Cyclic Loading Test for Beam-Column Connectionsof Concrete-Filled U-Shaped Steel Beams and

Concrete-Encased Steel Angle ColumnsHyeon-Jong Hwang1; Tae-Sung Eom2; Hong-Gun Park, A.M.ASCE3;

Seung-Hwan Lee4; and Hyoung-Seop Kim, P.E.5

Abstract: Prefabricated steel–reinforced concrete angle columns and concrete-filled U-shaped steel beams were recently developedfor efficient steel-concrete composite construction. In the present study, seismic details of TSC beam–PSRC column connections weredeveloped, taking into consideration constructability and cost efficiency. A cyclic loading test was performed on the beam-column connec-tions to investigate load-carrying capacity, deformation capacity, failure mode, and energy dissipation capacity. For the test parameters, theconnection type (interior or exterior) and the depth of the TSC beams were considered. The test results showed that the deformation andenergy dissipation capacities of the specimens satisfied the requirements for intermediate moment frames specified in the AISC standard.Further, the moment-carrying capacities predicted using plastic stress distribution were found to be in agreement with the test results. Thejoint shear capacities of the specimens were evaluated according to ASCE design guidelines. DOI: 10.1061/(ASCE)ST.1943-541X.0001242. © 2015 American Society of Civil Engineers.

Author keywords: Composite structure; Composite beams and columns; Beam-column connection; Cyclic test; Seismic performance;Metal and composite structures.

Introduction

In conventional concrete–steel composite construction, a wide-flange steel section encased in reinforced concrete is used forcolumns and awide-flange steel section, connected to a concrete slabwith shear studs, is used for beams. Recently, to improve structuralcapacity and cost efficiency, various types of composite columns andbeams have been explored. Fig. 1(a) shows a example of a prefab-ricated steel–reinforced concrete (PSRC) column used in construc-tion. In the PSRC column, four steel angles are placed at the cornersof the cross section and areweld-connected to transverse reinforcingbars or band plates. A PSRC column has two major advantages overa conventional concrete-encased steel (CES) column. First, becausethe steel sections and the reinforcing bars are prefabricated, rebarwork is unnecessary in the field. Second, the flexural strength

and stiffness of the PSRC column are increased under biaxialmoments by angles at the corners of the cross section. Montuoriand Piluso (2009), Campione (2010), Hwang et al. (2012), Kim et al.(2014), and Eom et al. (2014) have studied PSRC columns.

Fig. 1(b) shows an example of novel composite beams. TheT-type steel concrete (TSC) beam consists of a U-shaped steelsection (U-section) and infilled concrete. The U-section providesflexural and shear capacities to the beam, creating a form for theconcrete during construction. The infilled concrete can significantlyreduce deflection and floor vibration by increasing flexural stiffnessand prevents early local buckling of the U-section’s thin web andflange plates. It can also enhance fire resistance by providing a heatsink. Originally, Oehlers (1993), Oehlers et al. (1994), and Uy andBradford (1995a, b) proposed composite profiled beams that usesteel profiled sheets at the sides and bottom. Through experimentaland analytical studies, they evaluated the flexural and shearstrengths ductility, and long-term deflection of these beams.

Kim et al. (2006), Park et al. (2012), and Lee et al. (2013) per-formed cyclic loading tests on beam-column connections usingTSC beams. Kim et al. (2006) tested the seismic performance ofTSC beam–CES column connections; their results showed thatthe infilled concrete effectively prevented early local bucklingof the web and flange plates of the U-section and found damagein the beam-column joints to be insignificant. Excellent perfor-mance was thus achieved in compliance with the AISC 341-05specification for special moment frames (AISC 2005). Park et al.(2012) performed cyclic loading tests on TSC beam–RC columnconnections. Their specimens exhibited excellent ductility and en-ergy dissipation capacity, with load-carrying capacity degraded at a5% drift ratio because of local buckling of the web and flange platesin the TSC beam. Joint shear strength was evaluated by modifyingthe provisions of the ASCE design guidelines (ASCE 1994). Leeet al. (2013) also investigated the seismic performance of TSCbeam–steel column connections. Using special details to relocatethe plastic hinge of the TSC beams away from the column face,

1Assistant Professor, College of Civil Engineering, Hunan Univ., YueluMountain, Changsha, Hunan 410082, China; formerly, PostdoctoralResearcher, Dept. of Architecture and Architectural Engineering, SeoulNational Univ., 599 Gwanak-ro, Seoul 151-744, Korea. E-mail:hwanggun85@naver.com

2Assistant Professor, Dept. of Architectural Engineering, DankookUniv., 152 Jukjeon-ro, Gyeonggi-do 448-701, Korea (correspondingauthor). E-mail: tseom@dankook.ac.kr

3Professor, Dept. of Architecture and Architectural Engineering, SeoulNational Univ., 599 Gwanak-ro, Seoul 151-744, Korea. E-mail:parkhg@snu.ac.kr

4President, SEN Coretech, Inc., 6 Beodeunaru-ro 19 gil,Yeongdeungpo-gu, Seoul 150-040, Korea. E-mail: shlee@senkuzo.com

5Deputy Manager, SEN Structural Engineers Co., Ltd, 121-74,Dangsan-dong 6-ga, Yeongdeungpo-gu, Seoul 150-808, Korea. E-mail:hskim@senkuzo.com

Note. This manuscript was submitted on June 5, 2014; approved onDecember 9, 2014; published online on February 10, 2015. Discussion per-iod open until July 10, 2015; separate discussions must be submitted forindividual papers. This paper is part of the Journal of Structural Engineer-ing, © ASCE, ISSN 0733-9445/04015020(12)/$25.00.

© ASCE 04015020-1 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

they minimized damage in the joints, increasing deformationcapacity to the level of the special moment frame.

