Cyclic loading test for concrete-filled U-shaped steel ...senkuzo.com/sen/wp-content/uploads/2014/12/1000028.pdf · Cyclic loading test for concrete-filled U ... tested under cyclic

Engineering Structures 36 (2012) 325–336

Contents lists available at SciVerse ScienceDirect

Engineering Structures

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

Cyclic loading test for concrete-filled U-shaped steel beam–RC column connections

Hong-Gun Park a, Hyeon-Jong Hwang a,⇑, Cheol-Ho Lee a, Chang-Hee Park a, Chang-Nam Lee b

a Department of Architecture and Architectural Engineering, Seoul National University, Gwanak 599, Gwanak-ro, Kwanak-Gu, Seoul 151-744, South Koreab Sen Structural Engineers Co. Ltd., Dangsan-dong, Yeongdeungpo-gu, Seoul 150-808, South Korea

a r t i c l e i n f o

Article history:Received 11 July 2011Revised 12 December 2011Accepted 16 December 2011

Keywords:Composite beamRC columnSeismic performance testSpecial moment frame

0141-0296/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.engstruct.2011.12.033

⇑ Corresponding author. Tel.: +82 2 880 7053; fax:E-mail address: [email protected] (H.-J. Hw

a b s t r a c t

Concrete filled U-shaped steel beams are used to increase the flexural strength and stiffness of the steelbeams by using concrete fill. In this study, a seismic detail for the concrete filled U-shaped steelbeam–RC column connection was developed. A special detail using diagonal re-bars and welded re-barconnections was used to strengthen the beam–column joints. To verify the seismic performance of thebeam–column connection, two full-scale specimens were tested under cyclic loading. The test parameterwas the size of the U-shaped steel beam. The depths of the composite beams were 610 mm (steel platethickness = 6 mm) and 710 mm (steel plate thickness = 8 mm) including the slab depth of 160 mm. Thetest results showed that the specimens exhibited good strength, deformation, and energy dissipationcapacities. The deformation capacity exceeded 4% inter-story drift ratio. The primary failure mode of thespecimens was buckling and fracture of the steel plate in the beam.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction and 0.04 radians for a special moment frame. At the maximum

Conventional composite beams consist of a wide-flange steelsection and a concrete slab on the top. However, various types ofcomposite beams have recently been developed for better construc-tability and economy [1]. TSC beam is one of these composite beams(Fig. 1). The TSC beam is composed of a U-shaped thin steel platesection, infilled core concrete, and concrete slab. To manufacturethe U-shaped steel section, two thin J-shaped steel plates with6 mm or 8 mm thickness are cold-formed and flare-weldedtogether. The infilled concrete decreases the beam deflection andfloor-vibration by increasing the beam stiffness under the serviceload condition. Furthermore, it can increase the deformation capac-ity of the beam, preventing early local buckling of the thin steelplates. Similar to conventional composite beams, the concrete slabalso increases the beam stiffness by increasing the overall beamdepth. If necessary, the TSC beam is designed to transfer negativemoment to the beam–column joint by using tension re-bars in theslab.

The TSC beam is most efficiently used for simply supportedbeams which are subjected to positive moments along the beamlength. However, to extend its range of applications, moment con-nection details for the TSC beam–column connection need to bedeveloped, and the earthquake resistance should be evaluatedagainst cyclic loading. AISC seismic provisions [2] require that abeam–column connection be able to develop an inter-story driftratio greater than 0.02 radians for an intermediate moment frame,

ll rights reserved.

+82 2 871 5518.ang).

deformation, the load-carrying capacities of the beam and the con-nection should be at least 80% of their peak strengths.

The majority of existing studies have been performed for WF(wide flange section) steel beam–RC column connections [3–6].Researchers reported that the steel bearing plates and band plates(welded to the steel beam section at the column face), and theU-shaped ties (passing through holes in the beam web), success-fully strengthened the beam–column joint. Hence the test speci-mens exhibited satisfactory strength and deformation capacities,though the energy dissipation capacity slightly decreased due tothe slip deformation between the concrete and the steel section[3–6]. The maximum inter-story ratio of the specimens rangedfrom 4.0% to 5.5%. Ultimately, the test specimens failed due tothe fracture of the bottom flange after low cycle fatigue.

The present study focused on the TSC composite beam (usingU-shaped steel section–RC column connection). The seismic detailsfor the beam–column connection were developed and verified bycyclic loading tests. For this purpose, the following studies wereperformed.

(1) Seismic details for the TSC composite beam–RC column con-nection were developed, considering the re-bar prefabrica-tion construction method for columns.

(2) Two full-scale specimens were tested under cyclic loading.(3) The load-carrying capacity, maximum deformation, and

energy dissipation of the test specimens were evaluatedand compared with the requirements for earthquake-resisting moment frames [2].

(4) The shear strength of the panel zone was evaluated accord-ing to the ASCE guideline [7].

