Cyclic Loadind

J. Struct. Eng., 04015147

Cyclic Loading Test for Exterior Beam–Column Joints of CEFT Columns

Ho-Jun Lee1; Hong-Gun Park, M.ASCE2; Sung-Soon Park3; Sung-Bae Kim4; and In-Rak Choi5

Abstract: A cyclic loading test was performed to investigate the seismic performance of exterior beam-column joints of concrete-encased- and-filled steel tubular (CEFT) columns. Two specimens with steel beams and two specimens with precast concrete (PC) beams were tested under cyclic loading. The dimensions of the column cross section were 670 × 670 mm, which was a two-thirds-scale model of the prototype column. The test parameters for the steel beam and the PC beam were the beam depth (488 and 588 mm) and the flexural rebar ratio (1.1% and 1.5%), respectively. In the steel beam-CEFT column joints, continuity plates were used, while in the PC beam-CEFT column joints, couplers were used for the rebar connection. The steel beam-column connections failed due to early fracture of the continuity plate, and the PC beam-column connections failed due to rebar buckling and concrete spalling after beam yielding. The strength of the steel beam-column connections was evaluated using an existing design method, and the seismic performance of the PC beam-column connections was evaluated. On the basis of the test results, design considerations for beam-CEFT column joints were recommended. DOI: 10.1061/(ASCE)ST.1943- 541X.0001401. © 2015 American Society of Civil Engineers.

Author keywords: Beam-column joint; Concrete-encased and-filled steel tubular column; Steel beam; Precast concrete beam; Cyclic loading; Metal and composite structures.

Introduction

Recently, the use of large columns has increased in the construction of wholesale stores and warehouse buildings, which have long spans and tall story heights. In the construction of such large columns, the use of conventional reinforced concrete (RC) may not be economical due to the difficulty in the rebar fabrication and form- work. The use of precast concrete (PC) columns is also not advanta- geous due to the difficulty in the transportation and lifting of the heavy columns. As an alternative, hollow PC columns with reduced weight can be used. After the hollow PC columns are erected on the construction site, the hollow core can be easily filled with concrete. In manufacturing, however, due to the requirement of cross-ties, it is difficult to install and remove the inner form, which is used to form the hollow section. Considering the difficulty, a permanent thin steel tube can be used for the hollow PC column (Fig. 1). In the concrete-encased-and-filled tubular column (CEFT column), cross-ties are not necessary due to the existence of the steel tube, and the thin steel tube can be used as a structural element to resist

1Graduate Student, Dept. of Architecture and Architectural Engineer- ing, Seoul National Univ., 1 Gwanak-ro, Seoul 151-744, Korea. E-mail: [email protected]

2Professor, Dept. of Architecture and Architectural Engineering, Seoul National Univ., 1 Gwanak-ro, Seoul 151-744, Korea. E-mail: parkhg@ snu.ac.kr

3General Manager, Shinsegae E&C, 180 Jangchungdan-ro, Seoul 100-391, Korea. E-mail: [email protected]

4Manager, The Naeun Structural Engineering Co. Ltd., 21 Yangpyeong-ro, Seoul 150-105, Korea. E-mail: [email protected]

5Senior Researcher, Building Structure Research Group, POSCO Global R&D Center, 100 Songdogwahak-ro, Incheon 406-840, Korea (corresponding author). E-mail: [email protected]

Note. This manuscript was submitted on July 29, 2014; approved on July 14, 2015; published online on October 16, 2015. Discussion period open until March 16, 2016; separate discussions must be submitted for individual papers. This paper is part of the Journal of Structural Engineer- ing, © ASCE, ISSN 0733-9445/04015147(12)/$25.00.

member forces. The concrete encasement can develop additional strength and stiffness, providing local-buckling restraint and fire resistance to the steel tube.

However, because of the use of steel tube, the concrete encasement is separated from the core concrete. For this reason, the CEFT columns are susceptible to early spalling of the concrete encasement if adequate connections and reinforcement are not provided in the concrete encasement. Thus, the structural performance of the CEFT columns and the beam-column joints should be verified particularly under cyclic loading. The study of the CEFT column was performed in a previous paper (Park et al. 2015). Therefore, the present study focuses on the beam-column joint using the CEFT columns.

CEFT columns have been used in Japan for many decades, and the relevant test results were reported. However, the majority of the previous tests were performed for members (axial compression, flexural moment, and shear capacities), and the test results for beam-CEFT column joints were limited. Nakamura et al. (1999) tested steel beam-CEFT column joints using circular tubes, in which failure occurred in the columns. Fukuhara et al. (2010) tested steel beam-CEFT column joints using rectangular tubes, in which failure occurred in the beams. Although continuity plates were used in the joint, the load transfer mechanism between the steel beam and the CEFT column was not clarified. In the CEFT columns proposed in the present study, a more slender tube section and a thinner concrete encasement are used. Therefore, the structural design of the continuity plate is important in restraining the early damage of the concrete encasement and steel tube.

Recently, the use of CEFT columns has also been increasing in China. Liao et al. (2014) tested steel beam-CEFT column joints and RC beam-CEFT column joints. The major test parameters were the axial load level of the column and the moment capacity ratio between the beams and columns. However, since the dimensions of the steel tubes were small in the gross cross section, the advantages of the CEFT column were not fully used in the specimen.

Currently, specific design provisions for beam-CEFT column joints are not available. In the American Institute of Steel Construction

© ASCE 04015147-1 J. Struct. Eng.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


Fig. 1. Cross section of proposed CEFT column

(AISC) seismic provisions [ANSI/AISC 341-10 (AISC 2010)], comments were made only on the composite connections using RC columns, concrete-encased steel (CES) columns, and concrete- filled tubular (CFT) columns, on the basis of selected experimental studies.

