Seismic behavior of bidirectional bolted connections for CFT columns and H-beams

Engineering Structures 29 (2007) 395 407 www.elsevier.com/locate/engstruct Seismic behavior of bidirectional bolted connections for CFT columns and H-beams Lai-Yun Wu a,, Lap-Loi Chung b, Sheng-Fu Tsai a, Chien-Fan Lu a, Guo-Luen Huang a a Department of Civil Engineering, National Taiwan University, Taipei 106, Taiwan b National Center for Research on Earthquake Engineering, and Department of Civil Engineering, National Taiwan University, Taipei 106, Taiwan Received 17 February 2005; received in revised form 10 February 2006; accepted 4 May 2006 Available online 5 July 2006 Abstract Concrete-filled tubes (CFTs) have the advantages of both steel tubes and concrete. For that the confinement of the steel tube on the concrete would enhance the stiffness, strength and ductility of the concrete; also, the filled concrete can decrease the possibility of inward buckling of the steel. In this paper, bidirectional bolted beam-to-column connections for CFTs are proposed. A mechanical model is established to derive theoretical equations for calculating the stiffness, the yielding shear strength and ultimate shear strength of the panel zone. Also, a series of cyclic loading experiments has been conducted to verify it. The experimental results demonstrate that the bidirectional bolted connections have superior earthquake resistance in stiffness, strength, ductility and energy dissipation mechanism. These results indicate that the seismic resistance exceeds those specified in the seismic design codes of Taiwan and the US. Therefore, the bidirectional bolted beam-to-column connection has excellent seismic resistance, and this structural system can perform well as expected and be put into practice. c 2006 Elsevier Ltd. All rights reserved. Keywords: Seismic behavior; Concrete-filled tube (CFT); Bidirectional bolted beam-to-column connection; Earthquake resistance 1. Introduction Concrete-filled tubes (CFTs) have the advantages of both steel tubes and concrete. The steel tube provides the confinement and thus increases the stiffness and strength of concrete [1]. Meanwhile, the concrete reduces the possibility of the local buckling of the tube wall. Besides, the steel tube column eliminates the use of formwork during construction. In the Northridge Earthquake, US, 1994, many steel buildings suffered from brittle fracturing in the moment connections [2,3]. Therefore, ways to improve the design for the beam-to-column connections to avoid connections failure before the ductility of the beam, column and the panel zone is developed have been the focus of recent researches [4 8]. The beam-to-column connections used with CFTs can be classified broadly into two categories. The most convenient connection is to attach the steel beam directly to the skin of the steel tube or through the diaphragm plate. From the Corresponding address: 1, section 4, Roosevelt Road, Taipei 106, Taiwan. E-mail address: lywu@ntu.edu.tw (L.-Y. Wu). experimental and analytical results of Alostaz and Schneider [9, 10], severe tube wall distortions prohibited the development of the plastic flexural capacity of the beam and caused very large stresses and strains on the flange weld and tube wall. For the other connection category, the beam flange, web or the entire cross section is penetrated through the steel tube or the beam end is welded with anchorage which is embedded in the CFT column. In the research of Alostaz and Schneider [9,10], there were four methods. The first one was embedding weldable deformed bars. For the second method, headed studs were attached to the inside walls at the beam flange. The third method is extending the web plate into the concrete core with attached headed studs. For the fourth method, continuing the beam through the depth of the CFT column was considered to be the best rigid connection type. In the research of Schneider [11], the last connection type had the best seismic-resistant behavior. The related researches showed that the second category of beam-to-column connections have better seismic resistance. However, construction difficulty is the disadvantage of these 0141-0296/$ - see front matter c 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.engstruct.2006.05.007

396 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 Fig. 1. Schematic diagram of proposed beam-to-column connection. connection types. The connection type studied in this paper is the bolted type, as shown in Fig. 1. The ends of H-beams are welded with end-plates. The beam end-plates are then bolted to the rectangular steel tube in both east west and north south directions. However, the tie-rods locate at different elevations for the east west (E W) and north south (N S) directions. After the compressive strength of the concrete is fully developed, the tie-rods are pre-stressed. In addition, flange wing plates and upstanding ribs are welded to the end of the steel beam. The use of flange wing plates moves the plastic hinges away from the welding zone, and the reinforced upstanding rib reduces the prying action. After the columns