Notations

Abcs effective concrete shear section area of composite beam(mm2)

Abss effective web steel plate shear section area of compositebeam (mm2)

Ac effective area of column section excluding concrete cov-er thickness (mm2)

Acs effective shear section area of column (mm2)Ag gross area of column section (mm2)Ash cross-sectional area of ties in a layer (mm2)bp width of the core concrete in TSC beam (mm)bf width of TSC beam (mm)d depth of TSC steel beam (mm)df center-to-center distance between beam flanges (mm)dw overall beam depth including the slab depth (mm)Ebc concrete elastic modulus of composite beam (MPa)Ebs steel elastic modulus of composite beam (MPa)Ec elastic modulus of column (MPa)Fy yield strength of reinforcement (MPa)f 0c concrete strength (MPa)fysh yield strength of tie (MPa)fyw yield strength of web steel plate in TSC steel beam (MPa)Gbc concrete shear modulus of composite beam (MPa)Gbs steel shear modulus of composite beam (MPa)Gc shear modulus of column (MPa)Hc column height (mm)h column depth (mm)hs effective depth for Vs (distance between the longitudinal

re-bars in the compression and tension sides) (mm)Ibc concrete moment of inertia for composite beam cross

section (mm)Ibs steel plate moment of inertia for composite beam cross

section (mm)Ic moment of inertia for column cross section (mm)Lb net span of beam (mm)

Mn flexural plastic moment of beam based on material testresults (kN m)

Muc required moment strength (kN m)Mnc moment capacity (kN m)n the number of tie layers arranged in the depth of the

outer compression fieldtw thickness of web steel plate in TSC steel beam (mm)V lateral load applied to column (kN)Vb beam shear force (kN)Vc column shear force (kN)V 0c shear strength of outer concrete (kN)Vcf shear strength of outer concrete including the effect of

ties (kN)Vcs shear strength of core concrete (kN)Vp lateral load capacity calculated using the plastic mo-

ment of beam (kN)Vs shear strength of web steel plate (kN)V 0s shear strength of column ties (kN)x distance from cross beam face to column face (mm) (see

Fig. 14d)y length of cross beam (mm) (see Fig. 14d)Dt total inter-story drift (mm)Db inter-story drift caused by beam deformation (mm)Dbe inter-story drift caused by beam elastic deformation

(mm)Dbp inter-story drift caused by beam plastic deformation

(mm)Dc inter-story drift caused by column deformation (mm)Dpz inter-story drift caused by panel zone deformation

(mm)/ safety factorm poisson’s ratiohb rotation of plastic hinge in beam (%)hbe elastic rotation of plastic hinge in beam (%)

326 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

2. Test plan

Two cruciform specimens, S1 and S2, were prepared for thistesting (Table 1, Fig. 2). The specimens consisted of two concretefilled TSC steel beams, an RC slab, and an RC column. The specimenlength between the vertical supports was Lb = 6760 mm. The col-umn height from the bottom hinge to the lateral loading pointwas Hc = 2500 mm. Assuming an interior connection, cross-beamswere connected to the beams in the beam–column joint. The spec-imens were designed to satisfy the design concept of strongcolumn–weak beam (see (21) in Table 1). The TSC composite beam(calculation of flexural moment capacity, design of shear

Fig. 1. TSC beam section.

connectors, etc.) was designed according to AISC seismic provisions[2]. The design and details of the RC column and slab were basedon ACI 318 [8].

The properties of the specimens are listed in Table 1. The primarytest parameter was the size of the TSC beam: 450 � 270 � 6 mm forS1 and 550 � 270 � 8 mm for S2 (Fig. 1). The yield strengths for the6 mm and 8 mm thick steel plates were 466.1 MPa and 435.2 MPa,respectively. In S1, the width-to-thickness ratios were 45 and 75for the flange and the web, respectively. In S2, the ratios were33.8 and 68.8. The width-to-thickness ratios of the flanges satisfiedthe limit value of 2:26

ffiffiffiffiffiffiffiffiffiffiE=Fy

pð¼ 46:8Þ for S1) specified by AISC 360

[9], while the ratios of the webs did not satisfy the limit value of(3:00

ffiffiffiffiffiffiffiffiffiffiE=Fy

p¼ 62:1 for S1).

The dimensions of the column section were 800 � 800 mm.The concrete strengths were 35.0 MPa and 33.1 MPa for S1 andS2, respectively. For the longitudinal re-bars of columns, 20-D32(D32, area = 819 mm2, Fy = 541.2 MPa) were used. For the columnties, D13 (D13, area = 129 mm2) bars were placed at 150 mm ver-tical spacing. For the ties in the beam–column joint, D16 (D16,area = 199 mm2, Fy = 491.8 MPa) bars with 150 mm vertical spac-ing were used. The width and thickness of the composite metaldeck slab were 1690 mm and 160 mm. The width of the slabwas determined as one-fourth of the overall beam length6760 mm, according to AISC 360 [9]. The slab concrete strengthwas 32.6 MPa for S1 and 28.9 MPa for S2. The thickness of themetal deck steel plate was 1.6 mm and the depth of the metal

Table 1Summary of test program.

Specimen S1 S2

Beam (1) Concrete strength (MPa) 32.6 28.9(2) Slab re-bar 4-D25a+6-D10a

(3) Yield strength of re-bar D25a (MPa) 515.6(4) Yield strength of TSC steel plate (MPa) 466.1 435.2(5) TSC width (mm) 270(6) TSC depth (mm) 450 550(7) TSC steel plate thickness (mm) 6 8(8) TSC cross-section area (mm2) 8351 11135(9) Slab width (mm) 1690(10) Slab depth/metal deck depth (mm) 160/75(11) Stud spacingb (mm) 2-Ø16@115Composite section (12) Plastic positive moment capacityc Mþn (kN m) 1489 2152

(13) Plastic negative moment capacityc M�n (kN m) 1161 1668TSC steel section (14) Plastic positive moment capacityc Mþn (kN m) 546 929

(15) Plastic negative moment capacityc M�n (kN m) 546 929Column (16) Concrete strength f 0c (MPa) 35.0 33.1

(17) Column dimensions (mm) 800 � 800(18) Longitudinal reinforcement 20-D32a

(19) Yield strength of re-bars D32a (MPa) 541.2(20) Flexural moment capacityc (kN m) 3573 3498(21) Column moment capacity/beam moment capacity 2.70 1.83

a Re-bars D10 (Area = 71 mm2, Fy = 541.8 MPa), D25 (Area = 510 mm2, Fy = 515.6 MPa), D32 (Area = 819 mm2, Fy = 541.2 MPa).b Studs Ø16 (Area = 201 mm2, Fu = 440 MPa, shear strength Vsn = 80.4 kN).c Composite section capacity calculated based on material test results.