In the present study, to evaluate the seismic performance, exterior beam-CEFT column joints using steel beams or PC beams were tested under cyclic loading. To maximize the economy and the structural capacity of the hollow PC column, thin plates were used for the steel tube, and the thickness of the concrete encasement was minimized. For the connection, continuity plates and couplers were used in the steel beam-column joints and the PC beam-column joints, respectively. On the basis of the test results, the load- carrying capacity, deformation capacity, energy dissipation, and failure modes were investigated.

Test Program

Various design and construction techniques are currently being evaluated for the implementation of the proposed composite columns and joints. In the present study, possible connection methods and details were proposed and tested to help guide this effort. Two specimens (BC1 and BC2) with steel beams and two specimens (BC3 and BC4) with PC beams were prepared for testing. The steel beam-CEFT column joints were considered for fast construction. The PC beam-CEFT column joints were considered for economy. The cross section of the tested columns (670 × 670 mm) was a two-thirds-scale model of the prototype columns (1,000 × 1,000 mm). On the other hand, the dimensions of the prototype beams were used for the tested beams, to evaluate the strength of the beam-column joints considering unfavorable design conditions.

The test parameters were the beam depth (H-488 × 300 × 11 × 18 or H-588 × 300 × 12 × 20, where H-beamdepth × flange width × web thickness × flange thickness) for the steel beam- column joints, and the flexural rebar ratio (1.1% or 1.5% for the negative moment) for the PC beam-column joints. Identical CEFT columns were used for all the specimens. In Fig. 2, the length of the beam between the column center and the vertical loading point was le¼ 3; 515 mm. The net beam length was 3,180 mm. The length of the column between the top and bottom hinges was he¼ 3,500 mm. The net column height excluding the rigid ends was 2,000 mm. Cyclic vertical loading was applied at the beam end (Fig. 2). Table 1 presents the test parameter. Figs. 3 and 4 show the details of the specimens.

The moment capacities in Table 1 were calculated using the plastic stress distribution for the steel beams and the strain compatibility method for the PC beams and CEFT columns. In the exterior beam-column connections, the moment capacity ratios

Fig. 2. Test setup (dimensions in mm)

(2.26–3.15) of the columns to the beams were relatively high. Thus, the load-carrying capacity of the specimens was expected to be determined by the moment capacity of the beam if early failure of the connections (such as fracture of the continuity plates and excessive bar bond-slip) did not occur.

In the CEFT column, a rectangular built-up steel tube with 8-mm-thick steel plates was used. The thickness of the concrete encasement was 110 mm. The ratio of the hollow area to gross section area of the CEFT column was 0.45. In the concrete encasement, four D25 longitudinal bars were placed at the corners of the section, and D13 bars were used for the ties at the vertical spacing of 180 mm. To enhance the bond between the steel tube and concrete encasement, shear studs (ϕ13) were welded to the tube plates. The proposed CEFT column was expected to be susceptible to early cracking and spalling of the thin concrete encasement. Thus, to prevent such failure, eight D16 longitudinal bars were placed and tack-welded to the shear studs (Fig. 3). The longitudinal bars were welded to the beam flanges, which provide additional resistance to the flange bearing. The thickness of the cover concrete for the ties and studs was 30 mm.

Since a steel tube is used, design considerations for the CEFT column are more similar to those for the CFT column than to those for the CES column. Therefore, continuity plates, which are used for steel beam-CFT column joints, were used in specimens BC1 and BC2 to transfer the tensile force of the steel beam flanges to the joint. The thickness of the continuity plate was 25 mm, which was greater than the beam flange thickness of 18 and 20 mm for BC1 and BC2, respectively. To ensure good concrete filling, the continuity plate had a square hole of 200 × 200 mm at the center, and 40-mm-diameter holes at the corners of the plate. The top and bottom continuity plate (thickness = 25 mm) and a web plate (thickness = 11 and 12 mm for BC1 and BC2, respectively) were ex- tended to 500 mm from the column face to form brackets, which were connected to the beams using welding for the flanges and bolting for the webs. To avoid early spalling of the concrete encasement, ties were closely spaced (Fig. 3) at the top and bottom of the column framed into the joint.

Because of the change of the flange thickness, a plastic hinge may occur at the splice between the bracket and the beam, which may cause fracture of the splice or excessive slip of the bolt connection. This was investigated on the basis of the test results.


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


Fig. 3. Dimensions and details of specimens BC1 and BC2 (dimensions in mm)

The column tube was connected to the continuity plate and web of the bracket by complete-joint-penetration (CJP) groove welding. The flanges and the web of the bracket were connected to the beam by CJP groove welding and bolting, respectively. The column longitudinal bars (two D16) and the joint ties (D13, spacing = 180 mm) were connected to the flanges and the web of the bracket, respectively by external fillet welding. The electrode E71 T-1 with tensile strength of 584 MPa and CVN toughness of 86 J at 0°C was used. The concrete encasement and the filled concrete were placed at the same time after the fabrication of steel and rebars.

In specimens BC3 and BC4 (Fig. 4), U-shaped PC shells (Im et al. 2013) were used for the beams. In Korea and several other countries, the U-shaped PC beam is frequently used to enhance the integrity of the beam and the joint by using cast-in-place concrete and to reduce the lift weight. The concrete casting in the PC beam- CEFT column joints was performed following the actual practice: first, concrete was cast for the concrete encasement of the lower column and the PC beam shell; the remaining concrete was then placed after assembly of the lower column and the PC shell. To restrain spalling of the concrete encasement, ties were closely spaced at the top and bottom of the lower and upper columns. The seated length of the PC shells on the concrete encasement of the lower column was 40 mm.