have been erected in the field, the distance between two adjacent columns is slightly increased with a jack allowing enough space for the horizontal beam in a single segment to be hoisted in place and bolted to the columns. The beam can also be hoisted and assembled in two or three segments as in Fig. 2. The construction of bolted beam-tocolumn connections emphasizes that the welding is completed in a factory and no welding is necessary at the construction site. These construction procedures reduce the construction uncertainty and raise the construction quality for assuring the seismic-resistant performance of the concrete-filled steel tube column and beam. Bolted connections for CFT building structures were first proposed by Wu et al. [12]; several experiments with unidirectional bolted connections were performed, which proved that the bolted connection has good earthquakeresistance behavior. In this research, a mechanical model of a bidirectional bolted connection, which is closer to the real situation, is established to evaluate the stiffness, the yielding shear strength and ultimate shear strength of the panel zone of the connection. In the experiments, three sets of specimens are bolted in two directions. This is expected to be more realistic in simulating the behaviors of 3D buildings. Each specimen with different width-to-thickness ratios is acted upon with cyclic loading to investigate the stiffness, strength, ductility and energy dissipation of the bidirectional bolted connection for the CFTs. Fig. 2. Construction and erection of beam-to-column connection at a construction site. 2. Mechanical model for panel zone of bolted connections A rectangular concrete-filled steel tubular column is made of steel and concrete. Both materials contribute to the shear stiffness, yielding shear strength and ultimate shear strength of the beam-to-column connection. They can be assumed to have independent behavior [13]. The shear stiffness of the penal zone (K ) is the superposition of the shear stiffnesses of steel (K s ) and concrete (K c ) while the shear strength (V ) is the superposition of the shear strengths of steel (V s ) and concrete (V c ). 2.1. Steel tube The rectangular steel tube consists of column flanges and webs, and the panel zone is confined by four end-plates. The column flanges and the E W beam end-plates are combined tightly together under the compressive pre-stress of the bolts. Thus, they can be considered as a single body known as the generalized column flanges. Similarly, the column webs and the N S end-plates can be considered as a single body known as the generalized column webs. When the structure is subjected to horizontal loads, shear force is generated in the panel zone. This shear force causes shear deformation to the generalized column webs and flexural deformation to the generalized column flanges, as shown in Fig. 3. Thus, the shear stiffness and shear strength of the panel zone are contributed to by the shear behavior of the webs and the flexural behavior of the flanges. When the two generalized column flanges are subjected to shear force, they can be simulated as beams with fixed ends to resist flexural deformation. Hence, the shear stiffness of the two generalized column flanges ( K f ) due to flexural deformation is: 12E s Ī f K f = (2) (d b t bf ) 2 = 2E sb c t 3 f (d b t bf ) 2 (1)

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 397 shear stiffness of the two N S end-plates ( K ep ) is: [( K ep = 1 nd ) ( ) ] h 1 ndh 1 1 + d b K ep d b K eph [ = 1 nd h + nd ] 1 h K ep = r w1k ep (5) d b d b r A where K ep and K eph are the shear stiffnesses of the regions with and without holes, respectively, in the two N S end-plates. The shear stiffness of the two generalized column webs ( K wep ) is: K wep = K w1 + K ep. (6) Fig. 3. Schematic diagram for deformation of steel tube and end-plate at the panel zone. where E s is the Young s modulus of steel; d b is the depth of the H-beam; t bf is the thickness of the beam flange; b c is the width of the column; t f is the thickness of the generalized column flange which is the sum of the column flange thickness (t f ) and the beam end-plate thickness (t ep ), or t f = t f + t ep ; Ī f = b c t 3 f /12 is the moment of inertia of the generalized column flange. Because of the existence of holes, the steel column web is divided into two regions with and without holes. The shear stiffness of the region without holes (K w ) in the two column webs is: K w = 2A w G s = 2(d c 2t f )t w G s (2) where A w is the cross-sectional area of the region without holes in the column web; G s is the shear modulus of steel; d c is the depth of the column; and t w is the thickness of the column web. The shear stiffness of the region with holes (K wh ) in the two column webs is: K wh = 2A wh G s = 2(d c 2t f md h )t w G s = r A K w (3) where A wh is the cross-sectional area of the region with holes in the column web; m is the number of holes in a row of bolts; d h is the diameter of the holes; r A = 1 md h /(d c 2t f ) is the area reduction factor due to the existence of holes. Both the regions with and without holes