In the present study, the detailing method and seismic perfor-mance of the TSC beam–PSRC column connection were studied.New connection details were developed to facilitate fabrication ofthe beam-column joint. A cyclic loading test was performed for twofull-scale interior connections and an exterior connection to inves-tigate the load-carrying capacity, deformation capacity, and energydissipation capacity of the structure.

When compared with TSC beam–CES column, TSC beam–RCcolumn, and TSC beam–steel column connections tested in pre-vious studies, the following aspects were considered for the devel-opment of the TSC beam–PSRC column connection in this study:• In TSC beam–CES column or TSC beam–steel column connec-

tions, the U-section of the beam is directly welded to the steelcolumn so that the full moment capacity of the beam can betransmitted to the column. In a TSC beam–PSRC column con-nection, the U-section of the beam is not directly connected tothe corner steel angles. Thus, relatively complex connection de-tails are required, and so the degree of the moment transfer be-tween the TSC beam and the PSRC column should be evaluated.

• According to Eom et al. (2014), the PSRC column is susceptibleto early cover spalling as a result of the weak bond between thesteel angles and the cover concrete, even when experiencing mod-erate inelastic deformations. Such an early fracture of the coverconcrete may deteriorate the connection’s strength and deforma-tion capacity. Thus, the detrimental effect of premature cover con-crete spalling on connection behavior should be evaluated.

• The U-section, steel angles, reinforcing bars, and band plates areassembled together in the TSC beam–PSRC column joint. As aresult of the complexity of the connection details, the shear trans-fer mechanism of the beam-column joint should be clarified.

Connection Details

Fig. 2(a) shows the proposed details of an interior TSC beam–PSRC column connection. The PSRC column is fabricated bywelding four steel angles, four longitudinal bars, and transversebars. The U-sections of the TSC beam brackets are then framed tothe joint. Band plates and wing plates, shown in the figure, connectthe U-sections directly to the column angles. The TSC beam isconnected to the bracket using slip critical bolts. In the joint area,as shown, the web plates of the U-section are continuous through

the joint. However, the section’s top and bottom flanges are cut offat the face of the PSRC column for convenience during fabricationand for better concrete placement. Despite the discontinuous topand bottom flanges, the majority of the required moment can beresisted because the majority of the negative moment is resistedby the flexural rebars placed in the slab, and the positive designmoment is relatively small because of the effect of the gravity load.However, because the flanges of the U-section are cut off, brittlefailure due to stress concentration may occur at the joint. Such anoccurrence was evaluated in this test.

At the top and bottom of the U-sections, eight band plates arewelded to the column angles [Fig. 2(a)]. They act as temporarysupports for the TSC beams during construction. After concreteplacement, the band plates contribute to joint shear capacity and toconfinement of the concrete. The bottom flange of the section is notwelded to the bottom band plate, to prevent pullout failure of theband plate and early spalling of the cover concrete under cyclicloading. The top flange is welded to the top band plate as a safetymeasure for integrity under construction.

Eight wing plates are used to connect the web plates of theU-section to the column angles so that the shear of the TSC beamsis transferred to the PSRC column [Fig. 2(a)]. Fig. 2(b) showsthe details of an exterior TSC beam–PSRC column connection.Basically, the connection details are the same as those for aninterior connection. However, in the exterior face of the joint, largerband plates and additional transverse bars are used to enhance jointshear capacity because the exterior face of the connection is notconfined by any beam.

Fig. 3 illustrates the plastic stress distributions of the concreteand steel in order to calculate the plastic moment capacity of theTSC beam. Because the flange of the U-section is discontinuous inthe joint, only compressive stress is developed in it. In the web platethat is continuous within the joint, yield strength can be developedfor both tension and compression. For the flexural bars in the slab,the tensile yield strength is used for the negative moment. However,the compressive stress of the slab reinforcement for the positivemoment is neglected, as recommended in AISC 360-10 (AISC2010) and Eurocode 4 (EC4) (BS 2005).

Test Program

Test Specimens

Fig. 4 shows three full-scale TSC beam–PSRC column connectionspecimens in detail. Two cruciform interior connection specimens,I450 and I550, and a T-shaped exterior connection specimen, E550,were prepared for the test (“450” and “550” indicate the depth ofthe U-section in the TSC beam). These connections were designedaccording to AISC 360-10 (AISC 2010) and KBC 2009 [Architec-tural Institute of Kore (AIK) 2009]. To ensure the strong column–weak beam design concept, the sum of the upper and lower columnflexural capacities, ΣMnc, was greater than the sum of the leftand right beam flexural capacities, ΣMnb. The properties of thespecimens, including flexural moment capacity, are presented inTable 1.

Fig. 4(a) shows the configurations, reinforcement details, andsectional properties of specimen I550. The net height of the PSRCcolumn from the bottom pin support to the top loading point was3,060 mm. The net length of the TSC beam between the end sup-ports was 6,760 mm. The beams were spliced at 1,000 mm from thecolumn face using 50 M22 slip critical bolts (area of each bolt ¼380 mm2; design shear strength ¼ 83.6 kN). They were framedinto the beam-column joint from the four faces.

Fig. 1. (a) PSRC composite column; (b) TSC composite beam

© ASCE 04015020-2 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

The cross section of the PSRC column was 800 × 800 mm [seethe B-B section of Fig. 4(a)]. Four L-130 × 130 × 12 ðmmÞ steelangles (cross-sectional area ¼ 2,976 mm2 each; yield strength ¼394 MPa) were used at the corners of the cross section, and fourD32 bars (deformed bar diameter ¼ 32.3 mm; cross-sectionalarea ¼ 819 mm2 each; yield strength ¼ 550 MPa) were placed atthe center of the four column edges. The total steel ratio was 2.37%,including the four angles and the four D32 bars (total steel area ¼640,000 mm2). D13 bars (deformed bar diameter ¼ 12.7 mm;cross-sectional area ¼ 129 mm2 each; yield strength ¼ 519 MPa)

were used for rectangular hoops and crossties (i.e., intermediate tieswith 90° and 135° hooks at the ends) [Fig. 1(a)]. The vertical spac-ing was s ¼ 150 mm. Each rectangular hoop was fabricated bywelding four 600-mm D13 bars to the surface of the column angles(weld length lw ¼ 100 mm); the bar weld details were taken fromEom et al. (2014). Two crossties were used to enhance the shearcapacity and concrete confinement of the PSRC column. The D13hoop bars were essential to secure the bond (i.e., the shear transfer)between the steel angles and the surrounding concrete. The bondcapacity of the hoop bars welded to the steel plates was investigated