H.-G. Park et al. / Engineering Structures 36 (2012) 325–336 327

deck was 75 mm. The metal deck was connected to the top of theflange of the TSC beam by stud-welding. 2-/ � 16 studs with115 mm spacing (/16, area = 201 mm2, Fu = 440 MPa) werearranged along the beam length to develop the full compositepositive and negative moment capacities (Table 1 (11)). 4-D25(D25, area = 510 mm2, Fy = 515.6 MPa) were used for the negativeflexural moment of the TSC composite beam. Additionally, 6-D10(D10, area = 71 mm2, Fy = 541.8 MPa) were placed in the slabs. Allof the re-bars in the slab were considered in the calculation offlexural capacity of the TSC composite beam. The momentcapacities of the composite beam and RC column are presentedin Table 1 (12), (13), and (20). They are the plastic momentscalculated by using the material test results. Using the strongcolumn–weak beam concept, the sum of the upper and lower col-umn moment capacities (in this test, the column was notsubjected to axial load) was designed to be greater than thesum of the left and right composite beam moment capacities(Table 1 (21)).

Fig. 3 shows the test set-up. The column was supported by ahinge at the bottom. The beam ends were vertically supportedallowing the lateral movement of the specimen. To measure thereactions of the beams, 1000 kN load-cells were installed at thevertical supports. Four lateral supports were installed to preventthe transverse movement of the specimen, simulating the slab dia-phragm action (Fig. 2). A lateral load was applied to the top of thecolumn by an actuator. The loading program was planned accord-ing to AISC seismic provisions [2] (Table 2). After the inter-storydrift ratio exceeded 4.0%, two load cycles were applied at every1.0% increase. In this test, axial compressive load was not appliedto the columns because a very large force is required, and becausea low level of compressive force tends to prevent diagonal crackingat the panel zone. A–D in Fig. 3 indicates the measurement instru-ments including linear potentiometers and strain gauges. Using thelinear potentiometers, the story drift (C), nonlinear flexural defor-mation of the beam plastic hinge zones (A), shear deformation ofthe panel zone (B), and support movements (D) were measured.The strains of the TSC beam’s steel plate, slab re-bars, and rein-forcement at the beam–column joint were measured by the straingauges.

3. Connection details

In the existing methods for the WF steel beam–RC column con-nection, bearing plates and band plates, which are welded to thesteel beam at the column face, are used to prevent the concretebearing failure and the joint shear failure by developing a concretestrut mechanism in the panel zone [4–6]. In the proposed method,on the other hand, the shape of the steel beam section and the re-bar prefabrication method for columns were considered. For fastconstruction, the re-bar cage for the column and beam–columnjoint is prefabricated in the shop, and the steel brackets for theTSC beam connection are assembled with the re-bar cage. Afterthe re-bar cage is erected on the construction site, the TSC beamsare bolt-connected to the bracket. Considering the constructionmethod, prefabricated re-bars were used to strengthen thebeam–column joint. In the welding between the steel sectionsand re-bars, special heat treatment for alleviating developmentof residual stresses and tearing was not used.

Fig. 4 shows the connection details of specimen S1. The key ele-ments A1–D are marked in the figure. The TSC steel beam wasdirectly weld-connected to the column longitudinal bars by diago-nal re-bars to prevent slip between the steel beam and RC columnunder cyclic loading. The upper and lower diagonal re-bars, A1(D16) and A2 (D22 for specimen S1, D25 for specimen S2) werewelded to the column longitudinal bars and the top and bottomflanges of the TSC beam. Particularly, the lower diagonal re-barsA2 were integrated with two band re-bars B (D22 for specimenS1, D25 for specimen S2), by welding. L-shaped lateral re-bars(D16) for column ties were installed, passing through the holesin the TSC beam web. At the holes, the lateral re-bars were weldedto the web steel plates (C in Fig. 4). At the intersection between theTSC beam and the cross-beam, a 6 mm thick cover plate waswelded on the top flange of the TSC beam (D). The cross-beamwas also connected to the column longitudinal bars by the diago-nal re-bars. The cross-beam acts as a connector that enables thepanel zone concrete outside the TSC beam to provide shear resis-tance (see the section later in the paper on evaluation of joint shearstrength). If the cross-beam is not used, the bearing plate and/orcover plate used in the existing methods are required.

Fig. 2. Dimensions and details of specimen S1.

Fig. 3. Test setup and measurement of specimen S1.

328 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

The TSC beams were connected to the TSC brackets by M22 slip-critical bolts (Area = 380 mm2 and design shear strength =83.6 kN). In specimen S1, 8, 20, and 16 bolts were used for thetop flange, web, and the bottom flange, respectively. In specimen

S2, 8, 24, and 20 bolts were used. The details of specimen S2 werethe same as those of S1 except for the TSC beam’s depth, thediameter of the diagonal re-bars, and the number of slip-criticalbolts.