contribute in series to the shear stiffness of the column web. Thus, the shear stiffness of the two column webs ( K w1 ) is: K w1 = = [( 1 nd h d b ) 1 K w + [ 1 nd h + nd h d b d b r A ( ) ndh 1 d b K wh ] 1 ] 1 K w = r w1k w (4) where n is the number of rows of holes in the column web of the panel zone; r w1 is the first reduction factor for the shear stiffness of the two column webs due to the existence of holes. Similarly, both the regions with and without holes contribute in series to the shear stiffness of the N S end-plates. Thus, the The shear stiffness contributed by the rectangular steel tube (K s1 ) to the panel zone is the superposition of the shear stiffness of the generalized column webs ( K wep ) and that of the generalized column flanges ( K f ): K s1 = K f + K wep = K f + K w1 + K ep. (7) After pre-stressing, the maximum shear force that the two N S end-plates can resist is the friction F induced by the prestress T : F = 2T µ = 2(2mnt)µ = 4mntµ (8) where t is the pre-stress of each bolt; µ is the friction coefficient, which is usually between 0.3 and 0.4; the value 0.35 is adopted in this paper. From the specimens studied in this paper, before the region with holes in the N S end-plate yields, the shear force is already over the maximum static friction and the N S end-plate slips. Meanwhile, the shear strain γ 1 and the corresponding shear force V s1 in the panel zone, respectively, are: γ 1 = F/ K ep = F/(r w1 K ep ) (9a) V s1 = K s1 γ 1 = ( K f + r w1 K w )γ 1 + F = ( K f + K w1 + K ep )γ 1. (9b) After the two N S end-plates slip, the shear force contributed by the end-plates is the maximum friction. Therefore, before the region with holes in the column web yields, the shear stiffness contributed by the rectangular steel tube to the panel zone (K s2 ) is the superposition of the stiffness of the two generalized column flanges and that of the two simplex column webs: K s2 = K f + K w1 = K f + r w1 K w. (10) When the region with holes in the column web yields, the strength of the two column webs (V why ) is: V why = 2A wh τ sy = 2(d c 2t f md h )t w F y 3 (11) where τ sy = F y / 3 is the yielding shear stress of steel; F y is the yielding tensile stress of steel. The corresponding shear strain and the shear force of the steel tube in the panel zone,

398 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 respectively, are: γ 2 = V why K w1 = V why r w1 K w V s2 = K s2 γ 2 + F = ( K f + r w1 K w ) γ2 + F. (12a) (12b) Because of strain hardening of the material and restraint of the surrounding components, after reaching the yielding point, the shear stress strain curve extends continuously at smaller stiffness until the strain is four times the yielding strain [14]. Hence, the shear stiffness of the region with holes in the two column webs after yielding (K whp ) is: ( ) β 1 K whp = K wh = 3 ( β 1 3 ) r A K w (13) where β is the strain hardening factor which is the ratio of the ultimate tensile stress to the yielding one. Similarly, both the regions with and without holes still contribute in series to the shear stiffness of the column webs. Thus, after the region with holes yields, the shear stiffness of the two column webs ( K w2 ) is: K w2 = = where r w2 = [( 1 nd h d b [ 1 nd h d b + ) 1 K w + 3nd h ( ) ndh 1 d b K whp ] 1 ] 1 K w = r w2k w (14) (β 1)d b r A [ ] 1 1 nd h d b + 3nd h (β 1)d b r A is the second reduction factor for the shear stiffness of the column webs due to the existence of holes. After the stress of the region with holes in the column webs reaches the yielding point, the shear stiffness contributed by the rectangular steel tube (K s3 ) to the panel zone is the superposition of the shear stiffness of the column webs and that of the generalized column flanges: K s3 = K f + K w2 = K f + r w2 K w. (15) When the external loads increase continuously, the stress of the region without holes in the column web of the panel zone subsequently reaches the yielding point and the strength (V wy ) is: V wy = 2A w τ sy = 2(d c 2t f )t w F y 3. (16) The corresponding shear strain (γ 3 ) and shear force (V s3 ) of the steel tube in the panel zone, respectively, are: γ 3 = γ 2 + V wy V why K w2 = γ 2 + V wy V why r w2 K w V s3 = V s2 + K s3 (γ 3 γ 2 ) = K f γ 3 + V wy + F. (17a) (17b) After the region without holes in the column webs reaches the yielding point, the stiffness of the column webs vanishes. Hence, the shear stiffness contributed by the rectangular steel tube (K s4 ) to the panel zone is equal to the shear stiffness of the generalized column flange ( K f ): K s4 = K f. (18) When the external loads increase continuously, the stress of the generalized column flanges subsequently reaches the yielding point and the strength (V f y ) is: ( ) 16 b c t 2 f 4M f y 4F y V f y = (d b t bf ) = (d b t bf ) = 2b c t 2 f F y 3(d b t bf ) (19) where M f y is the yielding flexural strength of the generalized column flanges. The corresponding shear strain (γ 4 ) and shear force (V s4 ) of the rectangular steel tube in the panel zone, respectively, are: γ 4 = γ f y = V f y K f = F y(d b t bf ) 3E s t f (20) V s4 = V s3 + K s4 (γ 4 γ 3 ) = V f y + V