loads, global movements,
out-of-plane (OOP) brace buckling. The lateral loads were applied by a local movements, and strains. The applied load was measured by the in-
hydraulic actuator with a ± 154 mm (±10 in.) stroke and 2002 and ternal load cell in the actuator. Vertical deflection of the beam, story
2447 kN (450 and 550 kip) capacity in tension and compression, drift, and axial deformation and OOP deflection of the braces was mea-
respectively. sured with string potentiometers. Potentiometers were also used to
Earthquake demands are derived from the inertial loads, which are a measure any movement of the column bases, reaction blocks, and
function of the structural mass. In a building, the majority of the mass is other fixtures. All potentiometers were calibrated prior to each test.
the structural floor system and horizontally distributed building infra- Sets of multiple strain gauges were attached to all braces, beams, and
structure (other contributions to the mass include the weight of the columns in regions remaining elastic during the test to establish axial
other structural components and building façade). During earthquakes, loads, bending moments, and shear forces in these members.
these forces are transferred into the lateral-load frame through the A Northern Digital Incorporated (NDI) Optotrak optical measure-
composite-deck system and beams. To simulate this ideally in the labo- ment system was also used to monitor movements. Optotrak targets
ratory, the lateral loads would have been applied to the test frame were installed on the beam web, midspan gusset, and over the brace
through the beam, but this was not possible because doing so would un- length to monitor the in-plane and out-of-plane movements. Some of
realistically restrain the vertical deflections expected in the yielding these measurements duplicated measurements made by traditional in-
chevron beams. Instead, a load-transfer beam (see Figs. 2b and Fig. 4) strumentation. Areas where yielding was anticipated were
was used to transfer the loads from the actuator to the columns of the whitewashed prior to testing to enable determination of initiation and
frame. Pins were placed between the actuator and the transfer beam extent of yielding. Material properties were determined through tensile
and between the transfer beam and the columns to minimize bending tests for the steel sections and Table 2 summarizes the results of these
moments in the actuator and the tops of the columns. This arrangement tests. Strain data are not available from these coupon tests, but an elastic
induces an added overturning moment on the frame and increased modulus of 200,000 MPa (29,000 ksi) can be assumed for all steel used.
loads on the columns. The other end of the actuator was post-
tensioned to a reinforced-concrete reaction block, which was post- 3. Experimental results
tensioned to the strong floor.
The tests were completed between October 2016 and September
2017. In designing a chevron CBF, it is anticipated that the engineer
will select braces so that the design horizontal story shear in a given
bay and story will be less than 2Pcrcosθ of the braces, where θ is the
angle of the brace with respect to the horizontal. This represents the
equilibrium condition at initial brace buckling. After brace buckling,
the compressive force may decrease while the tensile force will

6
4
Drift (%)

2
0
-2
-4
-6
0 20 40 60 80
Half-cycle Number

Fig. 4. Annotated photo of a test specimen in the test apparatus. Fig. 5. Test loading protocol.
C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 433