The anchorage part of the beam longitudinal bars was prefab- ricated with the steel tube column by inserting into the steel tube and fixing with couplers before concrete casting (Fig. 4). The beam longitudinal bars were then connected to the anchorage bars by the couplers. The couplers not only connect the beam longitudinal bars, but also transfer the bar force to the steel tube by direct bearing.

Further, a part of the bar force is transferred to the core concrete by the bar bond. To supplement the loss of the steel tube due to the drilled holes for the rebar penetration, 10-mm-thick steel plates were fillet-welded to the surface of the steel tube in the joint. In the case of interior beam-column joints, the beam longitudinal bars can be continuous through the column tube, without using couplers. In this case, like ordinary RC beam-column joints, the forces of the beam longitudinal bars can be transferred to the joint by the bond of the deformed bars.

In specimen BC3, six D29 and four D29 bars were used at the top and bottom of the beam cross section, respectively. In BC4, eight D29 and six D29 bars were used. In order to restrain local buckling of the longitudinal rebars under cyclic loading, D13 hoop bars were placed at 100-mm spacings in the region of twice the beam depth (i.e., in the potential plastic hinge zone), which satisfied the special moment frame requirements in ACI 318-11 (ACI 2011). The thickness of the top cover concrete was 35 mm. For convenience in fabrication, transportation, and lifting, the PC beam was connected to the CEFT column at the column face. Thus, to transfer the beam shear force in the column surface, shear studs of 4 − ϕ19 (sectional area ¼ 283 mm2) were welded to the steel tube. The nominal shear capacity of the studs was 453 kN. To alleviate the stress concentration at the head of the shear studs, the upper and lower studs were designed with different lengths: 250 and 200 mm.

The material properties of the concrete, steel, and rebars are pre- sented in Table 1. All the values indicate the average of the results obtained from three compression or tension tests. The maximum size of coarse aggregates was limited to 19 mm, considering the thickness of the concrete encasement. The compressive strength


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

rig

hts

rese

rved

.


Fig. 4. Dimensions and details of specimens BC3 and BC4 (dimensions in mm)

of the concrete cylinders was measured on the day of testing. The compressive strengths of concrete were 35.0 and 43.5 MPa in the columns of BC1 and BC2, respectively, and 43.5 and 39.2 MPa in the 1st casting (lower column and U-shell beam) and 2nd casting (joint, upper column, and core of the beam), respectively, for BC3 and BC4. The yield strengths of steel plates were 430 MPa for 8-mm-thick tubes; 437 MPa (409 MPa) for 11 mm (12 mm) thick web plates; 382 MPa (406 MPa) for 18 mm (20 mm) thick flange plates; and 421 MPa for the continuity plates. The yield strengths of rebars were 479 MPa for D13 bars and 543 MPa for D29 bars.

Table 2 presents the displacement-control loading program for the test, which is specified in the Korean Building Code 2009 (AIK 2009) for steel structures. The loading program is the same as the seismic qualification loading sequence in section K2.4 b of ANSI/AISC 341-10 (AISC 2010). Cyclic vertical loading was applied at the beam end using an actuator. The loading was first applied in the downward direction (negative loading), which caused the negative moment in the beam. In order to restrain the lateral torsional buckling of the steel beams under cyclic loading, lateral supports were provided for the steel beams. As shown in Fig. 2, LVDTs were used to measure the displacement of the loading point, rigid-body motion, and deformation of the joint. Uniaxial strain gauges were used to measure the strains of the steel plates and rebars.

Test Results

Vertical Load-Displacement Relationship

Fig. 5 shows the vertical load-displacement relationship of the specimens at the loading point. In the figure, the drift ratio was calculated by dividing the vertical displacement by the effective beam length le¼ 3; 515 mm (from the column center to the loading point). Generally, BC1 and BC2 with steel beams showed brittle behavior caused by connection failure. On the other hand, BC3 and BC4 showed ductile behavior due to the development of a full plastic hinge zone at beam end. Pinching is also observed in the cyclic response of BC3 and BC4, which indicates that unfavorable bar bond-slip occurred in the beam-column joint.

In BC1 and BC2, the negative loading of the beam attained the maximum strength at 4% and 3% drift ratio, respectively. On the other hand, the positive loading reached the peak load at 3% drift ratio in both specimens. After the peak strength, the load-carrying capacity quickly decreased due to the failure of the connection between the steel beam and column tube. In BC3 and BC4, the negative loading of the beam reached the peak strength at 3% drift ratio, and reached the maximum deformation (corresponding to 80% of the peak strength) at 6% drift ratio. On the other hand, the positive loading reached the peak load at 4% and 3% drift ratio, respectively, and reached the maximum postpeak deformation at


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

rig

hts

rese

rved

.


Table 1. Properties of Test Specimens and Materials

Property

Beam

Specimen

BC1 BC2 BC3 BC4

Type Steel beams (H-section) PC beams (U-type shell)

Nominal dimensions (mm) 488 × 300 × 11 × 18 588 × 300 × 12 × 20 500 × 700 Yield/tensile strength of flange (MPa) 382=507 406=555 —Yield/tensile strength of web (MPa) 437=541 409=557 —Yield/tensile strength of continuity plate (MPa) 421=542 —Concrete strength (MPa) — — 43.5 (1st casting)

39.2 (2nd casting)Top rebars — — 6-D29 8-D29Bottom rebars — — 4-D29 6-D29Yield/tensile strength of rebars (MPa)Positive flexural capacitya (kN · m)

— — 543=676 1,166 1,688 591

843

Negative flexural capacitya (kN · m) (1,675 for bracket) (2,115 for bracket) 1,179 1,491Column

Nominal dimensions (mm) Steel tube (mm)

− 670 × 670

— 450 × 450 (8 mm thickness) Concrete strength (MPa) 35.0 43.5 43.5 (1st casting)

39.2 (2nd casting) Yield/tensile strength of steel tube (MPa) 430=538

Longitudinal rebars 4-D25, 8-D16Flexural capacitya (kN · m) 1,834 1,905 Column to beam moment ratio 3.15 2.26

aCalculated using actual material strengths. bNegative plastic moment of the beam was considered.