wy + F. (21) 2.2. Concrete The panel zone of the connection is bolted bidirectionally so that the concrete inside is not only loaded axially in the column but also confined by pre-stress both in the E W and N S directions. The Mohr Coulomb failure criterion is adopted to evaluate the ultimate shear stress. Considering the conditions of the concrete being loaded by uniaxial tension and compression, the relationship between σ 1 and σ 3 is: σ 1 f t σ 3 f c = 1 (22) where f c is the compressive strength of the concrete from a simple compression test, f t is the tensile strength from a simple tensile test. Define m to be the ratio of f c to f t. Eq. (22) can be rewritten as: m σ 1 σ 3 = f c. (23) According to the research by Richart [15], the ratio m is assigned as in 4.1. In this research, the axial compressive stress (σ x ), the lateral pre-stresses in the E W direction (σ y ), the lateral pre-stresses in the N S direction (σ z ), and the shear stresses τ xy, τ yz and τ zx, respectively, acted on the concrete in the panel zone are: P E c σ x = E s A s + E c A c T σ y = (b c 2t w )(d b t bf ) σ z = τ xy = τ cu τ yz = 0 τ zx = 0 T (b c 2t f )(d b t bf ) (24a) (24b) (24c) (24d) (24e) (24f)

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 399 where P is the axial compression acted on the concrete-filled steel tube; E c is the Young s modulus of concrete; A s and A c are the cross-sectional area of the steel tube and the concrete, respectively. From Eqs. (24a) to (24f), the principal stresses can be obtained as: [ (σ x + σ y ) + ] (σ x σ y ) + 4τxy 2 σ 1 = 1 2 (25a) σ 2 = σ z (25b) σ 3 = 1 [ ] (σ x + σ y ) (σ x σ y ) + 4τxy 2. (25c) After Eqs. (25a) and (25c) are substituted into the Mohr Coulomb failure criterion Eq. (23), the ultimate shear stress of the concrete can be determined as: τ cu = τ xy = 1 1 + m ( f c + σ x m σ y )( f c + σ y m σ x ). (26) When a rectangular section is subjected to shear force, the average shear stress is two thirds of the maximum shear stress. Therefore, the yielding shear strength (V cy ) and the ultimate shear strength (V cu ) of the concrete in the panel zone, respectively, are: V cy = 2 3 τ xy A c V cu = τ xy A c. (27a) (27b) There are in total 2n rows of holes for the bolts in the concrete of the panel zone in the E W and N S directions. The regions with and without holes contribute in series to the shear stiffness of the concrete so that the reduction factor for the shear stiffness of concrete (r c ) is: ( r c = 1 2nd h + 2nd ) 1 h. (28) d b d b r A Similarly, the ultimate shear stress ( V cu ) and yielding shear stress ( V cy ) of the concrete in the panel zone are, respectively, modified as: V cu = r c τ xy A c V cy = 2 3 r cτ xy A c = 2 3 V cu. (29a) (29b) It is assumed that the shear stress of concrete in the panel zone is ultimate ( V cu ) after the shear strain reaches γ 2. The elastic shear stiffness K c1, shear stiffness after yielding K c2, the yielding shear strain γ cy and the ultimate shear strain γ cu, respectively, are: K c1 = r c G c A c V cu K c2 = V cu V cy = 1 γ 2 γ cy 3 γ 2 γ cy γ cy = 2 τ xy 3 γ cu = γ 2 G c (30a) (30b) (30c) (30d) Fig. 4. Deformation of building structure subjected to seismic action. where G c = E c /2.3 is the shear modulus of concrete. 2.3. Steel tube and concrete The steel tube at the panel zone is divided mechanically into webs and flanges. The strength and strain at which the region with holes in the steel webs yields is defined as the yielding shear strength and yielding shear strain of the panel zone. The strength and strain at which the steel flanges yield is defined as the ultimate shear strength and ultimate shear strain of the panel zone. For the rectangular concrete-filled steel tube, the yielding shear strength (V y,pro ), shear stiffness (K pro ) and ultimate shear strength (V u,pro ) can be the superposition of those of the steel tube and of the concrete, respectively, as: V y,pro = ( K f + K w1 )γ 2 + F + V cu K pro = K f + K w1 + (F + V cu )/γ 2 V u,pro = V f y + V wy + F + V cu. (31a) (31b) (31c) Taking one of the specimens as an example, the computation of stiffness, yielding shear strength and ultimate shear strength for the panel zone is illustrated in the Appendix. 3. Experimental design 3.1. Design of testing frame In this study, a series of experiments has been designed to investigate the behavior of bolted beam-to-column connections for CFTs subjected to seismic cyclic loading. The experiment simulates a high-rise structure with a model that spans over 6 m with a height of 3.2 m. The specimen was 400 400 mm square CFT columns connected to H-shape steel beams (H500 200 10 16 mm) to simulate a cross-shape of beam-to-column connections. Under seismic actions, story drift occurs in the structure, and the inflection points of beam and column are assumed to occur at the mid-points. Taking a sub-structure as shown in Fig. 4, the panel zone of the connection is the center, and the midpoints of the two columns and the two beams become the boundaries. In this cross-shaped sub-structure, four hinges are used to simulate the upper, lower, left and right inflection points to perform a full-sized simulation study. The schematic diagram and dimensions of the testing frame are given in Fig. 5 and a photograph is shown Fig. 6. To simulate the upper and the lower inflection points of the columns, a segment of an H-beam is used. The H-beam segment