Table 2 (17.8 in.). Significantly less yielding occurred in the shear plate of this
Material properties for test specimens. specimen, but larger beam deflections and more beam yielding were
Specimen Section Specification Fy Fu noted. Bolt fracture did not occur.
(MPa) (ksi) (MPa) (ksi)
Chevron 3 had a W14 × 38 beam, which was approximately 35% of
AISC 2016 SCBF requirements, and its measured response is shown in
1 W14 × 120 A992 391 56.7 483 70.1
Fig. 6c. Up to the initiation of the buckling the behavior of Chevron 3
2 W14 × 61 A992 403 58.5 512 74.3
3,5 W14 × 38 A992 366 53.1 474 68.7 was nearly identical to Chevrons 1 and 2 but, after brace buckling, the
4 W14 × 26 A992 359 52.1 419 60.8 response of Chevron 3 was similar to Chevron 2. Chevron 3 reached a
6 W21 × 44 A992 400 58 500 72.5 maximum drift range of 7.2% prior to brace fracture, and had a maxi-
1–6 W12 × 50 A992 361 52.4 454 65.9 mum lateral resistance of 757 and –786 kN (170.1 and −176 kips).
1–4, 6 HSS4 × 4 × 5/16 A1085 431 62.5 516 74.8
5 HSS4 × 4 × 5/16 A500 Gr. C 467 67.7 572 82.9
The north brace fractured at the first −3.6% drift cycle, and the south
1–6 3/8-in. plate A572 Gr. 50 519 75.3 586 85 brace fractured at the second +3.6% drift cycle. The maximum OOP de-
flection of the brace prior to fracture was 498 mm (19.6 in.). Beam de-
flections and yielding were larger than observed for Chevron 2 (see
increase. As such, it is uncertain if the frame will develop a story shear Fig. 7d).
larger than 2Pcrcosθ. As a result, 2Pcrcosθ is the estimate of the design Chevron 4 utilized a W14 × 26 beam, which was approximately 23%
shear resistance of the braced frames in this research. of current SCBF requirements. Again, Chevron 4 has similar behavior to
The force-drift response of the six specimens are shown in Fig. 6. Chevrons 1, 2, and 3 up to the point of brace buckling, and its behavior is
Fig. 7 shows selected photographs of brace, column, and beam behavior illustrated in Fig. 6d. The north brace fractured at the first −4.4% drift
observed during the testing. The measured story shear force was nor- cycle, and the south brace fractured at the second +4.5% drift cycle. In-
malized by 2Pcrcosθ determined from the measured properties of the elastic behavior was similar to Chevrons 1 and 2 in that the braces began
steel. There is significant interest in the earthquake engineering com- to buckle OOP at approximately ±0.28% drift, and sustained
munity to determine post-fracture resistance and stiffness to enable col- progressively larger OOP buckling deformations each cycle, leading to
lapse resistance of SCBFs. Therefore, a series of larger inelastic cycles plastic-hinge formation in the center of the braces; this eventually
after brace fracture were applied to measure remaining stiffness and lead to tearing and fracture of the braces (as illustrated in Fig. 7b and
resistance. c). Connection damage included plate yielding, plate crack initiation,
The following paragraphs summarize each specimen individually. and limited weld tearing. The columns sustained yielding and local
All of the specimens reached the idealized lateral strength, although buckling in the plastic-hinge region. The frame had a lateral resistance
Chevron 4 barely reached this threshold. In all cases, the significant of 675 and –719 kN (151.7 and −161.6 kip). The maximum OOP deflec-
loss in lateral-load-carrying capacity resulted from brace fracture. Chev- tion of the brace observed prior to fracture was 470 mm (18.5 in.).
ron 1 showed significant increase in resistance after brace buckling, but Chevrons 1 through 4 were designed to study the impact of beam
it was also the only specimen to show gradual deterioration in strength strength on the seismic performance (strength, stiffness, deformation
prior to brace fracture. capacity and damage) of chevron-configured SCBFs. Chevrons 1 through
Chevron 1 met all of the current AISC SCBF Seismic Provisions [3]. Its 3 all developed and retained the target design resistance. In contrast,
force-displacement response is shown in Fig. 6a. Out-of-plane (OOP) Chevron 4 nearly reached this resistance at initial buckling but