Fig. 2. Details of TSC beam–PSRC column connections: (a) interior beam-column joint; (b) exterior beam-column joint

(a) (b)

Fig. 3. Plastic stress distributions of a TSC beam cross section at the column face: (a) positive moment; (b) negative moment

© ASCE 04015020-3 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

by Eom et al. (2014), who determined that the spacing, s ¼150 mm, and weld length, lw ¼ 100 mm, of the hoops gaveadequate shear transfer.

Fig. 4(a) also shows the cross section of the TSC beam (see theA-A section). The depth and width of the U-section were 550 mmand 300 mm, respectively, and the thickness of the plate was 6 mm.The thickness and width of the concrete slab were 150 mm and1,200 mm, respectively. The width of the slab was determined tobe less than one-fourth of the overall beam length of 6,760 mm,in accordance with AISC 360-10. The U-section of the beam wasfabricated using two L-shaped plates (i.e., a top flange and a web)and a 330-mm-wide bottom plate. The yield strength of the plateobtained from the material test was 450 MPa. For the ties to beable to prevent possible lateral deformations of the thin web platesduring concrete placement, L-50 × 50 × 6 angles were welded tothe top flanges at a spacing of 600 mm. To develop a full compositeaction with the concrete slab, ϕ16 shear studs (cross-sectionalarea ¼ 201 mm2 each; tensile strength ¼ 440 MPa) were used atthe top flange plates of the beam at a spacing of 125 mm [see the

top view and the A-A section in Fig. 4(a)]. Shear studs were notused at the web and bottom flange plates.

Four D25 bars (deformed bar diameter ¼ 25.4 mm; cross-sectional areaof abar¼ 510 mm2; yield strength¼ 460MPa) wereplaced in the concrete slab through the joint; the longitudinal barswere used to enhance the negative moment capacity of the TSCbeam. Four D10 bars (deformed bar diameter ¼ 9.5 mm; cross-sectional area ¼ 71 mm2 each; yield strength ¼ 524 MPa) wereadded for slab reinforcement; they were used at a spacing of200 mm for transverse reinforcement. The longitudinal D25 andD10 bars were placed beneath the transverse reinforcement so thatthe spalling of the cover concrete could be restrained.

Fig. 4(a) details the connection between the TSC beams and thePSRC columns [Fig. 2(a)]. Eight plates of 580 × 70 × 12 mm(cross-sectional area ¼ 840 mm2; yield strength ¼ 487 MPa) wereused for the band plates. The U-section of the beams was mountedon the bottom band plates without welding, whereas the top flangewas welded to the band plates. The band plates contributed to theshear capacity of the beam-column joint, as discussed in the section

Fig. 4. Dimensions and details of specimens: (a) interior beam-column joint (I550); (b) exterior beam-column joint (E550)

© ASCE 04015020-4 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Evaluation of Joint Shear Strength. Eight plates of 300 × 100 ×6 mm (yield strength ¼ 450 MPa) were used for the wing platesthat connected the web plates of the U-section to the column angles.

The details of the I450 specimen were the same as those for I550specimen except for the depth of the TSC beam. In the I450 speci-men, the web depth was reduced to 450 mm to investigate the effectof the beam’s flexural capacity. Generally, connection behaviorsuch as load-carrying capacity and ductility is significantly affectedby beam depth because the joint shear force and flexural momentincrease proportionally with it.

Fig. 4(b) details the exterior connection specimen, E550. Thenet height of the PSRC column was 3,060 mm and the net lengthof the TSC beam from the column face to the pin support was2,980 mm. The cross sections of the column and the beam werethe same as those of the I550 specimen [Fig. 4(a)]. Four D25 barswere placed in the concrete slab through the column and wereanchored by 90° standard hooks (ACI 318-11). Four D10 longitu-dinal bars were used for longitudinal reinforcement of the slabwithout an anchorage hook [see the slab rebars in Fig. 4(b)]. Theconnection details for the E550 specimen are shown in Fig. 4(b).The band plates (580 × 70 × 12 mm) and wing plates (300 × 100×6 mm) were used for the three joint faces where the beams wereframed. In the exterior face, the dimensions of the band plates

were increased to 580 × 140 × 12 mm, and three D13 bars wereused for transverse reinforcement at a spacing of 100 mm [referto Fig. 2(b)].

Table 2 presents the yield strengths (Fy, fy), tensile strengths(Fu, fu), and elongations at the fracture (εfr) of the steel sections,plates, and reinforcing bars used in the specimens. The mean valuesfor the material properties were obtained from the three directtension tests; the welding methods used for the fabrication of thespecimens are shown in Fig. 4. in In accordance with ANSI/AWSD1.4 [American Welding Society (AWS) 1998], YFW-C50DRelectrodes (FEXX ¼ 584 MPa and CVN86 J@0°C) were used forflare-bevel-groove welding between the angles and the transversebars and for fillet welding of the angles, U-sections, band plates,and wing plates. Concrete strengths were 30.8 MPa, 31.1 MPa, and30.1 MPa in specimens I450, I550, and E550, respectively.