400 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 Fig. 5. Schematic diagram of testing frame (Unit: mm). Table 1 Dimensions of specimens Specimen Column section (A572 Grade50) (mm) B/t Beam section (A36) (mm) FSBE6 400 400 6 6 66 H500 200 10 16 FSBE8 400 400 8 8 50 H500 200 10 16 FSBE10 400 400 10 10 40 H500 200 10 16 Table 2 Material properties of CFT column for each specimen Specimen Steel Concrete σ y (MPa) σ u (MPa) ε y (%) ε u (%) f c (MPa) Fig. 6. Photograph of testing frame. is bolted to the mid-point of the column, and the behavior of the minor axis of each H-beam segment to the bending moment is used to simulate the behavior of the inflection point. To apply the static loading to the structure, a horizontal crossbeam oriented in the N S direction is attached to the top of the column. Tie-rods with a diameter of 39 mm are used to connect the north and the south ends of the crossbeam to the strong floor. Pre-stressed forces are applied at the end of tie-rods to the strong floor with a jack such that the beam is subjected to a downward tension. The tension force is equivalent to the static load applied to the top of the cross-shaped structure. In this study, three specimens are penetrated with bolts to connect the column and beam in the E W direction, and column and beam end-plates in the N S direction. It is hoped to be more realistic in simulating behaviors of 3D buildings. The names of the specimens were FSBE. Letter F means filled with concrete, letter S means square section, letter B means bolted connection, E means end-plates in four faces and is FSBE6 424.4 535.6 0.20 22.73 27.95 FSBE8 458.6 631.8 0.21 20.00 28.16 FSBE10 436.5 577.3 0.24 20.00 33.41 the column thickness. The three specimens are designed to have column thickness of 6 mm, 8 mm and 10 mm, and are named FSBE6, FSBE8, and FSBE10, respectively. 3.2. Design of specimens The dimensions of each specimen are listed in Table 1. The rectangular steel tube is A572 (Grade 50) steel. To manufacture the rectangular column, two steel plates are cold bent into a U shape with the corner radius twice that of the thickness. Subsequently, two U-shaped steel plates are combined into a box steel tube by fully penetration welding with backing bars. The material properties of the CFT column for each specimen are listed in Table 2. The end-plates are connected with bolts on the rectangular concrete filled steel tubes. There are 4 bolts of 24 mm in diameter in the upper and lower sides of the upper and lower flanges, respectively, in total 16 of which are in the E W direction, made of A490 steel, penetrating the end-plates and

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 401 Fig. 7. Design of bolted beam-to-column connection (Unit: mm). Table 3 Material properties of the beam for each specimen Steel σ y (MPa) σ u (MPa) ε y (%) ε u (%) Web 360.1 540.8 0.23 20.00 Flange 256.0 417.5 0.13 26.16 are welded to reduce the prying actions of the end-plates and enhance the stiffness of the connection. The design of the beam is shown in Fig. 8; the material properties of the beam for each specimen are the same and are listed in Table 3. 3.3. The loading system Fig. 8. Design of beam end section (Unit: mm). the rectangular concrete filled steel tube as shown in Fig. 7. As the strength of filled concrete in the steel tube is developed fully, a pre-stress of 49 kn is applied to the bolt. The beams of the specimens are made of H-shaped A36 steel with cross section of H500 200 10 16 mm. The end of the H steel beam is fully penetration welded to end-plates of 25 mm in thickness and 400 mm in width. In order to avoid stress concentration in the beam end welding zone, the beam end is reinforced with flange wing plates of 16 mm in thickness, 50 mm in width and 200 mm in length; the length is enough to move the plastic hinges away from the welding zone. Upstanding ribs of 100 mm in width, 100 mm in height and 9 mm in thickness To apply the static loading to the structure, a horizontal crossbeam oriented in the N S direction is attached to the top of the column. 39 mm diameter tie-rods are used to connect the north and the south ends of the crossbeam to the strong floor. Pre-stressed forces of 1472 kn are applied at the ends of tie-rods to the strong floor with a jack such that the column is subjected to compression. Considering the difficulty of applying the vertical load and horizontal cyclic displacement to the column, the horizontal cyclic displacement is replaced by vertical cyclic displacements of the corresponding beam mid-points. The top of the substructure is controlled to be zero displacement by a horizontal actuator. The vertical actuators at the left and right sides of the structure are also controlled but in opposite directions in order to simulate the story drift caused by horizontal loading. Triangular waves are used for displacement control of the two vertical actuators in opposite directions, with the rate of story drift being fixed at 1.875 mm/s (the rate of story drift ratio 0.0625%/s). The magnitudes of story drift are 7.5 mm (0.25%), 11.25 mm (0.375%), 15 mm (0.5%), 22.5 mm (0.75%), 30 mm