Table 2Loading plan for test specimens.

Displacement (actuatorstroke), mm

Interstory driftratio (%)

The number of loadcycles

±9.375 ±0.375 6 cycles±12.5 ±0.5 6 cycles±18.75 ±0.75 6 cycles±25.0 ±1.0 4 cycles±37.5 ±1.5 2 cycles±50.0 ±2.0 2 cycles±75.0 ±3.0 2 cycles±100.0 ±4.0 2 cycles

H.-G. Park et al. / Engineering Structures 36 (2012) 325–336 329

4. Test results

4.1. Load-story drift relationship

Fig. 5 shows the lateral load-story drift relationships of thespecimens. The story drift ratio was calculated by dividing thenet lateral story drift by the column height. Since the specimenswere designed to satisfy the beam-sway mechanism (i.e. the strongcolumn–weak beam concept), the load-carrying capacity of thespecimens can be estimated with the plastic moments of thebeams ((12) and (13) in Table 1). Vþp and V�p in Fig. 5 indicate thepredicted lateral load-carrying capacities. As shown in the figure,the test results agree with the predicted load-carrying capacities.The maximum story drift ratios of Specimens S1 and S2 were 4%and 5%, respectively, which can be regarded as good deformationcapacities. The maximum deformation was defined as the post-peak deformation corresponding to 80% of the peak strength. Onthe other hand, the specimens showed pinching in the load-storydrift relationship, though it was not severe. For this reason, the en-ergy dissipation capacity decreased.

Fig. 6 shows the moment–rotation relationship at the beamplastic hinge zone in specimen S2. The beam moment at the col-umn face was estimated from the vertical reactions measured bythe load cells at the beam end. The rotation angles hb at the beam

Fig. 4. Details of beam–column conne

plastic hinges were calculated from the deformations measured atA in Fig. 3. As shown in Fig. 6, the rotation at the beam plastic hingezone exceeded 0.0400 radian. In S2, the maximum rotation at thebeam plastic hinge zones were 0.0377 radian for the positive mo-ment and 0.0423 radian for the negative moment. In S1, the max-imum rotations were 0.0321 radian for the positive moment and0.0415 radian for the negative moment. In Fig. 6, the measuredbeam moments agreed with the predicted flexural moment capac-ities Mþ

n and M�n (Table 1 (12), (13)). This result indicates that the

flexural moment capacity of the TSC composite beam can be accu-rately estimated by the plastic theory according to current designcodes.

Fig. 7 shows the variation of the average secant stiffness corre-sponding to the first load cycle at each inter-story drift ratio. Thesecant stiffness indicates the slope between the maximum positiveand negative story drifts in the load-story drift relationship. InFig. 7, the secant stiffness decreased as the inter-story drift in-creased. S2 showed greater stiffness than S1, due to the greaterbeam depth. The secant stiffness of S1 and S2 at 4% inter-story driftratio was approximately 35% of the initial stiffness.

4.2. Failure modes

Fig. 8 shows the crack and damage patterns in the specimens. Inthe beam–column joint of S1 (Fig. 8a and b), initial concrete crack-ing occurred at 0.75% story drift ratio and propagated to the overallpanel zone. At 3% story drift ratio, slab concrete cracking initiallyoccurred. At 5% story drift ratio, local buckling occurred in theweb and bottom flange of the TSC beam, and concrete cover spall-ing occurred in the column. In specimen S2, concrete cover spallinginitially occurred at 3% story drift ratio. Concrete cracking and con-crete cover spalling were severe (Fig. 8c and d). At 6% story drift ra-tio, the steel plates of the TSC beam were buckled, and the flexuralD25 re-bars in the slab were also buckled.

Fig. 9 shows the failure modes A–D of S1 and S2. The primaryfailure modes of S1 were steel plate buckling in the TSC beam(D), steel plate fracture at the welded connection between the

ction in specimen S1 (unit:mm).

Fig. 5. Load-story drift relationships.

Fig. 6. Flexural capacity of beams measured by load-cells in specimen S2.

Fig. 7. Variations of secant stiffness according to story drift angle.

330 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

bottom flange of the TSC beam and the band re-bars (E), and frac-ture of the column tie welded to the TSC beam web (B). On theother hand, concrete crushing at the column face below the TSCbeam (C) was restrained due to the diagonal re-bars connecting

the TSC beam and the column longitudinal re-bars. As mentioned,the width-to-thickness ratio of the web did not satisfy the require-ment of AISC 360 [9]. However, local buckling of the web plate oc-curred after a very large inelastic deformation (i.e. 5% story driftratio). Thus, the non-compliance of the width-to-thickness ratioof the web did not significantly affect the load-carrying capacityand deformation capacity of the specimen S1. This result indicatesthat the infilled concrete efficiently prevented early local bucklingof the thin web plate.

The failure modes of S2 were similar to those of S1. However,since the shear force of S2 was greater than that of S1 due to thegreater beam flange area, the beam–column joint was severelydamaged. Unlike S1, of the four D25 slab re-bars passing throughthe column section, the two at the center were buckled (A). How-ever, the re-bar buckling was not the primary cause of failure of thespecimen, because the re-bar buckling did not affect the tensilestrength of the re-bars, and the slab concrete was able to resistthe flexural compression under the positive moment, without theslab re-bars. In this test, axial force was not applied to the column.However, if a high axial force is applied to the column, the beam–column joint damaged by the high joint shear force may fail due tocrushing of the concrete. Thus, further studies are required toinvestigate the effect of axial force on the beam–column joint.