Testing Method

Fig. 5 shows the set-up for the cyclic loading test. An actuator(maximum loading capacity ¼ 2,000 kN; maximum stroke ¼�250 mm) was used at the top of the PSRC column. The loadingprogram was planned in accordance with AISC 341 (AISC 2005).Load cycles were repeated six times at the story drift ratios

Table 1. Properties of Test Specimens and Materials

Specimens I450 I550 E550

Beam type TSC beamSlab flexural bars 4-D25þ 4-D10Width of U-steel section (mm) 300 300 300Depth of U-steel section (mm) 450 550 550Thickness of plate in U-steel section (mm) 6 6 6Area of U-steel section (mm2) 5,256 6,456 6,456Slab width (mm) 1,200 1,200 1,200Slab thickness (mm) 150 150 150Stud spacing (mm) 2-Ø16@150 2-Ø16@125 2-Ø16@125Positive moment capacity Mþ

pa (kN · m) 811 1,119 1,115

Negative moment capacity M–pa (kN · m) 893 1,188 1,184

Column type PSRC columnSection dimensions (mm) 800 × 800

Longitudinal rebars 4-D32Steel angles (mm) L-130 × 130 × 12

Axial capacitya (kN) 22,209 22,365 21,845Moment capacitya (kN · m) 1,986 1,988 1,982Column–Beam flexural strength ratio 2.33 1.72 3.35aPlastic stress distributions and measured material strengths were used for calculation of axial and flexural capacities.

Table 2. Material Test Results

Plates and reinforcing bars Yield strength Fy or fy (MPa) Tensile strength Fu or fu (MPa) Elongation (%)

Angle section of PSRC column (SM490, 12 t)a 394 552 26.0U-section of TSC beam (SM490, 6 t)a 450 515 17.2Band plate (SM490, 12 t)a 487 541 21.5Wing plate (SM490, 6 t)a 450 515 17.2D32 (SD500 W)b 550 650 25.3D25 (SD400)b 460 613 25.3D13 (SD400 W)b 519 623 20.0D10 (SD400)b 524 650 18.7Concrete strength f 0

c 30.8 (I450), 31.1 (I550), and 30.1 (E550)aStructural steel for welding, complying with Korean Industrial Standard. Design yield and tensile strengths of SM490 were 325 MPa and 490 MPa,respectively.bReinforcing steel bars complying with Korean Industrial Standard (“400” and “500” indicate design yield strength in MPa. “W” indicates weldable steel.

© ASCE 04015020-5 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

δ ¼ �0.375, �0.5, and �0.75%; four times at δ ¼ �1.0%; andtwo times at δ ¼ �1.5, �2.0, �3.0, �4.0, and �5.0%. An axialload was not applied to the PSRC column because it was not fea-sible in the laboratory to apply a very large axial force that wasmeaningful to the full-scale test specimen. The column was pin-supported at the bottom; at the roller supports of the TSC beams,load cells were used to measure the vertical reactions. The cross-beams were not supported at the ends. The concrete slab waslaterally supported to prevent out-of-plane displacement. Linearvariable differential transformer (LVDTs) were used to measurethe horizontal displacement of the column and the slip deforma-tions at the column and beam supports. The steel section strainsand those of the reinforcing bars were measured during the testsusing uniaxial strain gauges.

Test Results

Lateral Load–Story Drift Ratio Relationship

Fig. 6 shows the lateral load–story drift ratio (P-δ) relationshipsof the specimens. The story drift ratio δ was calculated by divid-ing the lateral displacement at the loading point by the columnheight (3,060 mm). In the interior connection of the I450 specimen[Fig. 6(a)], the peak strengths Pþ

u and P−u occurred at δ ¼ �2.0%.

The load-carrying capacity was significantly degraded in thesecond load cycle at δ ¼ �3.0%. The damage modes of theI450 specimen at the end of the test are shown in Fig. 7(a). At δ ¼�3.0%, cover concrete spalling in the PSRC column occurredin the vicinity of the U-section, whose web plate buckled duringthe second load cycle. The spalling and buckling significantly de-creased load-carrying capacity, which was further degraded duringthe second load cycle with δ ¼ �4.0% because of a tensile fracture

in the buckled web plate [Fig. 7(a)]. The fracture was initiated at theweld joint between the web and the bottom flange and then propa-gated vertically.

Because the band plates were not connected to the bottom flangeof the U-section, the band plates were not damaged even after coverconcrete spalling. However, a gap occurred between the end ofthe U-section and the column face under positive moments, whichwas attributed to an anchorage slip of the web plates that wereexperiencing plastic strains inside the joint. As the anchorage slipincreased, development of the beam’s flexural capacities was de-layed. For this reason, there was significant pinching in the cyclicbehavior. The top band plates welded to the top flange of theU-section were pulled out under a negative moment. As a result,as shown in Fig. 7(a), significant vertical cracks occurred in theconcrete slab. The wing plates welded to the web of the U-sectionwere also pulled out, causing vertical cracks in the joint interface.

Figs. 6(b) and 7(b) show the test results for the I550 specimen.The load–drift ratio relationship and damage modes were similarto those for I450. However, because of a greater beam depth(550 mm), the shear demand for the beam-column joint signifi-cantly increased. As a result, cover concrete spalling was moresignificant and severe diagonal cracking occurred in the joint faceconnected to the beam, particularly at the bottom of the beam. Thecyclic behavior of I550, including strength degradation and pinch-ing, was similar to that of I450.

Figs. 6(c) and 7(c) show the test results for the E550 specimen.The load–drift ratio relationship and the damage modes weresimilar to those of I450 and I550. The peak strengths Pþ

u and P−u

developed at δ ¼ þ2.0 and −2.5%. After the peak strengths,the load-carrying capacity was significantly degraded because ofcover concrete spalling in the PSRC column and because of localbuckling and fracture of the web in the TSC beam. As shown in

Fig. 5. Test set-up and measurement: (a) interior connections (I450, I550); (b) exterior connections (E550)

Fig. 6. Lateral load–story drift relationships: (a) I450; (b) I550; (c) E550

© ASCE 04015020-6 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Fig. 7(c), the steel angles of the column were completely exposedbecause of the spalling at δ ¼ �4.0%. If the transverse bars hadbeen insufficient and a high axial load had been applied, local buck-ling might have occurred in the steel angles after spalling.