buckling was visually observed at ±0.28% drift, and gusset plates remained below this target resistance at increased deformation, as
began to yield due brace buckling deformation at approximately ± shown in Fig. 6d. Yielding of the beam and columns was more extensive
0.7% story drift. The maximum lateral resistance was 1020 and and severe in Chevron 4 than the prior tests. There was considerable
−1040 kN (229.3 and −233.8 kip). Significant yielding of the column torsion, axial, OOP P-δ deformation of the beam during later stages of
was noted adjacent to the top of the corner gusset plates at approxi- the Chevron 4 test, suggesting that this specimen did not meet seismic
mately ±1.9% story drift; it is postulated that the base fixity resulted performance expectations.
in high moment demands there (see Fig. 7a). Local flange buckling The test program included two additional specimens. Chevron 5 had
(also seen in Fig. 7a) was noted at that location at about ±2.2% drift. identical frame geometry, column size, brace size, and connection de-
Local cupping of the braces was noted with increasing deformation tails to Chevron 3, but it utilized A500 Grade C braces instead of
(see Fig. 7b). The south brace fractured first at 3% drift, and the north A1085 braces to study the impact of brace specifications. The W14
brace fractured second at −3.3% drift (see Fig. 7c). Cracking of gusset- × 38 beam had 35% of the strength required by the AISC SCBF require-
plate-to-framing member welds initiated prior to brace fracture, but ment for unbalanced load [3]. The behavior of this specimen is illus-
the cracks remained small and stable due to proper sizing and use of de- trated in Fig. 6e. Chevron 5 reached a maximum drift range of 6.6%
mand critical weld material (AWS E71T-8). Significant local yielding of and had a maximum lateral resistance of 780 and –774 kN (175 and
the shear plates was noted and one shear-plate bolt fractured simulta- −174 kips) as shown in Fig. 6e. The north brace fractured at −3.0%
neously with brace fracture. Large OOP deformations of the braces drift, and the south brace fractured at +3.6% drift. This behavior is sim-
were noted, and the maximum OOP deflection observed prior to brace ilar to that of Chevron 3 except that the inelastic deformation prior to
fracture was 439 mm (17.3 in.). The maximum drift range of the brace fracture was 9% smaller than that achieved with Chevron 3.
frame prior to brace fracture was 6.3%. Limited elastic beam deflection The sixth specimen was designed to consider the impact of beam
was noted, but deflections were small and no significant yielding oc- stiffness on the response. Chevron 6 had the same frame geometry, col-
curred until after the first brace fractured. umn size, brace size, and connection details as Chevron 2 but utilized a
Chevron 2 had a W14 × 61 beam, which has approximately 55% of W21 × 44 beam instead of a W14 × 61. Its measured response is shown
the chevron beam resistance required by the 2016 SCBF provisions; its in Fig. 6f. The resistance of the W21 × 44 was approximately 56% of that
measured behavior is shown in Fig. 6b. The initiation of brace buckling required by the 2016 AISC SCBF requirements for the unbalanced load
and the yielding of the columns was very similar to that noted in Chev- from the braces of chevron braced frames. All other components and
ron 1, except that beam deflections and story drifts were slightly larger welds met AISC seismic requirements [3]. Chevron 6 reached a maxi-
for each event after brace buckling. The south brace fractured first at mum drift range of 6.0% and had a maximum lateral resistance of 911
+3.3% drift half-cycle, and the north brace fractured at the second and –910 kN (204.8 and −204.6 kips). The south brace fractured first
−3.3% drift half-cycle; the resulting drift range was 6.6%. The maximum at +3.0% drift, and the north brace fractured second at −3.0% drift.
story shears prior to brace fracture were 919 and –888 kN (206.6 and Chevron 6 achieved almost equal maximum lateral resistance and
−199.6 kip). The maximum OOP deflection of the brace was 452 mm slightly lower drift range as Chevron 2, which sustained a maximum
434 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