4.3. Contributions to inter-story drift

The overall inter-story drift Dt of the specimen comprises thedeformations of the column Dc, beam Db, and panel zone Dpz

(see Fig. 10). Further, the story drift Db caused by the beam defor-mation consists of the elastic deformation Dbe and the plasticdeformation Db (Db = Dbe + Dbp). Therefore, the overall inter-storydrift Dt of the specimen can be defined as

Dt ¼ Dc þ Dbe þ Dbp þ Dpz ð1Þ

Since concrete cracking did not occur in the column, it can beassumed that the column deformation remains in an elastic state.Thus, the inter-story drift caused by the column can be calculatedas follows (Fig. 10a)

Dc ¼V

3EcIcH3

c1 þ H3c2

� �þ V

GcAcsðHc1 þ Hc2Þ ð2Þ

where V is the lateral load applied at the top of the column, Ec is theelastic modulus of the concrete, Ic is the moment of inertia of thecolumn section, Hc1 and Hc2 are the heights of the upper and lowercolumns, respectively, and Acs is the effective shear section, which isdefined as the gross section area divided by 1.2, considering

Fig. 8. Cracking damage patterns at beam–column joints.

Fig. 9. Failure modes of specimens.

Fig. 10. Estimations of story drift developed by column, beam, and panel zonedeformations.

H.-G. Park et al. / Engineering Structures 36 (2012) 325–336 331

non-uniform shear stress distribution. Gc is the shear modulus,Gc = Ec/2(1 + v), where v is the Poisson’s ratio of the concrete (=0.2).

The inter-story drift Dbe caused by the beam elastic deformationcan be calculated as follows (Fig. 10b)

Dbe ¼V

3ðEbcIbc þ EbsIbsÞHc

Lb

� �2

L3b1 þ L3

b2

� �

þ VðGbcAbcs þ GbsAbssÞ

Hc

Lb

� �2

ðLb1 þ Lb2Þ ð3Þ

where Ebc is the concrete elastic modulus, Ibc is the moment of iner-tia of the concrete section in the composite T-beam section, Ebs isthe steel elastic modulus, Ibs is the moment of inertia of the TSCsteel section, Lb1 and Lb2 are the net lengths of the beams, Hc isthe overall height of the specimen, Gbc is the concrete shear modu-lus, Abcs is the effective concrete shear section of the composite

332 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

beam, Gbs is the steel shear modulus, Abss and is the effective shearsection of the TSC steel section, i.e. the area of the web steel plate.

Fig. 10c shows the inter-story drift Dbp caused by the beamplastic deformation. In the figure, hb indicates the inelastic rotationin the plastic hinge zone (the length = 450 mm for S1 and 550 mmfor S2), which was estimated from the linear potentiometer A inFig. 3. Here, hb is defined as the average of the inelastic rotationsof the two plastic hinges.

The plastic rotation is calculated eliminating the elastic rotationhbe from the inelastic deformation, hb–hbe. In S1, the maximumplastic rotations at the beam plastic hinge zone were 0.0273 radianfor the positive moment and 0.0367 radian for the negative mo-ment. In S2, the maximum plastic rotations were 0.0330 radianfor the positive moment and 0.0376 radian for the negative mo-ment. In Fig. 10c, the imaginary vertical displacement dbp at thelocations of the beam supports is calculated as

dbp ¼ hb � hbeðLb � aÞ=2 ð4Þ

where a = the panel zone depth. Then, the inter-story drift Dbp

caused by the plastic hinge rotations can be calculated as follows:

Dbp ¼Hc

Lb

Xdbpðhb � hbeÞ Hc � a

Hc

Lb

� �ð5Þ

The inter-story drift developed by the panel zone deformation iscalculated as follows:

Fig. 11. Contributions of column, bea

Dbz ¼cpz

2ðHc � bÞ � bþ ðLb � aÞHc

Lb� a

Hc

Lb

� �

¼ cpz Hc � aHc

Lb� b

� �ð6Þ

The shear angle at the joint cpz is estimated ascpz ¼ dj � d0j

� � ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffia02 þ b02

p=2a0b0 , where dj and d0j are the diagonal

deformations measured by the two diagonal linear potentiometers(B) at the panel zone (Fig. 3), and a0, b0, are the horizontal andvertical distances between the ends of a diagonal linearpotentiometer.

Fig. 11a shows the ratio of each contribution calculated fromEqs. (1)–(6) to the overall inter-story drift measured from the test-ing of specimen S1. As the overall inter-story drift increased, thecontribution of the beam, particularly the contribution of the beamplastic deformation, increased. At 4.1% story drift ratio, the contri-bution of the beam was 3.4% drift ratio (elastic deformation = 0.4%and plastic deformation = 3.0%), which corresponds to 83% of theoverall story drift. The sum of all the contributions calculated fromEqs. (1)–(6) was close to 100%, which indicates that the calculationresults agree with the measured overall story drift.

In S2, the panel zone deformation was not properly measuredbecause of the severe concrete damage in the panel zone. Thus,the contribution of the panel zone was not estimated. Fig. 11bshows the contributions of the beam plastic hinge rotations for

m, and panel zone to story drift.

Fig. 13. Energy dissipation per load cycle of S1, S2, and plastic hinge zone of S2.