Beam Reactions

Fig. 8 shows the vertical reactions at the end supports of the TSCbeam. The reactions were measured from the load cells installed atthe beam ends (see LC1 and LC2 in Fig. 4) to verify the positiveand negative flexural moment capacities of the TSC compositesection using complex connection details in the joint. The negativeand positive values denote downward and upward reactions, re-spectively. Thus, negative and positive reactions were produced bythe negative and positive moments of the beams, respectively. Fig. 8shows that the beam reaction–story drift ratio relationships werevery similar to those for the P-δ relationships shown in Fig. 6.The negative reactions were 8–29% greater than the positive reac-tions, indicating that the negative moment capacities of the TSC

beams for I450, I550, and E550 were 8–29% greater than thepositive moment capacities.

Steel Plate Strains

Figs. 9(a and b) show the web and flange strains of the U-sectionin the E550 specimen. The vertical and horizontal axes indicate thesteel strains and the story drift ratios, respectively. The web strainwas measured at the bottom of the web, 225 mm from the columnface; the flange strain was measured at the bottom flange, also225 mm from the column face. As shown in Fig. 2, the bottomflange of the U-section was cut off at the column face whereas theweb was connected to the joint. However, in Fig. 9 both flange andweb strains were less than the yield strain of 0.00225. This indi-cates that, although web yielded inside the joint, its tensile stressquickly decreased because of the effect of the flanges as the dis-tance from the column face increased. In Fig. 9(b), as a result ofthe flange plate cut-off, the flange plate’s peak tensile strain wasonly half of the web strain [Fig. 9(a)] and was less than the peak

Fig. 7. Damage modes at the end of tests: (a) I450; (b) I550; (c) E550

Fig. 8. Vertical reactions at beam ends: (a) I450; (b) I550; (c) E550

© ASCE 04015020-7 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

compressive strain. The results for I450 and I550 were similar tothose for E550.

Evaluation of Structural Performance

Load-Carrying Capacity

The load-carrying capacity of the specimens was predicted as fol-lows. First, the TSC beam’s positive and negative plastic momentcapacities, Mþ

p and M–p, were calculated from the plastic stress dis-

tributions, as shown in Fig. 3. The specimens’ positive and negativestrengths, Pþ

n and P–n, were then calculated from the static equilib-

rium between the lateral load and the support reactions as follows:

Pþn orP−

n ¼ Mþp þM−

p

L − hc

�LH

�ð1Þ

where L = length of the TSC beam between the end supports;H = net column height; and hc = column depth.

Fig. 6 compares the predicted strengths, Pþn and P–

n, with thetest results, Pþ

u and P–u.

Fig. 8 compares the predicted beam reactions, Rþn (or R–

n)[= Mþ

p ðorM−p Þ=Ls, Ls ¼ 2,980 mm] with the test results. As

shown in Figs. 6(a and b), Pþn and P–

n of the interior connections(I450 and I550) were in agreement with the test strengths, Pþ

u andP–u, showing a safety margin of about 5%. In the exterior connec-

tion of E550 [Fig. 6(c)], the positive test strength Pþu was equivalent

to Pþn whereas the negative test strength P–

u was 13% greater thanP–n. This result was also observed in the beam reactions [Fig. 8(c)].

Even for I450 and I550, the negative test strengths R–u were about

12% greater than R–n [Figs. 8(a and b)] because the contribution of

the top flanges welded to the band plates was not considered in thecalculation of negative plastic moments.

Deformation Capacity

Deformation capacity δu (or maximum story drift ratio) wasdefined as the postpeak story drift ratio corresponding to 80%of the nominal strength, 0.8Pþ

n and 0.8P–n (AISC 2005). In

Figs. 6(a and b), the deformation capacity of specimens I550 andI450 was δu ¼ 4.0%, satisfying the minimum requirements for spe-cial moment frames specified in AISC 341 (AISC 2005). For E550,on the other hand, the deformation capacity was δu ¼ þ3.0% forthe positive loading and −4.0% for the negative loading. This waslower than the requirement for special moment frames, but satisfiedthe requirement for intermediate moment frames (AISC 2005).

Local Buckling of Steel Plates

Local buckling of the thin steel plates in the U-section was evalu-ated following Oehlers et al. (1994). Because of the restraint pro-vided by the core concrete, local buckling occurred outward fromthe concrete and the fold lines of the plates acted as fixed supports.The b=t ratio of a steel plate corresponding to local buckling isdefined as follows:

bt¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffikπ2Es

12Fyð1 − ν2Þ

sð2Þ

where b and t = width and thickness of the steel plate, respectively;k ¼ 10.67 for the restrained steel plate; Es = elastic modulus of thesteel (205 GPa); Fy = yield strength of the steel; and ν = Poisson’sratio of the steel (= 0.3).

For the web plates in the U-section, the b=t ratio was estimatedas 66.3, which was less than the actual b=t ratio of the web platesused for specimens I450, I550, and E550 (75.0, 91.7, and 91.7).This result indicates that local buckling occurred in the web plates(Fig. 7). However, the b=t ratio and k values in Eq. (2) correspondto steel plates subjected to uniformly distributed stresses. Thus, forweb plates subjected to linearly distributed stresses, Eq. (2) under-estimates the b=t ratio. For this reason, local buckling of steel platesin the test specimens occurred only at large inelastic deformationsafter the peak strength (Fig. 6).

For the bottom flange in the U-section, the actual b=t ratio was50.0, which was less than the required b=t ratio of 66.3. Thus, localbuckling did not occur in the bottom flange.