Fig. 6. Normalized lateral force versus story drift.

of 909 kN (206.6 kips) and 6.6% drift range. The damage progression 4. Comparison of test specimens
(brace buckling, plastic hinging, and fracture) of Chevron 6 was similar
to Chevron 2. Connection damage again included plate yielding, plate Table 3 compares salient maximum measured data points for each
cracking, and weld tearing, but the weld tearing in Chevron 6 was less test, including beam deflection, drift range and strength. The study pa-
severe than that of Chevron 2. The columns sustained yielding and rameters were evaluated using this comparison data.
local buckling in the plastic-hinge region. The beam deflection was
slightly smaller for Chevron 6 than for Chevron 2 because of the in- 4.1. Beam deflection
creased elastic stiffness. This smaller beam deflection increased the ten-
sile force and deformation demand in the brace, resulting in brace Fig. 8a compares the beam deflections of Chevrons 1 through 4 as a
fracture at smaller frame drifts. function of story drift. Chevron 1 had small, elastic deflection, since
C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 435

(a) Column yielding adjacent to gusset (b) Local cupping of brace


(Chevron 1) (typical of all tests)

(c) Brace fracture (typical of all tests) (d) Beam yielding (Chevron 3)

Fig. 7. Photos of observed behavior.

beam yielding was minimized. At low drift levels (less than 1%), Chev- 4.2. Lateral resistance
rons 1 and 2 had similar deflections, since these deflections were largely
elastic. At drift levels exceeding 2% story drift (4% drift range), the beam Table 3 compares the maximum lateral resistance and deformations
deflections of the four specimens were quite different, with the largest prior to brace fracture for Chevrons 1 through 4. As shown in the
deflection approximately 3 times the smallest (Chevron 4 in compari- table and Fig. 6, the resistance of the frame decreased with reduced
son to Chevron 1). Although the deflections were increased with in- beam strength and increased beam yielding. Fig. 6 shows that only
creased beam yielding, they were still moderate in magnitude, even at Chevron 4, which had 23% of the currently required beam strength,
story drifts beyond 3%, since the maximum deflection-to-span ratio failed to meet the target design resistance through significant inelastic
was approximately 1/40 and the maximum residual deflection-to- deformation. In addition to demonstrating that beam yielding does
span ratio was approximately 1/120. Fig. 8b compares beam deflections not significantly impact resistance, the results also show that an in-
of Chevron 2 and 6. This figure provides insight into the beam deflection crease in beam yielding (following brace buckling) results in an in-
as a function of beam stiffness rather than strength, since both beams crease in drift range prior to brace fracture. For example, compare
had approximately 55% of the currently required beam strength. As ex- Chevrons 1, 2, and 3. The minimum drift-range capacity was 6.0% for
pected, the beam deflections of Chevron 6 (KF of 0.10) were smaller Chevron 1 and it should be noted that this drift range is somewhat
than the beam deflections of Chevron 2 (KF of 0.08), but the difference larger than that achieved in testing of other SCBF systems. This
at larger drift ranges was smaller than expected: 10 mm (0.4 in.). This shows that deflection of chevron beams reduces the fracture potential
difference was entirely due to the variation in moment of inertia and for the brace at a given drift, because beam deflection increases com-
elastic deflection of the beam under the unbalanced load. pressive deformation but decreases tensile elongation of the brace in