H.-G. Park et al. / Engineering Structures 36 (2012) 325–336 333

S1 and S2. The contribution of the beam plastic hinge rotation in S2was less than that in S1. This result indicates that in S2, the contri-bution of the panel zone deformation was significant because ofthe greater shear force.

4.4. Measurement by strain gauges

Fig. 12 shows the strains of the slab re-bars and TSC beam mea-sured by strain gauges. Fig. 12a shows the strain of the slab re-barlocated at 45 mm from the column face, and the strain of the bot-tom flange in the TSC beam. In the bottom flange, after the strainreached 0.015, the strain did not decrease to zero under reversedloading, due to the buckling of the steel plate, and the strain con-tinued to increase to 0.04. Under the positive moment, since theneutral axis was located in the slab section, a large tensile strainoccurred at the bottom flange, which caused the buckling of thebottom flange under cyclic loading. In the flexural re-bars, themaximum strain was 0.015.

Fig. 12b shows the strains of the diagonal re-bars connectingthe steel beam and the column longitudinal bars. The strain wasless than the yield strain (�0.002) corresponding to the specifiedyield strength 400 MPa, which indicates that the diagonal bars inan elastic state successfully transferred the shear forces of thebeams to the column longitudinal re-bars.

Fig. 12. Strain of re-bars and beam steel plates in S1.

Fig. 13 shows the energy dissipation per load cycle in S1, S2, andthe plastic hinge zone of S2. The energy dissipation was estimatedfrom the area of each hysteresis loop shown in Figs. 5 and 6. The en-ergy dissipation of S1 and S2 suddenly decreased after the maxi-mum deformations (4% in S1 and 6% in S2). In S2, the energydissipated by the beam plastic rotation was more than 80% of theoverall energy dissipation. This result implies that 20% of the overallenergy dissipation was dissipated by the re-bars and steel plates inthe beam–column joint. It should be noted that shear-slip or bond-slip do not contribute to the energy dissipation per load cycle.

4.5. Effective stiffness of beams

The effective stiffness of beams should be carefully defined toaccurately estimate the deflection of the structure under serviceloading. At the proposed beam–column connection, the flexuralstiffness of the T-shaped beam decreases due to two reasons: (1)only a part of the slab width is directly connected to the columnface; and (2) concrete cracking occurs at the beam end. Thus, theproposed connection should be regarded as a semi-rigid connec-tion, as defined in AISC 360 [9]; the decrease of the beam flexuralstiffness at the beam end should be considered in the analysis anddesign of the structure.

Fig. 11b shows the inelastic rotation at the beam end, whichwas measured at A in Fig. 3. The inelastic rotation at the serviceload (corresponding to 60% of the ultimate load), excluding theelastic deformation, was hbp = 0.0041 radian for S1 and 0.0038 ra-dian for S2. Using the inelastic rotation, the effective rotationalstiffness at the beam end can be estimated: kb = M/hbp, whereM = the beam end moment. Using the measured inelastic rotation,the rotational stiffness at the beam end was estimated askb = 137 MN m/rad for S1 and 233MN m/rad for S2.

Usually, in the estimation of deflection of concrete structures,the average effective stiffness over the beam length is used, consid-ering the effect of concrete cracking. In the present study, the aver-age effective stiffness of the T-shaped beam cross-section isdefined as EIe = aEbcIbc + EbsIbs, where a = stiffness reduction factor.Using the measured overall deflection of the beam, a was esti-mated as 0.15 in average (0.14 for S1 and 0.16 for S2). The kb orEIe value can be used to calculate the deflection of a structure un-der service loading.

5. Evaluation of joint shear strength

Generally, the failure of the steel beam–RC column joints iscaused by either shear failure of the panel zone or bearing failureof the column concrete. In the proposed connection method, con-

Fig. 14. Shear transfer mechanism at beam–column joint [7].

Table 3Shear strength of the panel-zone estimated by ASCE design guidelines.

Specimens S1 S2

Demand Muc (kN m) Eq. (11b) 2065 2824Vsdf (kN m) (contribution ratio) 978 (41.7%) 1494 (43.3%)Vcs (0.75dw) (kN m) (contribution ratio) 978 (41.7%) 1109 (32.2%)Vcfd (kN m) (contribution ratio) 389 (16.6%) 844 (24.5%)Capacity Mnc (kN m), Eq. (11c) 2345 3447Capacity/demand (Mnc/Muc) 1.14 1.22

334 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

crete bearing failure was prevented by the use of the welded diag-onal re-bars which directly transferred the beam shear force to thecolumn longitudinal re-bars. Therefore, in the present study, onlythe joint shear failure mode was considered.

The joint shear strength of the test specimens was evaluatedaccording to the design guideline for steel beam–RC column con-nections by ASCE [7]. The proposed details for the TSC beam–RCcolumn connection differed from the connection details consideredby the ASCE design guidelines: the shape of the steel beam, theconnection details of the panel zone, and the use of the slab forthe composite action. Thus, the relevant equations and parameterswere modified to apply the design guidelines to the proposed con-nection method.

In the ASCE design guidelines, the shear capacity of a steelbeam–RC column connection comprises the contributions by thesteel web plate (Vs), the compression strut of the core concreteinside the TSC beam section (Vcs), and the compression field mech-anism outside the TSC beam section (Vcf) (Fig. 14).