Energy Dissipation Capacity

Fig. 10 compares the energy dissipation capacity of the threespecimens. As shown in Fig. 10(a), the energy dissipation per loadcycle, ED, defined as the area enclosed by a complete load cycle,was calculated for the second load cycle at each story drift ratio toconsider cyclic strength degradation. The ED values for I550 withthe greater beam depth were 45% greater than those for I450 withthe lower beam depth. Those for the exterior connection of E550were approximately half of those for the interior connection ofI550. The ED values increased proportionally to the story drift ratiountil reaching δ ¼ 3.0%. Then they decreased as a result of strengthdegradation.

Fig. 10(b) shows the energy dissipation ratio κ at each driftlevel. This ratio was defined as the ratio of the actual energy dis-sipation ED per load cycle to the idealized elastic–perfectly plasticenergy dissipation Eep∶κ ¼ ED=Eep. The κ values were calculated

Fig. 9. Strains of web and flange of U-steel section in E550: (a) web; (b) bottom flange

© ASCE 04015020-8 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

for the second load cycle at each drift level. Because all specimensexperienced severe pinching in cyclic behaviors, as shown in Fig. 6,the κ values were significantly less than the κ ¼ 0.5–0.6 of con-ventional composite connections using steel beams (Xuemei andGustavo 2004; Cheng and Chen 2005). A notable distinction be-tween the κ values for I550 and I450 with different beam depthswas not evident. On the other hand, E550, when subjected to lessjoint shear, exhibited 17% greater κ values on average.

Discussion

In the proposed method, the flanges of the U-section were in-tentionally not connected to the joint to achieve better concreteplacement and to improve convenience during fabrication. Thediscontinuity of the flange plates was expected to cause a local frac-ture in the web plate because of the high concentration of stress atthe joint interface. However, the test results showed that the anchor-age slip of the web plate inside the joint actually alleviated thestress concentration. For this reason, the maximum deformationof the test specimens reached the level of intermediate momentframes. However, the slip caused the energy dissipation capacityto decrease.

Axial load was not applied to the PSRC columns in the presentstudy. Axial load has both positive and negative effects on connec-tion behavior. It is beneficial in decreasing diagonal shear crackingin the beam-column joint, but it accelerates spalling of the coverconcrete in the joint as well. Thus, degradation of connectionstrength may occur earlier. However, in this test the specimensfailed at the end of the beams (i.e., yielding, anchorage slip, buck-ling, fracture of web plates) rather than in the beam-column jointpanel. For this reason, the effect of axial load on the test specimensis expected to be limited.

Shear Capacity of Beam-Column Joint

Based on the test results, the joint shear strength of the specimenswas evaluated by modifying the existing methods proposed bythe ASCE Task Committee (ASCE 1994) and by Park et al.(2012). Fig. 11 shows the three mechanisms that contribute to jointshear strength: web shear yielding of the U-section [Fig. 11(a)],direct strut action of the infilled concrete inside the U-section[Fig. 11(b)], and strut-and-tie action between the concrete (outsidethe U-section) and the band plates [Fig. 11(c)]. In a U-section, theshear strength, Vws, of the two web plates can be calculated asfollows [Fig. 11(a)]:

Vws ¼ ð2Þ × 0.6Fy tw hwj ð3Þ

where Fy = yield strength of the web; tw = web thickness; and hwj =effective horizontal length of the web in the joint in the direction ofthe shear.

For the interior joint [Fig. 2(a)], the distance between the wingplates can be regarded as hwj because the wing plates welded to thewebs act as vertical stiffeners. For the exterior joint [Fig. 2(b)],however, hwj should be decreased to the distance between the wingplate and the web of the cross beam.

The shear strength, Vcs, of the direct concrete strut can becalculated as follows [Fig. 11(b)]:

Vcs ¼ 1.7ffiffiffiffiffif 0c

pbic hc ≤ 0.5f 0

cbichb for the interior × joint ð4aÞ

Vcs ¼ 1.2ffiffiffiffiffif 0c

pbichc ≤ 0.5f 0

cbichb for the exterior × joint ð4bÞ

where f 0c = concrete strength in MPa; bic = width of the infilled

concrete (i.e., the inner distance between the two webs ofthe U-section); hc = depth of the PSRC column; and hb = overalldepth of the TSC beam, including the concrete slab.

The concrete shear strengths for the interior joint (1.7pf 0c) and

the upper limit (0.5f 0cbichb) in Eq. (4a) were defined according to

ASCE Task Committee provisions (ASCE 1994). The concreteshear strength of Eq. (4b), 1.2

pf 0c, was adopted from ACI 318-11

for an exterior joint with three sides confined.The shear strength, Vst, of the strut-and-tie action can be approx-

imately calculated as follows [ASCE 1994; Fig. 11(c)]:

Vst ¼ 0.4ffiffiffiffiffif 0c

pboho ð5Þ

where bo = effective width of the joint concrete contributing thestrut-and-tie mechanism; and ho = effective length of the joint con-crete in the direction of the shear.

As shown in Fig. 11(c), the diagonal compression of the con-crete strut is resisted by the top and bottom band plates. Thus, thedistance between the angles that can be used to find hobo is definedas follows [Fig. 11(c)]:

bo ¼ baf − bic − 2tw ð6Þ

where baf = distance between the corner angles in the directionorthogonal to the shear; and bic and tw = widths of the infilled con-crete and the web of the U-section, respectively.

Test results showed that the PSRC columns were prone to coverconcrete spalling. Therefore, the cover concrete was not includedin the width of the joint concrete contributing to the strut-and-tiemechanism. The shear strength of Eq. (5) was applicable to both theinterior and the exterior joints.