Table 3
Selected maximum measurements.

Specimen Chevron 1 Chevron 2 Chevron 3 Chevron 4 Chevron 5 Chevron 6

Beam size W14 × 120 W14 × 61 W14 × 38 W14 × 26 W14 × 38 W21 × 44

Brace A1085 A1085 A1085 A1085 A500 Gr. C A1085


Relative beam strength 115% 55% 35% 23% 35% 56%
Max. lateral resistance (kN) −1040 919 −786 −719 780 911
Max. beam deflection (mm) 28 46 76 97 64 36
Max. beam residual deflection (mm) 1.0 13 45 62 38 10
Drift at 1st brace fracture (%) 3.0 3.3 −3.6 −4.4 −3.0 3.0
Drift at 2nd brace fracture (%) −3.3 −3.3 3.6 4.5 3.6 −3.0
Max. drift range (%) 6.3 6.6 7.2 8.9 6.6 6.0
Corner gusset OOP deflection (mm) −61 −64 −64 −71 −58 −64
436 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

Fig. 8. Measured midspan deflections of chevron beam: (a) Chevrons 1 through 4 (b) Chevrons 2 and 6.

the hinge region. Fracture is driven by tension demands, which are re- strength is normalized to Pcre in both directions; Py is denoted with a
duced in this configuration. Chevrons 2 and 6 investigate the impact dashed line on the positive side of the y-axis. The south brace was out
of beam stiffness. The peak resistances were essentially the same. of phase with nearly identical behavior. Chevron 1 developed the com-
The maximum deformation prior to brace fracture was lower for pressive buckling load and nearly developed the tensile yield force, and
Chevron 6 because the decreased elastic defection increased the ten- significant tensile elongation was noted (primarily as yielding around
sile elongation and induced earlier fracture. the plastic hinge of the buckled brace). Chevrons 2 through 4 developed
the compressive bucking force, sustained increasing brace shortening
4.3. Unbalanced force deformation with a reduction in beam strength, and had decreasing ten-
sile force and elongation with decreased beam strength. It is noted that
The brace forces were determined from the strain gages placed in re- the elastic stiffness of the beams was also systematically reduced in the
gions that remained elastic, and unbalanced forces on the beam were test program as a byproduct of the strength reduction; reduced beam
computed from these measured brace forces. Fig. 9 is a plot of these un- stiffness has been shown to be similarly correlated with increased com-
balanced forces for Chevrons 1 through 4 as a function of story drift, pressive shortening and decreased tensile elongation [16]. In addition,
where the measured unbalanced forces are normalized by those re- Chevrons 2 through 4 had beam-to-brace stiffness ratios (KF) between
quired by current AISC SCBF requirements. Of significant note is that 0.03 and 0.08, suggesting that the beam stiffness would not be large
the idealized AISC forces were not achieved for any of the specimens, in- enough to induce brace tensile yielding, even if the beams were stron-
cluding Chevron 1. The unbalanced force demand decreased with de- ger. Still, these test results clearly show that, for chevron SCBFs with
creased beam strength and increased beam yielding, but this yielding beams, the maximum tensile force developed in the brace is
relationship was not linear. For example, the unbalanced forces for limited by the beam plastic resistance. Chevron 4 had no significant ten-
Chevrons 3 and 4 were very similar despite a significant change in sile elongation, and its maximum brace tensile force was essentially the
beam resistance. The plot suggests that Chevron 3 represents a reason- same magnitude as its buckling force.
able design limit for chevron SCBFs with yielding beams.
5. Numerical simulation
4.4. Brace elongation and shortening
To further study the impact of beam yielding on chevron SCBFs, a
Fig. 10 shows the axial force versus elongation or shortening hyster- small parametric study was conducted using high-resolution nonlinear
etic curves of the north braces of Chevrons 1 through 4. The brace finite-element models (FEM) (see Fig. 11). The initial phase of this study
validated the model using the test results.
The modeling was conducted in Abaqus [26]. The full specimen and
setup was modeled using four-node quadrilateral shell elements with
reduced integration (Element Type S4R). Significant yielding as well
as local and global geometric nonlinearity was modeled. However,
brace fracture and weld cracking were not directly simulated. Instead,
values of equivalent plastic strain were used to numerically estimate
the onset of these behaviors; this approach has been successfully used
in prior research by the authors [27]. Note that simulation of brace frac-
ture is critical in determining the seismic performance of the systems, as
it determines the drift-range capacity.
Because plastic strain is mesh dependent, a systematic and consis-
tent modeling approach for mesh size was used. A moderately coarse
mesh (up to 50 × 50 mm) was used in regions of elastic behavior, and
a fine mesh (19 × 19 mm) was used for braces in their buckled region.
An intermediate mesh size was used in other locations with yield defor-
mations with appropriate transitions to accommodate changes in
mesh size.
Fig. 9. Normalized vertical unbalanced forces for Chevrons 1–4. A bilinear kinematic-hardening material model was used with the
measured material properties (see Table 2). The elastic modulus was
C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 437