H.-G. Park et al. / Engineering Structures 36 (2012) 325–336 335

Fig. 14a shows the shear resistance of the web steel plate. Theshear strength of the web steel plate is defined as follows:

Vs ¼ 0:6f ywtwhs ð7Þ

where fyw is the yield strength, and tw is the plate thickness. In thecase of S1, tw = 2 � 6 mm for the two web plates. hs is the effectivedepth for Vs. In the present study, hs was defined as the distance be-tween the column longitudinal bars in the compression and tensionsides, hs = 662 mm for S1.

Fig. 14b shows the compression strut mechanism of the coreconcrete inside the TSC steel beam section. The shear strength isdefined as follows:

Vcs ¼ 1:7ffiffiffiffiffiffiffiffif 0cbp

qh 6 0:5f 0cbpdw ð8Þ

where f 0c (MPa) is the concrete strength. 1:7ffiffiffiffif 0c

pindicates the shear

strength of the well confined concrete (ACI 352R-02 [10]). In theproposed connection method, the core concrete was well confinedby the U-shaped steel beam and the slab concrete. bp is the widthof the core concrete, h is the depth of the column, and dw is theoverall depth of the TSC composite beam including the slab depth.In S1, bp = 270 mm, h = 800 mm, and dw = 600 mm.

Fig. 14c shows the compression field mechanism of the concreteoutside the TSC steel section. The shear strength Vcf is defined asthe sum of the contributions of the outer concrete and lateral re-bars (i.e. ties).

Vcf ¼ V 0c þ V 0s 6 1:7ffiffiffiffiffiffiffiffif 0cbo

qh ð9aÞ

V 0s ¼ Ashfyshn ð9bÞ

V 0c ¼ 0:4ffiffiffiffiffiffiffiffif 0cbo

qh ð9cÞ

where V 0c is the shear strength of the outer concrete, and V 0s is theshear strength of the lateral re-bars.

In Eq. (9b), Ash is the cross-sectional area of ties in each layer, fysh

is the yield strength, and n is the number of layers arranged in theouter compression field depth d (n = 4 in S1). Ash should be morethan 0.004 (b–bf) sh, where sh is the spacing of tie layers. (b–bf) isthe overall width of the outer concrete, where bf is the width ofthe core concrete. In S1, (b–bf) = 518 mm, sh = 150 mm, and the tiesin the panel zone, 2-D16 (Ash = 398 mm2), satisfy the requirementof 0.004 (b–bf) sh = 311 mm2.

In Eq. (9c), bo is the effective width of the outer concrete com-pression field. In the ASCE design guideline, the outer concretecompression field is considered only when the connection has ashear transfer mechanism between the steel beam and the outerconcrete in the joint, by using a center steel column or extendedbearing plates. In the proposed connection method, the cross-beams can provide the shear transfer mechanism. The ASCE designguideline specifies the effective width of the outer concrete com-pression field as follows (Fig. 14 d)

bo ¼ Cðbm � bf Þ < 0:5d ð10aÞC ¼ ðx=hÞðy=bf Þ ð10bÞbm ¼ ðbf þ bÞ=2 < ðbf þ hÞ < 1:75bf ð10cÞ

Originally, x and y are defined using the dimensions of thecross-section of a core steel column. In the present study, theyare defined by the dimensions of the cross beam. x is the distancefrom the cross-beam face to the column face (=535 mm in S1), h isthe depth of the column, y is the length of the cross-beam(=800 mm in S1), bf is the TSC beam width. For S1, b0 = 223 mmin Eq. (10a). The depth of the outer concrete compression fieldwas defined as the TSC steel beam depth d , because the compres-sion field is confined by the slab concrete at the top and by theband re-bars at the bottom (Fig. 14c). According to the ASCE designguidelines, additional ties are required above and below the outer

concrete compression field depth. The minimum required area ofthe ties is Atie = Vcf/fysh, which should be placed within a verticaldistance 0.4d above and below the compression field depth. Atthe top of the compression field region, which is confined by theslab concrete, ties are not required. Below the compression fieldregion, two layers of 2-D22 (1548 mm2) and one layer of 2-D10(142 mm2) were used within 0.4d = 178 mm (see Fig. 4). However,the ties did not satisfy the requirement of Atie = Vcf/fysh(=2426 mm2). In this case, the ASCE guideline requires thatthe shear resistance Vcf be reduced to the ratio of the tie area tothe required area. In the case of S1, Vcf (1690 mm2/2426 mm2) � 1251 kN = kN.

In the ASCE guideline, the safety of a beam–column joint sub-jected to shear force is evaluated as follows:

Muc 6 /Mnc ð11aÞMuc ¼

XMp þ Vbh� Vcd� Vbhs ð11bÞ

Mnc ¼ Vsdf þ Vcsð0:75dwÞ þ Vcf d ð11cÞ

In Eq. (11a), Muc is the strength demand, and Mnc is the capacity ofthe joint. This equation was derived from the equilibrium of the ver-tical forces applied by shear (see ASCE guidelines [7]). In Eq. (11b),Mp is the beam plastic moment, Vb is the beam shear force, h is thecolumn depth, Vc is the column shear force, and d is the beam depth.In this evaluation, the strength reduction factor / = 1.0 was used.

In Eq. (11c), df is the center-to-center distance between thebeam flanges, and 0.75dw is the moment arm for the TSC compositebeam section. d is the outer concrete compression field depth. Incase of S1, df = 440 mm, 0.75dw = 450 mm, and d = 446 mm.