402 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 Fig. 9. Loading process for displacement control (% radian denotes 10 2 rad). (1.0%), 45 mm (1.5%), 60 mm (2.0%), 90 mm (3.0%), etc., as shown in Fig. 9. When the story drift ratio is below or equal to 1.0%, the same magnitude of displacement is applied for three cycles. When the story drift ratio increases above 1.5%, only two cycles are applied for each magnitude. When the displacement reaches 60 mm (2.0%), the story drift ratio increases in order by 1.0% till the strength of the specimens has reduced to 80% of the maximum strength. 4. Experimental results and discussion There were three sets of specimens with different thickness. The experimental results are discussed as (1) experiment courses and failure modes, (2) strength and ductility, (3) angular displacement and energy dissipation, and (4) stiffness and shear strength of panel zone, described in the following sections. 4.1. Experiment courses and failure modes For the specimen FSBE6, when the beam mid-points angular displacement reached 2%, the lower column face near the panel zone slightly bulged outward and the whitewash of the beam webs started to peel off. When the angular displacement reached 3%, the beam webs showed slight uneven deformation. As the angular displacement reached 4%, the Fig. 10. Failure condition of specimen FSBE6. column near the panel zone showed significant local buckling and its flanges showed slight buckling deformation. As the angular displacement reached 5%, the bending buckling of beam flanges became obvious and the whitewash of the beam webs peeled off severely. As the angular displacement reached 6%, the beam flanges showed torsional buckling and the webs showed local buckling. The experiment stopped at the angular displacement of 7%. The failure condition is shown in Fig. 10 and the relationship between its loading and angular displacement is shown in Fig. 11. For the specimen FSBE8, when the angular displacement reached 3%, the beam flanges showed slight local buckling. As the angular displacement reached 4%, the local buckling became obvious and the column web started to buckle locally. As the angular displacement reached 5%, the beam flanges showed torsional buckling. The experiment stopped at the angular displacement of 7%. The failure condition is shown in Fig. 12 and the relationship between its loading and angular displacement is shown in Fig. 13. For the specimen FSBE10, when the angular displacement reached 4%, the beam flanges showed local buckling, while the loading applied in the beam-end and panel zone was maximum. As the angular displacement reached 5%, the loading began Fig. 11. Relationship between loading and angular displacement for east and west beams of specimen FSBE6 (% rad denotes 10 2 rad).

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 403 Fig. 12. Failure condition of specimen FSBE8. Fig. 14. Failure condition of specimen FSBE10. Table 4 Experimental results of flexural strength of beam for each specimen Specimen West East +M max (kn m) M max (kn m) +M max (kn m) M max (kn m) FSBE6 795 777 801 783 FSBE8 806 806 842 818 FSBE10 824 818 854 830 to lower. The experiment stopped at the angular displacement of 7%. The failure condition is shown in Fig. 14 and the relationship between its loading and angular displacement is shown in Fig. 15. 4.2. Discussions on strength and ductility The hysteretic loops of the loading and angular displacements in the E W beam ends are shown in Figs. 11, 13 and 15. The experimental values of flexural strength are listed in Table 4. It demonstrates that the lower the width-to-thickness ratio is, the higher the ultimate flexural strength becomes. Table 5 lists the yielding angular displacement, ultimate angular displacement and ductility ratio of each specimen. The Table 5 Experimental results of ductility ratio of connection for each specimen Specimen West East θ y (% rad) θ u (% rad) θ u /θ y θ y (% rad) θ u (% rad) θ u /θ y FSBE6 1.06 5.61 5.29 1.01 5.43 5.38 FSBE8 1.02 5.10 5.00 0.97 4.95 5.10 FSBE10 0.96 5.00 5.21 0.82 4.84 5.90 θ y : yielding angular displacement; θ u : ultimate angular displacement; θ u /θ y : ductility ratio; % rad denote 10 2 rad. ratio of ultimate shear strain to the yielding one is defined as ductility ratio of the panel zone. It is observed that the thicker the steel tube is, the lower the yielding and ultimate angular displacements become. Besides, the ductility ratios of all specimens are over 5, which shows that all specimens had good ductility. 4.3. Angular displacements and energy dissipations of beam, column and panel zone Table 6 lists the angular displacements of the beam, column and panel zone for each specimen when the maximum story Fig. 13. Relationship between loading and angular displacement for east and west beams of specimen FSBE8 (% rad denotes 10 2 rad).