1,841 3.12b 2.47b

□ □

Table 2. Cyclic Loading Program

Number of cycles Drift ratio (%)

Displacement at loading point (mm)

was close to 1.0 until the drift ratio reached 3% and 2%, respectively. After the peak load, significant strength degradations occurred. In the PC beam-column connections, the strength under the negative loading gradually decreased as the drift ratio increased,

6 cycles ±0.375 13.2 6 cycles ±0.5 17.6 6 cycles ±0.75 26.4 4 cycles ±1.0 35.2 2 cycles ±1.5 52.7 2 cycles ±2.0 70.3 2 cycles ±3.0 105.5 2 cycles ±4.0 140.6 2 cycles ±5.0 175.8 2 cycles ±6.0 210.9 2 cycles ±7.0 246.1

6% drift ratio when local buckling occurred in the top bars of the beam. Although the plastic moment capacity of the beam was almost identical in BC1 and BC3 (Table 1), the actual load-carrying capacity of BC1 was greater than that of BC3, due to the greater strength of the steel bracket used in BC1.

Table 3 presents the summary of the test results. As shown in Fig. 6, the yield displacement Δywas defined using the secant stiffness connecting the origin and 75% of the peak load (Park 1988). Maximum displacement Δuwas defined as the postpeak displacement corresponding to 80% of the peak strength (Park 1988). The maximum drift ratios were 3.4–4.6% and 5.5–6.2%, for the steel beam-column joints and the PC beam-column joints, respectively. Ductility values (μ ¼ Δu=Δy) were μ = 2.08–2.45 and

3.60–5.72, respectively. Fig. 7 shows the strength degradation due to cyclic loading at each

drift ratio. The strength ratio was defined as the ratio of the 2nd-load cycle strength to the 1st-load cycle strength at each drift ratio. Generally,as the drift ratio increased, the strength ratio decreased. In BC1 and BC2 with the steel beam, the strength ratio


due to the bond-slip of rebars and spalling of the cover concrete at the bottom of the beam [Fig. 7(a)]. On the other hand, in the positive loading direction, the strength degradation was not significant [Fig. 7(b)].

Damage Modes

Fig. 8 shows the damage of the specimens at the end of the test, and Fig. 9 shows the failure modes of the specimens. In Fig. 8, the concrete encasement of BC1 and BC2 was severely damaged, which indicates that large deformation occurred in the steel tube. At the beginning of the test (0.375% drift ratio), fine diagonal cracks of less than 0.1 mm occurred in both the joint faces parallel and orthogonal to the beam direction. As the displacement increased, the number of cracks increased. At 1.5% drift ratio, initial yielding

of the beam flange of BC1 occurred. In BC1 and BC2, the concrete cracks were less than 0.3 mm at 1.5% and 1.0% drift ratios, respectively. At 2.0% and 1.5% drift ratios in BC1 and BC2, respectively, local fracture occurred at the corners of the transition between the continuity plate and the bracket flange, which was detected from the strain gauge measurement. For this reason, the concrete cracks significantly increased particularly at the joint face connected to the beam. The maximum crack width was 0.6 mm. However, even after the local fracture, the load-carrying capacity increased due to the effect of stress redistribution (Fig. 5). At 3% and 2% drift ratios in BC1 and BC2, respectively, spalling of the cover concrete occurred [Figs. 9(a and b)]. Further, at 4% and 3% drift ratios in BC1 and BC2, respectively, the concrete encasement was completely delaminated, the inner steel tube yielded, and the specimens exhibited the maximum load. The steel tube showed a significant out- of-plane deformation in the beam direction.


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


Fig. 5. Vertical load-displacement relationships of specimens

Table 3. Summary of Test Results

Load-carrying capacity Deformation capacity Yield stiffness

Specimen

Predicted strength, Pn(kN)

Maximum strength Pu, kN (%)

Yield displacement Δy, mm (%)

Maximum displacement Δu, mm (%)

Ductility μ (¼ Δu=Δy)

ky(¼ Pu=Δy) (kN=mm) Failure mode

BC1 (+)

435

455 (3.0)

63.2 (1.8)

155 (4.4)

2.45

7.20

Failure of

(−) BC2

−435 −448 (3.9) −70.6 (2.0) 162 (4.6) 2.29 6.35 continuity plate

(+) 630 539 (2.4) 59.2 (1.7) 123 (3.5) 2.08 9.12 Failure of (−)

BC3 −630 −521 (2.9) −55.6 (1.6) 120 (3.4) 2.16 9.37 continuity plate

(+) 186 242 (4.0) 38.3 (1.1) 219 (6.2) 5.72 6.32 Beam flexure (−)

BC4 −370 −409 (3.0) −46.2 (1.3) 210 (6.0) 4.55 8.85

(+) 256 340 (3.1) 54.1 (1.5) 195 (5.5) 3.60 6.28 Beam flexure (−) −469 −528 (3.0) −54.5 (1.6) 213 (6.1) 3.91 9.69

In BC3 and BC4, plastic hinges developed at the beam ends. The failure mode of BC4 was similar to that of BC3. Thus, only the failure mode of BC3 is shown in Fig. 8. In Fig. 8, the top of the beam was severely damaged due to local buckling of the top longitudinal bars. Furthermore, flexural cracks developed more at the

top of the beam than at the bottom of the beam. At the bottom of the beam, the beam-column interface was locally damaged.