Table 3 presents the strength demands and capacities of speci-mens S1 and S2. The ratios of the capacity to the demand were 1.14for S1 and 1.22 for S2. As presented in the table, the shear strengthwas provided mainly by the TSC beam web plates and the core con-crete strut. In the outer concrete compression field, S1 did not sat-isfy the required area of ties, while S2 satisfied the required area,which results in the difference in Vcf. The sum of the shear resis-tance provided by the concrete was more than 56% of the overallshear strength of the connection. This result indicates the impor-tance of the shear transfer mechanism between the TSC steel sec-tion and the concrete in the joint.

As mentioned, the proposed details for the TSC beam–RC col-umn connection differed from the connection details consideredby the ASCE design guidelines. Thus, in this evaluation, several de-sign parameters were defined differently, which may cause impro-per evaluation of the joint shear strength. Accordingly, furtheranalytical and experimental studies are required to confirm theproposed evaluation method.

6. Conclusions

The seismic performance of the concrete filled U-shaped steelbeam (TSC beam)–RC column connection was studied. For thecomposite action, the TSC steel beam was integrated with the in-filled core concrete, slab concrete, and slab re-bars. Consideringthe U-shaped steel section and the re-bar prefabrication construc-tion method for columns, a new connection detail was used for thebeam–column joint: The TSC steel beam was directly weld-con-nected to the column longitudinal bars, by using diagonal re-barsand banded re-bars placed at the beam–column joint. Two full-scale specimens, S1 and S2, using 450 mm and 550 mm steel beamdepths, respectively, were tested under cyclic loading. The primarytest results are summarized as follows:

1. All specimens were designed according to the strong column–weak beam concept. The test results showed that plastic hingesoccurred at the beam ends as expected.

336 H.-G. Park et al. / Engineering Structures 36 (2012) 325–336

2. The load-carrying capacities of the specimens exceeded thestrengths predicted from the beam plastic moments. The max-imum story drift ratios of Specimens S1 and S2 were 4% and 5%,respectively.

3. The TSC composite beams showed good rotation capacities atthe beam plastic hinge zones. In S1, the maximum rotationswere 0.0321 radian for the positive moment and 0.0415 radianfor the negative moment. In S2, the maximum rotations at thebeam plastic hinge zones were 0.0377 radian for the positivemoment and 0.0423 radian for the negative moment.

4. The positive and negative plastic moments of the TSC compositebeam were accurately predicted by the plastic theory, accordingto current design codes.

5. Large plastic strains developed in the bottom flange of TSCbeam, which caused fracture of the bottom flange at the maxi-mum deformation of the specimen.

6. The specimens showed good energy dissipation capacity, mini-mizing shear-slip and bond-slip deformations at the beam–col-umn joint. In the case of S2, 80% of the overall energy wasdissipated by the beam plastic rotation, and 20% was dissipatedby the re-bars and steel plates in the beam–column joint.

7. The core concrete prevented premature local buckling of 6 mm(the width-to-thickness ratio = 45 and 75 for the flange and theweb, respectively) or 8 mm (the width-to-thickness ratio = 33.8and 68.8) thin steel plates used in TSC steel beam until the ten-sile strain of the plate reached 0.015.

8. Although buckling of the slab re-bars occurred in S2, the loadcarrying capacity of S2 did not decrease because of the slab con-crete. However, when this connection is used for exterior col-umns, it is desirable to provide lateral re-bars to preventbuckling of the slab re-bars.

9. The shear strength of the panel zone was evaluated by the ASCEdesign guidelines. The evaluation result indicates that the shearcapacity was greater than the shear demand. This result agreed

with the failure mode of the test specimens that failed due to frac-ture of the steel plates in the TSC steel beam. However, since thedetails for the TSC beam–RC column joint differed from those con-sidered in the ASCE design guidelines, further studies are requiredto verify the strength of the proposed TSC beam–column joint.

Acknowledgment

This research was financially supported by Sen Structural Engi-neering Co. Ltd. and The Small and Medium Business Administra-tion in Korea, and the authors are grateful to the authorities fortheir support.

References

[1] Morino S. Recent developments in hybrid structures in Japan-research, designand construction. Eng Struct Elsevier 1997;20(4):336–46.

[2] American Institute of Steel Construction (AISC). Seismic provisions forstructural steel buildings (ANSI/AISC 341-05). Chicago; 2005.

[3] Sheikh TM, Deierlein GG, Yura JA, Jirsa JO. Beam–column moment connectionsfor composite frames: Part 1. J Struct Eng ASCE 1988;115(11):2858–76.

[4] Gustavo P, James KW. Seismic response of exterior RC column-to-steel beamconnections. J Struct Eng ASCE 2000;126(10):1113–21.

[5] Xuemei L, Gustavo JP. Seismic behavior of reinforced concrete column–steelbeam subassemblies and frame systems. J Struct Eng ASCE 2004;130(2):310–9.

[6] Cheng C, Chen C. Seismic behavior of steel beam and reinforced concretecolumn connections. J Construct Steel Res 2005;61(5):587–606.

[7] ASCE Task Committee on Design Criteria for Composite Structures in Steel andConcrete. Guidelines for design of joints between steel beams and reinforcedconcrete columns. ASCE J ASCE 1993;127(1):2330–57.

[8] American Concrete Institute (ACI). Building code requirements for structuralconcrete (ACI 318-08) and commentary. ACI Committee 318, ACI; 2008.

[9] American Institute of Steel Construction (AISC). Specification for structuralsteel buildings (ANSI/AISC 360-10). Chicago; 2010.

[10] ACI-ASCE Committee 352. Recommendations for design of beam–columnconnections in monolithic reinforced concrete structures. ACI 352R-02, ACI;2002. p. 1–37.