404 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 Fig. 15. Relationship between loading and angular displacement for east and west beams of specimen FSBE10 (% rad denotes 10 2 rad). Table 6 Angular displacement of the beam, column and panel zone at maximum story drift Specimen θ Beam /θ Total (%) θ Column /θ Total (%) θ PZ /θ Total (%) FSBE6 72.40 24.85 2.75 FSBE8 82.26 15.62 2.12 FSBE10 82.71 15.48 1.81 Table 8 Theoretical values of ultimate shear strength and experimental values of maximum shear strength for the panel zone Specimen V u,pro (kn) V max,exp (kn) FSBE6 3090 3249 4.89% FSBE8 3268 3290 0.69% FSBE10 3468 3524 1.59% (V u,pro V max,exp ) V max,exp 100% Table 7 Theoretical and experimental values of elastic stiffness for the panel zone Specimen K pro ( 10 6 kn/rad) K exp ( 10 6 kn/rad) FSBE6 2.884 2.962 2.63% FSBE8 2.947 3.228 8.71% FSBE10 3.115 3.368 7.51% (K pro K exp ) K exp 100% drift was reached. From the table, the percentages of total displacement contributed by the beam, column and panel zone are 70% 80%, 15% 25% and 1% 3%, respectively. It shows that the beam contributed the most displacement to the structural system. The results also show that the thicker the tube is, the more the displacement contributed by beam becomes. Dissipated energy is defined as the area of the hysteretic loop formed by the relationship of force and displacement. Fig. 16 shows the energy dissipation proportions of the beam, column and panel zone. The figure shows that the beam contributed the most energy dissipation, which has the same tendency as in Table 6. The thicker the tube is, the higher the percentage of energy dissipation contributed by beam becomes. 4.4. Relationships between forces and deformations of the panel zone The experimental stiffness is achieved from regression analysis within the elastic range. Table 7 lists the theoretical and experimental values of elastic stiffness of the panel zone. It can be seen that most of the predicted values are more conservative; therefore, the stiffness of the beam-to-column connection would not be over-estimated. In the experiment, the ultimate shear strength might not be reached. The maximum strength developed in the panel zone is defined as the maximum shear strength of the panel zone. Table 8 lists the theoretical values of ultimate shear strength and experimental values of maximum shear strength for the panel zone. It can be seen that the theoretical values of ultimate shear strengths are somewhat higher than the experimental values of the maximum shear strength; thus the recommended values are conservative and worth referring to. Besides, it is investigated that the higher the width-to-thickness ratio with a thinner steel tube wall is, the more likely the local buckling to column webs in the panel zone will be. Furthermore, it causes the stiffness and strength to decrease. The thicker the tube walls of the specimens are, the later the local buckling occurs. 5. Conclusions The focus of this study is the seismic behavior of bidirectional bolted connections for CFTs. For the bolted beamto-column connection, the main tube is welded in a factory and the beam-to-column connection is field-bolted, which would increase the convenience of fabrication and decrease the construction uncertainty. Besides, the confinement effects of pre-stress induced by bolts can increase the stiffness, strength and ductility of panel zones. The installation of end-plates in the H-beam is another characteristic in this design. The flange wing plates and upstanding ribs are welded to the end of the steel beam. The use of a flange wing plate can move the plastic hinges away from the welding zone, and the reinforced upstanding rib can reduce the prying action.

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 405 Fig. 16. Proportion and distribution of energy dissipation for bolted beam-to-column connections (% rad denotes 10 2 rad). A mechanical model for the panel zone is established to calculate the yielding shear strength and ultimate shear strength. For a series of experiments that was performed, the results obtained agreed with the theoretical ones. Thus, the model is a valuable reference for future research and application. There are three sets of specimens in the experiment. The experiment results indicate that the stiffness, strength, ductility and energy dissipation mechanism of the bolted connection have superior earthquake resistance. Even though the story drift ratio exceeds 7%, the structure still stands. In all specimens, the plastic angular displacements reached 6%, which met the seismic specifications used in Taiwan and the US. The pre-stressed bidirectional bolts would enable the panel zone to substantially enhance its strength, so the energy is mainly dissipated by plastic hinges induced in the beams. The thicker the tube is, the higher the percentage of energy dissipation contributed by beam becomes. For the bolted beam-to-column connections proposed in this study, the welding work is completed in a factory and only bolting work is needed in the field; thus the construction