At the beginning of loading (0.375% drift ratio), a few flexural cracks of less than 0.1 mm developed within the distance of 1,000 mm from the column face. The number of cracks gradually


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

rig

hts

rese

rved

.


Fig. 6. Definitions of yield displacement and maximum displacement

increased and the cracks were propagated toward the joint as the displacement increased. At 0.75% drift ratio, a gap was observed at the joint interface where the PC beam shell was seated on the concrete encasement of the lower column. The development of the gap is attributed to the bond-slip of the beam bottom bars. At 1.0% drift ratio, the top and bottom bars first yielded. The flexural cracks in the beam grew to 0.7 mm. The maximum crack width in the

beam-column joint was 0.3 mm. After yielding of the flexural bars, cracks were concentrated at the beam end. As the shear force in the joint was resisted mainly by the inner CFT, shear cracks in the joint face did not further propagate. At 3% drift ratio, spalling of concrete occurred at the bottom of the PC beam shell. The spalling of the bottom cover concrete was the main cause of the degradation of the negative load-carrying capacity. At 3% drift ratio, the negative load-carrying capacity reached the maximum load, and the maximum crack width exceeded 1.0 mm.

Ultimately at 6% drift ratio, the load-carrying capacity significantly decreased due to the local buckling of the top flexural bars. Since the bottom bars had a relatively thick concrete cover of the PC shell, local buckling of them did not occur. Failure of the couplers did not occur until the end of the test.

In the exterior joint face at the opposite side of the beam connection, spalling of the cover concrete occurred at the location of the bottom flexural bars, and the bottom couplers were exposed [5% drift ratio, Fig. 9(c)]. This result indicates that the couplers did not effectively prevent bond-slip of the flexural bars subjected to compression. Under repeated cyclic loading, residual tensile strain increased, particularly in the bottom bars, which had less area than the top bars. Due to the force equilibrium in the cross section, the residual strain of the bottom bars was greater than that of the top bars with the greater area. Thus, as the bar length increased in the

Fig. 7. Cyclic strength degradation at 2nd load cycle: (a) negative loading; (b) positive loading

Fig. 8. Damage after the end of the test


(a) (b)

Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

rig

hts

rese

rved

.


(a) (b) (c)

Fig. 9. Failure modes of test specimens: (a) BC1; (b) BC2; (c) BC3, BC4

joint, the couplers were not completely in contact with the steel tube, and only the bar bond provided resistance against the compressive force. This result indicates that when high ductility is required under cyclic loading, the compression development length of the bottom bar needs to be satisfied in the joint.

Strains of Rebars and Steel Plates

Fig. 10 shows strains of the rebars and steel plates in BC1, BC2, and BC4. The strains of BC3 were similar to those of BC4.

In Fig. 10(a), the bracket flange of BC1 [No. 1 in Fig. 10(a)] remained elastic under cyclic loading, while the beam flange (No. 2) at the splice experienced large inelastic deformation. Yielding of the beam flange occurred at 1.5% drift ratio. In Fig. 10(d), both the bracket and beam flange of BC2 remained almost in the elastic range until the end of the test. In BC4 [Fig. 10(g)], both the top and bottom longitudinal bars exhibited large inelastic strains. In the figure, the strain of the top bar decreased when it was subjected to compression under reversed loading. On the other hand, the strain of the bottom bar did not significantly decrease under reversed

(f)

Fig. 10. Strains of rebars and steel plates: (a) strains of bracket and beam flanges (BC1); (b) strains of ties (BC1); (c) strains of tube flanges (BC1); (d) strains of bracket and beam flanges (BC2); (e) strains of ties (BC2); (f) strains of tube flanges (BC2); (g) strains of beam rebar (BC4); (h) strains of ties (BC4); (i) strains of studs (BC4)


(g) (h) (i)

(a) (b) (c)

(d) (e)

Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

rig

hts

rese

rved

.


γ ¼ ðδ − δ Þffiaffiffi

2ffiffiffiffiffiffiffiffiffiffi

2ffiffi

loading. This result indicates that significant bond-slip occurred in the bottom bars inside the joint. The maximum tensile strain of the top flexural bars, which failed in the local buckling, was about 4.7%.

Figs. 10(b, e, and h) exhibit the strains of the hoops in the joints of BC1, BC2, and BC4, respectively. The strains gradually increased as the cyclic loading proceeded. The strain significantly increased when diagonal cracking occurred in the joint. The strains of BC1 and BC2 reached the yield strain, while the strain of BC4 remained elastic.

Figs. 10(c and f) show the strains of the column tube plate near the continuity plate in BC1 and BC2, respectively. The strains were significantly greater than the yield strain, which indicates that the steel tube experienced a significant out-of-plane deformation in the beam direction. Fig. 10(i) shows the strain histories of the shear studs installed at the column surface for BC4. Ultimately, the strains were close to the yield strain.

Shear Deformation of Joint Panel Zone

Fig. 11(a) shows the average shear strain γpof the joint panel. The shear strain has positive values under the positive loading of the beam. The average shear strain of the panel zone was calculated as p

þ b=2ab, where δ1 and δ2 are diagonal deformations of the LVDTs in the joint panel and aand bare the vertical and horizontal distances between the LVDTs (Fig. 2).

The joint shear behavior of BC1 and BC2 was similar in both the positive and negative directions due to the symmetric section of the beam. On the other hand, in BC3 and BC4, the shear strains show greater values in the negative direction because the negative moment capacity of the PC beams was greater than the positive moment capacity (i.e., the joint shear force under the negative loading was greater).