Fig. 10. Energy dissipation per load cycle

© ASCE 04015020-9 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

According to the ASCE Task Committee (ASCE 1994) andPark et al. (2012), the contribution of the strut-and-tie mechanismcan be increased by providing transverse reinforcement in thejoint. However, the use of transverse reinforcement in the joint areais limited because of the complexity of the details in the joint(Fig. 2). Therefore, the contribution of transverse bars was ignoredin Eq. (5). For the strut-and-tie mechanism in that equation, the

tension force caused by the concrete strut should be resisted bythe top and bottom band plates. The required cross-sectional area,Abp, of each band plate is thus calculated as follows:

Abp ≥ Vst

2Fybpð7Þ

where 2Fybp = yield strength of the top and bottom band plates.According to the ASCE Task Committee (ASCE 1994), joint

shear capacity and demand can be expressed as moments fromthe static moment equilibrium at the joint [Fig. 11(d)]X

Muc −X

Vbho=2 ≤ ϕMnc ð8aÞ

XMuc ¼

XMbp þ

XVbhc=2 − Vchb ð8bÞ

Mnc ¼ Vwsdw þ Vcsð0.75hbÞ þ Vstd ð8cÞ

where (ΣMuc–ΣVbho=2) andMnc = joint shear demand and capac-ity, respectively; ΣMpb and ΣVb = sums of the plastic momentsand shear demands, respectively, of the TSC beams framing intothe joint [in the calculation of ΣMpb and ΣVb, the moment andshear directions shown in Fig. 11(e) are presumed to be positive];Vc = column shear demand; dw = center-to-center distance betweenthe top and bottom flanges; 0.75hb = approximate moment armfor the TSC beam section; and d = center-to-center distance be-tween the top and bottom band plates.

Table 3 lists joint shear demands and capacities for theI450, I550, and E550 specimens—(ΣMuc–ΣVbho=2), Vwsdw,Vcsð0.75hbÞ, Vstd, and Mnc—that were calculated fromEqs. (3)–(8). The joint shear capacities were provided primarilyby the web shear yielding (Vwsdw) and the direct strut mechanism(0.75Vcshb). Either web shear yielding or the direct strut mecha-nism accounted for approximately 39.9–44.2% of the total jointshear capacity. The strut-and-tie action related to the band platescontributed only 13.0–17.3% of the total joint shear capacities.

(a)

(b)

(c)

(d)

(e)

Fig. 11. Evaluation of joint shear strength: (a) web shear yielding(Vws); (b) concrete diagonal strut (Vcs); (c) strut-and-tie action (Vst);(d) effective width for interior joint (bo); (e) joint shear demand

Table 3. Joint Shear Strength (kN · m and %)

Shear strengthI450

(kN · m)I550

(kN · m)E550

(kN · m)

Shear-moment demand,ΣMuc–ΣVbho=2

a1,704 2,307 1,184

Shear-moment capacity byweb plate, Vwsdw (ratio)b

965 (43.2) 1,185 (44.2) 854 (42.8)

Shear-moment capacityby diagonal strut,Vcs (0.75hb), (ratio)

c

978 (43.8) 1,147 (42.7) 796 (39.9)

Shear-moment capacityby compression field,Vstd (ratio)d

291 (13.0) 350 (13.1) 345 (17.3)

Shear-momentcapacity, Mnc

2,235 2,683 1,995

Capacity/demand,Mnc=Muc

1.64 1.53 2.22

aPMp ¼ 1,704, 2,307, and 1,184 kN · m for I450, I550, and E550,respectively;

PVb ¼ 572, 774, 397 kN (I450, I550, E550); and Vc ¼ 632,

855, 439 kN (I450, I550, E550).bFy ¼ 450 MPa; tw ¼ 2 × 6 mm; hs ¼ 680, 680, 490 mm (I450, I550,E550); and dw ¼ 438, 538, 538 mm (I450, I550, E550).cf 0

c ¼ 30.8, 31.1, 30.1 MPa (I450, I550, E550); bic ¼ 288 mm; h ¼800 mm; and hb ¼ 600, 700, 700 mm (I450, I550, E550).dbaf ¼ 680 mm; tw ¼ 6 mm; and d ¼ 508, 608, 608 mm (I450,I550, E550).

© ASCE 04015020-10 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

The ratio of joint shear capacity to demand, Mnc=Muc, of I450,I550, and E550 were 1.64, 1.53, and 2.22, respectively. Thus, thejoint shear failure was expected not to occur in the joint, which wasin agreement with the test result. To confirm the joint shear capacity(Vws þ Vcs þ Vst), further tests are required for specimens with agreater joint shear demand.

Summary and Conclusions

In the present study, a beam-column connection method forTSC beams and PSRC columns was developed. In the proposedmethod, relatively complex details, such as band plates and wingplates, were used to connect the steel angles of the column and theU-section of the beam. In particular, for fabrication and concreteplacement, the flanges of the U-section in the beam were intention-ally not connected to the joint. Cyclic loading tests for three full-scale connection specimens were performed.

The primary test results are summarized as follows:• The load-carrying capacity reached peak strengths at story drift

ratios δ ¼ 2.0–3.0%. After the peak strength, capacity decreasedbecause of the web local buckling of the U-section at the columnface. Ultimately, the specimens failed because of a tensile frac-ture of the buckled web plates that was initiated at the weld jointbetween the web and the bottom flange of the U-section.

• The deformation capacity reached δu ¼ 4.0% for the interiorconnections of the I450 and I550 specimens. It reached δu ¼3.0% for the exterior connection of E550.

• The specimen test strengths were on average 5% greater than thenominal strengths predicted from the plastic stress distributionsof the beam section, indicating that the discontinuity of the ofthe TSC beam flanges did not significantly affect the load-carrying capacity of the beam and connection.

• As yielding occurred in the web plate of the beam, significantweb plate anchorage slip occurred inside the joint. For this rea-son, the energy dissipation capacity of the specimens decreased,showing a pinching in cyclic behavior. However, the anchorageslip mitigated possible stress concentration that may have beencaused by the discontinuity of the flange plate, which contrib-uted to the increase in specimen deformation capacity.

• The joint shear strength of the TSC beam–PSRC column con-nection was evaluated using a modified ASCE Task Committeerecommendation. The web shear yielding of the U-section, thedirect strut mechanism of the infilled concrete, and the strut-and-tie mechanism formed by the band plates were consideredin the evaluation of joint shear capacity. The results showed thatpredicted joint shear capacity was 1.53–2.22 times greater thanthe joint shear requirement. This result agreed with the testresults showing no shear failure in the joint.