Fig. 10. Measured brace axial force versus axial deformation.

200,000 MPa (29,000 ksi) with a Poisson's ratio of 0.3 for all members. explicitly modeled as shown in Fig. 11, and the column displacements
Approximately 35,000 elements were required to model each were constrained to match those of the transfer beam where pins
specimen. were located in the test setup. Bilinear spring elements connected the
The column base plate connections were modeled as restrained in all beam web to the shear plate at each bolt location in the global
directions, and OOP restraints were imposed on beams and columns at X-direction, and a translational constraint was imposed in the other
the locations of the OOP lateral-torsional buckling restraint. Welded two directions. The yielding resistance of the bilinear spring was defined
joints were simulated using tie constraints restricting relative transla- by the bolt bearing and shear capacities. In the experiments, the braces
tion and rotation between joined nodes. The load-transfer beam was were fabricated with a downward eccentricity of 3 mm (0.125 in.) to

Fig. 11. Overview of finite-element model.


438 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

initiate buckling. Therefore, a negative Z-direction (OOP) 3-mm offset measurements at larger drift demands, because gusset-plate weld
was modeled to initiate buckling without causing significant change to cracking initiates at about 2% drift and limits the strength. However,
the brace buckling capacity. as mentioned previously, this cracking was not simulated in the
Nonlinear analyses were performed prior to testing to aid in Abaqus model. The simulation of Chevron 1 underestimated the
instrumenting and testing each specimen. Ultimately, the models post-buckling compressive resistance and OOP deflection of the
were validated using the test results. The local and global yielding brace again due to the fact that gusset-plate weld cracking was
mechanisms and deformations were compared for all test speci- not simulated in the model. Fig. 12b shows that the lateral resis-
mens; Figs. 12 and 1 show a few comparisons made to validate the tance for Chevron 3 was quite accurately predicted at all deforma-
models for Chevrons 1 and 3. Chevron 1 is shown because it had tions. The analyses slightly overestimate the vertical deflection of
the largest discrepancies between measured and computed behav- the beam and the OOP deflection of the brace at large story drifts,
ior despite having limited yielding. Chevron 3 was selected as a again due to gusset-plate weld cracking.
specimen with large inelastic deformation near the limit of tolera- Fig. 1 compares the computed von Mises stress fields in the brace
ble yielding in chevron beams. Fig. 12a shows that the model pre- with observed brace buckling deformation at specific story drifts. OOP
dicted the lateral resistance of Chevron 1 well. The computed deformations of the braces were large and significantly impacted the
lateral resistance was somewhat larger than experimental global frame behavior, local beam behavior, and gusset-plate rotation

Fig. 12. Comparison of measured and computed response of Chevrons 1 and 3.


C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 439

Fig. 13. Local behavior of brace for Chevron 1 (observed response above and simulated response below).

demands. The figure shows good correlation between the observed and provides good predictions of global and local behavior of the chevron
numerically simulated buckled shape and OOP deformation of the brace CBFs. Given this conclusion, it was justified to use the numerical simula-
at moderate drift levels. Brace fracture was preceded by local cupping tion methodology to extend the experimental study with a limited pa-
deformation of the brace, as shown in Fig. 1b. Brace tearing and fracture rameter study.
was not modeled, but Fig. 1c shows serious necking predicted at the Of primary importance for the system is the minimum limit of the
story drift and location of brace fracture. All of these mechanisms required beam strength. The experimental study only focused on a spe-
were well predicted by the simulation models. In summary, these com- cific range of beam strengths; this was extended with the FEM paramet-
parisons demonstrated that the nonlinear simulation methodology ric study.