To evaluate the joint shear deformation of BC1 and BC2, with a steel beam, the joint shear force-shear strain relationship was compared with the result of an existing prediction model [Fig. 11(b)]. Fukumoto and Morita (2005) proposed a prediction model that can describe the elastoplastic behavior of the joint panel in steel beam- CFT column connections. The model was followed the guidelines in Recommendations for Design and Construction of Concrete Filled Steel Tubular Structures (AIJ 2008). Fukumoto (2012) also proposed a prediction model for the joint panel of steel beam-CEFT column connections by modifying the existing model.

Fig. 11(b) compares the joint shear strains of BC1 and BC2 with the predictions (Fukumoto and Morita 2005; Fukumoto 2012). In the predictions of the model for CFT column (Fukumoto and Morita 2005), the average joint shear strain and joint shear strength at the yield point were predicted as (0.00378, 2,641 kN) for BC1

and (0.00378, 2,777 kN) for BC2. In the predictions of the model for the CEFT column (Fukumoto 2012), the yield points were predicted as (0.00378, 4,877 kN) and (0.00378, 5,511 kN). This result indicates that the joint shear strength was significantly increased by the concrete encasement. However, in the tests, failures of BC1 and BC2 occurred due to fracture of the connection between the continuity plate and the tube, rather than due to the shear strength of the joint panel. Thus, in Fig. 11(b), in the range of the experimental shear strains, the predictions showed only the elastic behavior. Nevertheless, Fig. 11(b) shows that the initial stiffness of the test specimens agreed with the predictions.

Evaluation of Structural Performance

Load-Carrying Capacity

Fig. 5 shows the predicted ultimate loads Pþn and P−

n in the positive and negative directions. The predicted loads were calculated based on the plastic moment of the beam: Pn¼ Mb=l0, where Mb= the moment capacity of the beam and l0 ¼ 2,680 mm for BC1 and BC2 and 3,180 mm for BC3 and BC4. In the case of steel beam- column joints, the peak strength of BC1 with the smaller steel beam agreed with the predicted load, while in BC2 with the larger steel beam, the peak loads were less than the predictions (Fig. 5). This is because in the case of BC2, the experimental peak loads were determined by the strength of the continuity plate-to-tube connection rather than by the plastic moment of the beam. In the case of BC3 and BC4 with the PC beam, the test peak strengths agreed with the predictions.

Connection Strength of BC1 and BC2

The test results of BC1 and BC2 showed that failure occurred at the connection between the column tube and the continuity plate. To evaluate the safety of the connection, an existing design method for CFT columns was applied to the CEFT columns, neglecting the contribution of the concrete encasement, which showed early spalling. In the Recommendations for Design and Construction of Concrete Filled Steel Tubular Structures (AIJ 2008), a design model (Fukumoto 2007) based on yield line theory was recommended for the design of the steel beam-to-tube connection. Fig. 12 shows the load transfer mechanism of the beam-to-tube connection using a continuity plate. The total tensile strength Ttotal of the connection is defined as the sum of the following contributions: strength of the continuity plate Tc; out-of-plane strength of the tube flange in the horizontal and vertical direction Txand Ty, respectively; and the strength of the tube web Tw

(a) (b)

Fig. 11. Shear strain of joint: (a) average shear strains of joint specimens; (b) comparisons of test results and predictions


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


(a) (b)

Fig. 12. Load transfer mechanism between beam flange and tube (adapted from Fukumoto 2007): (a) steel tube; (b) continuity plate

Table 4. Strengths of Beam Flange-to-Tube Connections

Predictions (Fukumoto 2007) Test results

Yield strength (2% drift Contribution of components for

yield/ultimate strength (kN) Yield/ultimate ratio for BC1 and 1.5%

drift ratio for BC2) Ultimate strength

moment capacity Py My Pu Mu

Specimen Ty Tx Tw Tc Ttotal Mc;y=Mc;u(kN · m) (kN) (kN · m) My=Mc;y (kN) (kN · m) Mu=Mc;u

BC1 74=139 68=128 836=1,911 2,104=2,709 2,246=2,976 1,096=1,452 (+) 407 1,294 1.18 (+) 455 1,447 1.00 BC2 74=139 68=128 836=1,911 2,104=2,709 2,246=2,976 1,321=1,750 (+) 446 1,418 1.07 (+) 539 1,714 0.98

Ttotal ¼ Tyþ minðTx; TwÞ þ Tc ð1Þ

Txand Twact in the same load transfer path. Thus, the smaller of Tx and Twwas used. In Fig. 12(b), the resistance of the continuity plate is divided into area Asubjected to direct tension and area Bsubjected to less tensile deformation. In the present study, considering the air holes at the corner of the continuity plate, only area Awas considered.

Table 4 presents the calculated strengths of the connections according to the AIJ recommendations. For the ultimate state, the total tensile strength of the beam flange-to-tube connection was Ttotal ¼ 2,976 kN in both BC1 and BC2. The majority of the resistance was attributed to Tc¼ 2,709 kN of the continuity plate, because relatively thin steel plates were used for the column tube. The resultant moment capacity of the connection was calculated as Mc;u¼ Ttotal hb, where hb= depth of the steel beam: Mc;u¼ 1,452 kN · m for BC1 and Mc;u¼ 1,750 kN · m for BC2. As pre- sented in Table 4, the predicted strengths agreed with the ultimate strength of the specimens: Mu¼ 1,447 kN · m for BC1 and Mu¼ 1,714 kN · m for BC2. This result indicates that in the evaluation of the strength of steel beam-CEFT column connection, the contribution of the concrete encasement should be neglected.