Acknowledgments

This research was financially supported by grants from theR&D Policy Infra Program (Code 11-Technology Standardization-09-01, Ministry of Land, Transportation and Maritime Affairs ofKorea), the Small and Medium Business Administration inKorea (No. 00045821), the Integrated Research Institute forConstruction and Environmental Engineering, Seoul NationalUniversity Research Program (Ministry of Education and HumanResources Development), and the Basic Science Research Pro-gram through the National Research Foundation of Korea(No. 2012R1A1A1003282, Ministry of Education, Science, andTechnology). The authors are grateful to the authorities for theirsupport.

Notation

The following symbols are used in this paper:Abp = cross-sectional area of the band plate;Fy = yield stress of steel plates in the TSC beam;

Fybp = yield stress of the band plate;H = net column height;L = length of the beam between the vertical supports;

Mnc = flexural moment capacity of the column;Mp = flexural moment capacity of the TSC beam;Muc = flexural moment demand of the column;Pn = lateral load-carrying capacity predicted using the plastic

moment capacity of the beam;Pu = lateral peak strength;Vb = shear demand of the beam;Vc = shear demand of the column;Vst = shear strength of the compression field mechanism;Vcs = shear strength of the core concrete;Vws = shear strength of the web steel plate;baf = distance between the corner angles in the direction

orthogonal to the shear;bic = width of the infilled concrete;bo = effective width of the concrete compression field;d = center-to-center distance between the top and bottom band

plates;dw = distance between the top and bottom flanges of the

U-section;f 0c = compressive strength of the concrete;

fy = yield stress of the rebar;hb = overall depth of the TSC beam, including slab

thickness;hc = depth of the column;ho = distance between the corner angles in the direction parallel

to the shear;hwj = effective depth for Vws;s = spacing of the transverse bars;tw = thickness of the web plate;δu = maximum story drift ratio; andκ = energy dissipation ratio.

References

AIK (Architectural Institute of Korea). (2009). “Korean building code andcommentary.” Seoul.

AISC. (2005). “Seismic provisions for structural steel buildings.” ANSI/AISC 341-05, Chicago.

AISC. (2010). “Specification for structural steel building.” ANSI/AISC360-10, Chicago.

ASCE. (1994). “Guidelines for design of joints between steel beams andreinforced concrete columns.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(1994)120:8(2330), 2330–2357.

AWS (American Welding Society). (1998). “Structural welding code-reinforcing steel.” AWS D1.4, Miami, FL.

BSI (British Standards Institution). (2005). “Eurocode 4: Design ofcomposite steel and concrete structures.” London, 90.

Campione, G. (2010). “R/C columns strengthened by means of steel anglesand battens: Testing, modeling and design.” Stud. Res. Politecnico diMilano, 30, 42–72.

Cheng, C., and Chen, C. (2005). “Seismic behavior of steel beam andreinforced concrete column connections.” J. Constr. Steel Res., 61(5),587–606.

Eom, T. S., Hwang, H. J., Park, H. G., Lee, C. N., and Kim, H. S. (2014).“Flexural test for steel-concrete composite members using prefabricatedsteel angles.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X.0000898,04013094.

© ASCE 04015020-11 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Hwang, H. J., Eom, T. S., Park, H. G., Lee, C. N., and Kim, H. S.(2012). “Compression test for prefabricated composite columns usinghigh-strength steel angles.” J. Korean Soc. Steel Constr., 24(4),361–369 (in Korean).

Kim, C. S., Park, H. G., Chung, K. S., and Choi, I. R. (2014). “Eccentricaxial load capacity of high-strength steel-concrete composite columnsof various sectional shapes.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X.0000879, 04013091.

Kim, S. B., Ham, J. T., Lee, C. N., and Kim, S. S. (2006). “Study on thestructural behavior of TSC beam-to-SRC column connection.” J. Arch.Inst. Korea, 22(6), 55–62 (in Korean).

Lee, C. H., et al. (2013). “Cyclic seismic testing of composite concrete-filled U-shaped steel beam to H-shaped column connections.” J. Struct.Eng., 10.1061/(ASCE)ST.1943-541X.0000635, 360–378.

Monturi, R., and Piluso, V. (2009). “Reinforced concrete columns strength-ened with angles and battens subjected to eccentric load.” Eng. Struct.,31(2), 539–550.

Oehlers, D. J. (1993). “Composite profiled beams.” J. Struct. Eng.,10.1061/(ASCE)0733-9445(1993)119:4(1085), 1085–1100.

Oehlers, D. J., Wright, H. D., and Burnet, M. J. (1994). “Flexural strengthof profiled beams.” J. Struct Eng., 10.1061/(ASCE)0733-9445(1994)120:2(378), 378–393.

Park, H. G., Hwang, H. J., Lee, C. H., Park, C. H., and Lee, C. N. (2012).“Cyclic loading test for concrete-filled U-shaped steel beam-RC columnconnections.” Eng. Struct., 36, 325–336.

Uy, B., and Bradford, M. A. (1995a). “Ductility of profiled compositebeams. Part I: Experimental study.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(1995)121:5(876), 876–882.

Uy, B., and Bradford, M. A. (1995b). “Ductility of profiled compositebeams. Part II: Analytical study.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(1995)121:5(883), 883–889.

Xuemei, L., and Gustavo, J. P. (2004). “Seismic behavior of reinforced con-crete column-steel beam subassemblies and frame systems.” J. Struct.Eng., 10.1061/(ASCE)0733-9445(2004)130:2(310), 310–319.

© ASCE 04015020-12 J. Struct. Eng.

J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

SEO

UL

NA

TIO

NA

L U

NIV

ER

SIT

Y L

IB o

n 02

/12/

15. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.