Fig. 14. Effect of increased demand-to-capacity ratio.


440 C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441

Fig. 14 shows the global results of the study. As noted earlier, chevron CBFs with identical columns, connections, frame geometry,
2Pcrcosθ was taken as the expected minimum design resistance for and A1085 HSS4 × 4 × 5/16 braces, but they employed different beam
these chevron CBFs. In these plots, this resistance using the measured sections to evaluate the effect of beam yielding on frame performance.
yield stress of A1085 HSS4 × 4 × 5/16 braces was used to normalize The final two tests addressed the relative performance of A500 HSS
the computed lateral resistance for the experimentally and numerically braces and beam depth and stiffness. Experimentally validated,
simulated single-story chevrons with reduced beam resistance as a high-resolution finite-element models were used to further evaluate
function of the normalized beam strength in Fig. 14a. The “Beam and extend the experimental results through a parameter study. These
Strength Ratio” is the inverse of the results of beam interaction equation nonlinear computer models simulated the cyclic frame deformations
(see Table 1) when the beam is analyzed for SCBF unbalanced loads cur- up to brace fracture of all frames. The observed local and global response
rently required in design. was compared to the numerically simulated behavior to verify the accu-
The experimental results are plotted as square symbols in the fig- racy of the models.
ures, and they compare well with analytical predictions, which are plot- A number of conclusions were drawn from this research:
ted as lines with circular symbols. The figure shows that the lateral
resistance steadily decreases with decreasing beam strength, and • The inelastic story drift achieved prior to brace fracture was larger for
crosses below the target lateral resistance (1.0 in the figure) between chevron CBFs with beam yielding than for chevron CBFs without beam
Chevrons 3 and 4, which corresponds to a beam resistance that is ap- yielding, and drift capacity increased with increased beam yielding.
proximately 30% of that currently required. • Chevron CBFs achieve larger story drifts prior to brace fracture than
Fig. 14b shows that the maximum beam deflection increases with other SCBF configurations because the chevron beam deflects down-
decreasing beam strength, as expected, however the response is not lin- ward. This deflection increases compressive deformations of the
ear. The simulated results are slightly larger than the experimentally braces but reduces tensile force and elongation of the brace.
measured deflections because the shear-plate connection provides • The maximum lateral resistance of the frame decreases with increased
some rotational resistance. This rotational resistance becomes more sig- beam yielding, but the change in lateral resistance is smaller than the
nificant with smaller beam sizes. Again, there is a large change in the reduction in beam resistance. The lateral resistance remains larger
beam deflection when the beam strength is approximately 30% of than the design lateral resistance through large inelastic deformations
what is currently required. if the beam strength is larger than approximately 30% of the current
Fig. 14c and d shows the maximum axial force and deformation of SCBF requirement.
the braces in tension and compression. These plots clearly show that • Vertical deflections of the beam increase with decreased beam
tensile elongation is significantly decreased and the compressive short- strength, as expected, but the deflections were smaller than antici-
ening is significantly larger as the beam strength is reduced. The maxi- pated given the reduction in beam resistance. Residual deflections
mum tensile force in the brace decreases as beam strength is reduced, after lateral load was removed were typically less than 50 mm
but it stabilizes at a magnitude similar to the compressive buckling (2 in.) or 1/120 of the span length, provided the beam had more
force at very low beam resistance. than 30% of the current SCBF requirement.
All four plots show slight-to-moderate changes in slope as the beam • Braces buckle in compression and sustain similar degradation in com-
strength ratio falls between 0.2 and 0.3. As stated previously, the focus pressive force regardless of beam strength. However, the maximum
of this research has been evaluation of the unbalanced beam forces tensile brace force approaches the magnitude of the brace buckling
due to post-buckling behavior of the braces (see Fig. 1c) and the impact force with weaker chevron beams.
of beam strength on demand and ductility. However, the beam must • There are two design demand limit states (Fig. 1b and c) that must be
still resist the full load case shown in Fig. 1b, and this force is not re- considered in beam design. The results suggest the chevron CBF be-
duced by beam yielding, where the upper bound on the force in the ten- havior will be influenced by axial yielding of chevron beams if the
sile brace is reduced. Further, the full load case becomes relatively beam strength is less than about 25% of current SCBF requirements.
constant as the tensile and compressive brace forces converge for
beams with a low strength relative to current requirements. If the Based on the current work, it appears that the unbalanced load re-
beam is overstressed under this load condition, axial shortening and quirement can be reduced to approximately one-third of the current re-
elongation of the beam and OOP P-δ deformation must be expected. quirement. Limited yielding of the beam increases the inelastic
Such beam deformations were observed during testing of Chevron 4, deformation capacity of chevron CBFs, which is important under the
though they were limited in magnitude. The W14 × 26 had axial stress maximum-considered earthquake loading. This yielding may reduce
levels in the range of 105 to 140 MPa (15 to 20 ksi) at the maximum the resistance, but the resistance remains stable through large deforma-
brace forces, and beams with a Beam Strength Ratio of 0.1 to 0.15 tions. The combined observations of increased inelastic deformation ca-
would have much higher axial stress levels. This high stress is particu- pacity and retained resistance show that chevron CBFs with limited
larly critical in this research program, because the frames had no slab, beam yielding should have no adverse effects on life safety and collapse
which will contribute resistance to these axial forces in an actual build- prevention. Therefore, it is recommended that controlled yielding be
ing. Hence, this beam must be designed accounting for both demand permitted for chevron SCBFs.
limits, with particular attention to the combined loading for each. This research on single-story chevron braced frames shows that
Chevrons 1, 2, 3, and 4 all achieved inelastic story drifts larger than chevron beam yielding could be beneficial to chevron braced frame per-
achievable with other SCBF configurations, which have been tested pre- formance, since it improves the economy and seismic performance
viously by the authors and others. This occurs because the brace with while retaining architectural advantages of the system. However, the
yielding chevron beams does not develop large tensile force and elonga- conclusions from this research must be further qualified. In particular,
tion, but develops larger compressive shortening. Even under low-cycle multi-story chevrons will have very different beam-column connec-
fatigue conditions, brace fracture is driven largely by the tension strain, tions since gusset plates will be required at these locations, and concen-
and hence larger local deformations are required in the brace before tration of inelastic damage to a single story and a wide range of other
fracture occurs with this beam deflection. multi-story system concerns could have significant impact; these issues
have not been addressed here. Further, the composite slab present in
6. Summary and conclusions most applications will provide additional resistance and lateral restraint
to the beam. As a result, experimental and analytical research on multi-
Six large-scale single-story chevron braced frames were tested story chevrons with composite slabs, as well as nonlinear dynamic anal-
under inelastic cyclic loading to large drifts. The first four tests were ysis of seismic response of a wide range of chevron systems, is needed.
C.W. Roeder et al. / Journal of Constructional Steel Research 159 (2019) 428–441 441