Fig. 13 shows a comparison between the moment capacities and demands of BC1 and BC2, which were based on the strengths of the beam and the beam-tube connection. The moment in the vertical axis was normalized by the beam plastic moment. In BC1, the moment based on the connection strength was slightly higher than the moment demand corresponding to the beam plastic hinge moment. This result indicates that the initial yielding occurred at the beam plastic hinge zone, and the tensile failure of the connection can occur as the plastic hinge moment of BC1 is increased by the cyclic strain hardening effect. The failure mode agrees with the actual failure mode of BC1. In BC2, on the other hand, the moment based on the connection strength is significantly smaller than the moment corresponding to the beam plastic moment capacity. Thus, the

failure mode of BC2 is determined by the connection strength. The failure mode agrees with the test result.

However, in the actual design, such failure modes of BC1 and BC2 should be avoided because the failure at the connection causes severe spalling of concrete encasement and inelastic deformation of steel tube in the joint, which is difficult to repair, and decreases

Fig. 13. Moment capacity and demand of BC1 and BC2


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


test results. In this study, the reference load cycle was chosen as the second load cycle at 4.0% drift ratio because only two load cycles were used in the loading program (Table 2). BC3 and BC4 satisfied the acceptance criteria for both stiffness and energy dissipation.

Design Considerations

Fig. 14. Acceptance criteria for stiffness and energy dissipation in ACI 374.1-05

the deformation capacity of the beam-column joint. Under repeated cyclic loading, the plastic moment of the beam can be increased by the cyclic strain hardening effect. Thus, the connection between the beam and tube needs to be conservatively designed with a greater strength. In steel beams, the strain hardening factor is generally taken as 1.1 [ANSI/AISC 341-10 (AISC 2010)].

Seismic Design Code Conformance of BC1 and BC2

The seismic performance of BC1 and BC2 was evaluated according to their conformance to the requirements of ANSI/AISC 341-10 (AISC 2010). The seismic design code requires demonstration of 80% of the nominal beam flexural strength at a 4% drift ratio for composite special moment frames (C-SMF). In the case of BC1, the requirement was satisfied, though ultimately it failed due to the fracture of the continuity plate. In the case of BC2, on the other hand, the majority of the inelastic deformation was developed by the out-of-plane deformation of the steel tube, rather than by the inelastic deformation of the beam. Thus, the performance of BC2 was not evaluated.

Seismic Design Code Conformance of BC3 and BC4

In order to evaluate the seismic performance of BC3 and BC4 with the PC beams, the test results were compared with the requirements specified in ACI 374.1-05 (ACI 2005). The acceptance criteria of ACI 374.1-05 (ACI 2005) are shown in Fig. 14. The secant stiffness k3.5 between −0.35% and þ0.35% drift ratios should be greater than 0.05ki (ki = initial stiffness) in the third load cycle of the 3.5% drift ratio. The energy dissipation ratio β (¼ Ed=Ed0, where Ed= actual energy dissipation and Ed0 = idealized energy dissipation assuming elastoplastic behavior) should be greater than 0.125 in the third load cycle of the 3.5% drift ratio. Table 5 presents the

Table 5. Evaluation of Seismic Performance of BC3 and BC4 (ACI 374.1-05)

Stiffness (kN=mm) Energy dissipation (kN · mm)

Actual

On the basis of the test results, the following design considerations for steel beam-CEFT column joints are recommended:

1. In order to restrain early fracture of the continuity plate at the joint, the ultimate strength of the connection between the beam and tube should be greater than the expected maximum demand force, which includes the effects of the material overstrength and strain hardening of steel beam under cyclic loading. The required tensile strength is expressed as 1.1Mpc;exp=hb, where Mpc;exp is the expected flexural moment at the column face corresponding to the beam plastic moment. The factor 1.1 indicates a strain hardening effect of the steel beam [ANSI/AISC 341-10 (AISC 2010)]. Otherwise, the maximum demand force needs to be decreased by using a plastic relocation method such as the reduced beam section connection.

2. Although fracture of the continuity plate (thickness = 25 mm) occurred in the small-scale test specimens, in the prototype column (cross section: 1,000 × 1,000 mm), the strength of the continuity plate Tc is increased by the greater dimensions. Thus, if the same steel beams are used, the 25-mm-thick continuity plate is expected to provide adequate strength for the steel beam-prototype column joint.

3. Although the web of the bracket was welded to the tube plate, due to the out-of-plane deformation of the thin tube plate and early delamination of the cover concrete, the plastic moment of the web may not be properly transferred to the column. Thus, for safe design, the continuity plate (i.e., the connection strength) and the bracket flange need to be designed for the full beam plastic moment including the contribution of the web.

4. In order to prevent any stress concentrations, the use of a tapered bracket for smooth transition from the continuity plate to the beam flange is recommended. Further, it is recommended to use a circular opening for placing the core concrete rather than a rectangular one so that strength loss of the continuity plate can be minimized.

5. If properly reinforced with longitudinal bars, ties, and studs, 110-mm-thick concrete encasement can be used for the flexure-axial strength of the CEFT columns, and for the shear strength of the beam-column joint. However, the contribution of the thin concrete encasement should be neglected in the evaluation of the ultimate tensile strength of the connection between the steel beam and tube.

Design considerations for PC beam-CEFT column joints are recommended as follows:

1. To transfer the beam shear force in the column surface, the use of multiple shear studs with different lengths is recommended (in the test specimens, ϕ19 studs with lengths of 250 and

Specimen

BC3

ki 0.05kik3.5 at 4.0%

k3.5 = 0.05ki

Elastoplastic energy, Ed0

energy, Ed

β ¼ Ed= Ed0

200 mm were used). 2. For exterior joints, couplers can provide adequate tension

anchorage for the beam longitudinal bars. However, since cou- (+) 8.62 0.431 1.53 3.55 133,059 36,712 0.276 (−) 12.3 0.625 1.55 2.48

BC4 (+) 9.28 0.464 2.22 4.78 163,544 46,026 0.281 (−) 13.8 0.690 2.10 3.04

plers cannot provide compression anchorage, compression bar development length should be satisfied, particularly when high ductility is required under cyclic loading.