These topics are addressed in additional research described in forth- [12] Y. Balazadeh-Minouei, S. Koboevic, R. Tremblay, Seismic assessment of existing steel
Chevron braced frames, J. Struct. Eng. 144 (2018), 04018046. https://doi.org/10.
coming papers. 1061/(ASCE)ST.1943-541X.0002037.
[13] ASCE, Seismic evaluation and retrofit of existing buildings, ASCE/SEI 41-13,
American Society of Civil Engineers, Reston, VA, 2013.
Acknowledgments [14] F. McKenna, M.H. Scott, G.L. Fenves, Nonlinear finite-element analysis software ar-
chitecture using object composition, J. Comput. Civ. Eng. 24 (2010) 95–107,
This research was funded by the American Institute of Steel Con- https://doi.org/10.1061/(ASCE)CP.1943-5487.0000002.
[15] I.F. Khatib, S.A. Mahin, K.S. Pister, Seismic Behavior of Concentrically Braced Steel
struction. Mr. Tom Schlafly was the program manager for this work.
Frames, in: UCB/EERC-88/01, Earthquake Engineering Reserach Center, Berkeley,
James Malley, Rafael Sabelli, and Tom Sabol served as technical advisors CA, 1988.
to the research team. This financial support and technical advice are [16] M. D'Aniello, S. Costanzo, R. Landolfo, The influence of beam stiffness on seismic re-
greatly appreciated and gratefully acknowledged. sponse of chevron concentric bracings, J. Constr. Steel Res. 112 (2015) 305–324,
https://doi.org/10.1016/j.jcsr.2015.05.021.
[17] S. Costanzo, M. D'Aniello, R. Landolfo, Critical review of seismic design criteria for
References chevron concentrically braced frames: the role of the brace-intercepted beam, Ing.
Sismica. 33 (2016) 72–89.
[1] ICBO, Uniform Building Code, International Conference of Building Officials, (Whit- [18] C.W. Roeder, E.J. Lumpkin, D.E. Lehman, A balanced design procedure for special
tier, CA) 1988. concentrically braced frame connections, J. Constr. Steel Res. 67 (2011)
[2] AISC, Seismic provisions for structural steel buildings, ANSI/AISC 341-97, American 1760–1772, https://doi.org/10.1016/j.jcsr.2011.04.016.
Institute of Steel Construction, Chicago, IL, 1997. [19] C. Terpstra, Impact of Beam Strength on Seismic Performance of Chevron Concentri-
[3] AISC, Seismic provisions for structural steel buildings, ANSI/AISC 341–16, American cally Braced Frames, University of Washington, 2017http://hdl.handle.net/1773/
Institute of Steel Construction, Chicago, IL, 2017. 40001.
[4] S. Costanzo, M. D'Aniello, R. Landolfo, Seismic design criteria for Chevron CBFs: [20] S.M. Ibarra, Experimental Investigation of Chevron Special Concentrically Braced
European vs north American codes (Part-1), J. Constr. Steel Res. 135 (2017) Frames with a Yielding Beam Plastic Mechanism, University of Washington, 2018.
83–96, https://doi.org/10.1016/j.jcsr.2017.04.018. [21] C.W. Roeder, D.E. Lehman, J.W. Berman, S.M. Ibarra, C. Terpstra, A.D. Sen, T. Li, R. Liu,
[5] SEAOC, Recommended Lataeral Force Requirements and Commentary, Structural Investigation of Chevron-Configured Special Concentrically Braced Frames: A Report
Engineers Association of California, Sacramento, CA, 1990. to the American Institute of Steel Construction, (Seattle, WA) 2017.
[6] SEAOC, Recommended Lateral Force Requirements and Commentary, Structural En- [22] AISC, Specification for structural steel buildings, ANSI/AISC 360–16, American Insti-
gineers Association of California, Sacramento, CA, 1996. tute of Steel Construction, Chicago, IL, 2017.
[7] D.A. Foutch, S.C. Goel, C.W. Roeder, Seismic testing of full-scale steel building—part I, [23] E.J. Lumpkin, P.-C. Hsiao, C.W. Roeder, D.E. Lehman, C.-Y. Tsai, A.-C. Wu, C.-Y. Wei,
J. Struct. Eng. 113 (1987) 2111–2129, https://doi.org/10.1061/(ASCE)0733-9445 K.-C. Tsai, Investigation of the seismic response of three-story special concentrically
(1987)113:11(2111). braced frames, J. Constr. Steel Res. 77 (2012) 131–144, https://doi.org/10.1016/j.jcsr.
[8] C.W. Roeder, Seismic behavior of concentrically braced frame, J. Struct. Eng. 115 2012.04.003.
(1989) 1837–1856, https://doi.org/10.1061/(ASCE)0733-9445(1989)115:8(1837). [24] D.E. Lehman, C.W. Roeder, D. Herman, S. Johnson, B. Kotulka, Improved seismic per-
[9] T. Okazaki, D.G. Lignos, T. Hikino, K. Kajiwara, Dynamic response of a chevron con- formance of gusset plate connections, J. Struct. Eng. 134 (2008) 890–901, https://
centrically braced frame, J. Struct. Eng. 139 (2013) 515–525, https://doi.org/10. doi.org/10.1061/(ASCE)0733-9445(2008)134:6(890).
1061/(ASCE)ST.1943-541X.0000679. [25] ATC, Guidelines for cyclic seismic testing of components of steel structures for build-
[10] A.D. Sen, C.W. Roeder, J.W. Berman, D.E. Lehman, C.-H. Li, A.-C. Wu, K.-C. Tsai, Experimen- ings, Rep. ATC-24, Applied Technology Council, Redwood City, CA, 1992.
tal investigation of chevron concentrically braced frames with yielding beams, J. Struct. [26] Dassault Systèmes, Abaqus Unified FEA, 2013.
Eng. 142 (2016), 04016123. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001597. [27] J.-H. Yoo, C.W. Roeder, D.E. Lehman, Analytical performance simulation of special
[11] R. Tremblay, N. Robert, Seismic performance of low- and medium-rise chevron concentrically braced frames, J. Struct. Eng. 134 (2008) 881–889, https://doi.org/
braced steel frames, Can. J. Civ. Eng. 28 (2001) 699–714, https://doi.org/10.1139/ 10.1061/(ASCE)0733-9445(2008)134:6(881).
l01-038.

You might also like