3. To enhance the deformation capacity of the PC beams, local buckling of the top flexural bars should be restrained. In the


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.


bottom bars having a relatively thick concrete cover of the PC shell, local buckling does not occur.

Conclusions

A cyclic loading test was performed for exterior beam-to-column joints using concrete-encased-and-filled tubular (CEFT) columns. Two specimens using steel beams and two specimens using PC beams were tested. The major findings of this study are summa- rized as follows:

1. In BC1 and BC2 with steel beams, brittle failure occurred in the continuity plate-to-column tube connection. Consequently, the concrete encasement and steel tubes of the CEFT column were severely damaged.

2. BC3 and BC4 with PC beams showed good deformation capacity, developing the full moment capacity in the beam plastic zone. However, in large inelastic deformations, significant bond-slip occurred in the beam rebars inside the joint, which decreased the energy dissipation capacity. Damage to the concrete encasement and steel tube was relatively limited.

3. Due to the lack of concrete integrity in the beam-column tube connection of BC3 and BC4, the bottom part of the PC beam was vulnerable to early spalling. Further, significant plastic elongation occurred in the beam bottom bars inside the joint. Thus, the bar couplers were no longer effective for the compression anchorage of the beam bars.

4. In BC1 and BC2, the strength of the beam flange-column tube connection was evaluated by using an existing design method for CFT columns. The evaluations of the ultimate strength and failure mode agreed with the test results.

5. The energy dissipation of the PC beam-column connections, BC3 and BC4, was decreased by the bond-slip of the beam flexural bars. However, the test results satisfied the acceptance criteria of ACI 374.1-05 (ACI 2005).

6. On the basis of the test results, several design recommendations were made: for the design of steel beam-CEFT column joints, the use of tapered continuity plates and circular concrete placement cutouts in continuity plates; and for the design of PC beam-CEFT column joints, the use of compression development lengths for the top rebars and multiple shear studs with different lengths for shear transfer between the beam and joint.

Acknowledgments

This research was financially supported by Shinsegae E&C. Authors are also grateful to grants from the High-Tech Urban

Development Program (09 R&D A01) and the R&D Policy Infra Program (Code 11-Technology Standardization-09-01), funded by the Ministry of Land, Infrastructure, and Transport (MOLIT) of the Korean Government.

References

ACI (American Concrete Institute). (2005). “Acceptance criteria for

moment frames based on structural testing and commentary.” ACI 374.1-05, Detroit.

ACI (American Concrete Institute). (2011). “Building code requirements for structural concrete and commentary.” ACI 318-11, Detroit.

AIJ (Architectural Institute of Japan). (2008). “Recommendations for design and construction of concrete filled steel tubular structures.” Tokyo (in Japanese).

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

AISC (American Institute of Steel Construction). (2010). “Seismic provisions for structural steel buildings.” ANSI/AISC 341-10, Chicago.

Fukuhara, T., Miyauchi, Y., Hiroshige, T., and Muto, K. (2010). “Seismic behavior of reinforced concrete covered CFT structural elements. Part 2: Performance of steel beam-column joints.” Summaries of Technical Papers of Annual Meetings, Architectural Institute of Japan, Tokyo, 1361–1362 (in Japanese).

Fukumoto, T. (2007). “Local elasto-plastic behavior of steel beam to concrete-filled square steel tube column moment connections.” J. Struct. Constr. Eng., 617, 177–184 (in Japanese).

Fukumoto, T. (2012). “Elasto plastic behavior on panel zone within concrete-encased and concrete-encased-and-filled steel tubular column beam-to-column moment connections with steel beam.” J. Struct. Constr. Eng., 77(678), 1339–1345 (in Japanese).

Fukumoto, T., and Morita, K. (2005). “Elastoplastic behavior of panel zone in steel beam-to-concrete filled steel tube column moment connections.” J. Struct. Eng., 10.1061/(ASCE)0733-9445(2005)131:12(1841), 1841–1853.

Im, H. J., Park, H. G., and Eom, T. S. (2013). “Cyclic loading test for reinforced-concrete-emulated beam-column connection of precast concrete moment frame.” ACI Struct. J., 110(1), 115–126.

Liao, F. Y., Han, L. H., and Tao, Z. (2014). “Behaviour of composite joints with concrete encased CFST columns under cyclic loading: Experi- ments.” Eng. Struct., 59, 745–764.

Nakamura, Y., Matsuo, A., and Ueura, K. (1999). “Ultimate strength of plastic deformation capacity of CFT columns with covering RC considering material combination.” J. Technol. Des., 7, 39–44 (in Japanese).

Park, H. G., Lee, H., Choi, I., Kim, S., and Park, S. (2015). “Concrete-filled steel tube column encased with thin precast concrete.” J. Struct. Eng.,

10.1061/(ASCE)ST.1943-541X.0001303, 04015056. Park, R. (1988). “State-of-the-art report on ductility evaluation from labo-

ratory and analytical testing.” Proc., 9th World Conf. on Earthquake Engineering, Vol. 8, International Association for Earthquake Engineer- ing (IAEE), 605–616.


Dow

nloa

ded

from

asc

elib

rary

.org

by

Gaz

i Uni

vers

ites

i on

10/2

2/15

. Cop

yrig

ht A

SC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Documents

Cyclic Loadind