quality can be controlled easily. From the theoretical and experimental results, bolted beam-to-column connections have an excellent earthquake resistance with respect to stiffness, strength, ductility, and energy dissipation. Besides, the failure mode can be predicted from the design of the connection. Therefore, the bolted beam-to-column connection has excellent seismic resistance, and the structural system can perform well as expected and be put into practice. Acknowledgements The authors deeply appreciate the financial support offered by the National Science Council, ROC, and the National Center for Research on Earthquake Engineering, ROC. Appendix. Worked example for the mechanical properties of a panel zone Taking specimen FSBE6 as an example, the computation process of the yielding shear strength (V y,pro ), shear stiffness (K pro ) and ultimate shear strength (V u,pro ) of bidirectional bolted connections for CFT columns and H-beams is illustrated. Dimensions: The width of the column: b c = 0.4 m The depth of the column: d c = 0.4 m The thickness of the column flange: t f = 0.006 m The thickness of the column web: t w = 0.006 m The depth of the H-beam: d b = 0.5 m The thickness of the H-beam flange: t bf = 0.016 m The thickness of the H-beam web: t bw = 0.016 m The thickness of the beam end-plate: t ep = 0.025 m The diameter of the holes: d h = 0.03 m The number of rows of holes in the column web of the panel zone: n = 2 The number of holes in a row of bolts: m = 4

406 L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 Material properties: The Young s modulus of the steel: E s = 2.001 10 11 N/m 2 The yielding tensile stress of the steel: F y = 4.244 10 8 N/m 2 The Poisson s ratio of the steel: v s = 0.3 The friction coefficient between end-plate and steel tube: µ = 0.35 The strain hardening factor of the steel: β = 1.3 The Young s modulus of the concrete: E c = 2.484 10 10 N/m 2 The compressive strength of the concrete: f c = 2.795 10 7 N/m 2 The Poisson s ratio of the concrete: v c = 0.15 Loading: The pre-stress of each bolt: t = 4.9 10 4 N The axial compression acted on the CFT column: P = 1.472 10 6 N 1. Calculate the shear stiffness of the two generalized column flanges ( K f ) from Eq. (1): K f = 2E sb c t 3 f (d b t bf ) 2 = 2.036 107 N/rad where t f = t f + t ep = 0.031 m. 2. Calculate the shear stiffness of the region without holes (K w ) in the two column webs from Eq. (2): K w = 2(d c 2t f )t w G s = 3.583 10 8 N/rad where G s = E s 2(1+v s ) = 7.696 1010 N/m 2. 3. Calculate the shear stiffness of the region with holes (K wh ) in the two column webs from Eq. (3): K wh = 2(d c 2t f md h )t w G s = 2.475 10 8 N/rad. 4. Calculate the shear stiffness of the two column webs ( K w1 ) from Eq. (4): [( K w1 = 1 nd ) ( ) ] h 1 ndh 1 1 + d b K w d b K wh = 3.401 10 8 N/rad. 5. Calculate the pre-stress of all bolts (T ) and the friction force between end-plate and steel tube (F) from Eq. (8): T = 2mnt = 7.840 10 5 N F = 2T µ = 5.488 10 5 N. 6. Calculate the strength of the two column webs (V why ) when the region with holes in the column web yields from Eq. (11): V why = 2(d c 2t f md h )t w F y 3 = 7.880 10 5 N. 7. Calculate the corresponding shear strain of the steel tube (V why ) in the panel zone from Eq. (12a): γ 2 = V why = 0.002317 rad. K w1 8. Calculate the strength (V wy ) when the stress of the region without holes in the column web of the panel zone reaches the yielding point from Eq. (16): V wy = 2(d c 2t f )t w F y 3 = 1.141 10 6 N. 9. Calculate the strength (V f y ) when the stress of the generalized column flanges reaches the yielding point from Eq. (19): V f y = 2b c t 2 f F y 3(d b t bf ) = 2.247 105 N. 10. Calculate the axial compressive stress (σ x ) and the lateral pre-stresses in the E W direction (σ y ) acted on the concrete in the panel zone from Eqs. (24a) and (24b): P E c σ x = = 6.493 10 6 N/m 2. E s A s + E c A c T σ y = (b c 2t w )(d b t bf ) = 4.175 106 N/m 2 where A c = (b c 2t f )(d c 2t w ) = 0.15054 m 2, A s = b c d c A c = 0.00946 m 2. 11. Calculate the ultimate shear stress of the concrete (τ cu ) in the panel zone from Eq. (26): τ cu = τ xy = 1 1 + m ( f c + σ x m σ y )( f c + σ y m σ x ) = 8.645 10 6 N/m 2. 12. Calculate the reduction factor for the shear stiffness of concrete (r c ) from Eq. (28): ( r c = 1 2nd h + 2nd ) 1 h = 0.9030 d b d b r A where r A = 1 md h /(d c 2t f ) = 0.6907. 13. Calculate the ultimate shear stress ( V cu ) of the concrete in the panel zone from Eq. (29a): V cu = r c τ xy A c = 1.175 10 6 N. 14. For the rectangular concrete-filled steel tube, the yielding shear strength (V y,pro ), shear stiffness (K pro ) and ultimate shear strength (V u,pro ) can be the superposition of those of the steel tube and of the concrete from Eqs. (31a) to (31c): V y,pro = ( K f + K w1 )γ 2 + F + V cu = 2.559 10 6 N K pro = K f + K w1 + (F + V cu )/γ 2 = 2.884 10 9 N/rad V u,pro = V f y + V wy + F + V cu = 3.090 10 6 N. References [1] Fujimoto T, Nishiyama I, Mukai A. Test results of CFT beam-tocolumn connection. US Japan cooperative earthquake research program: Composite and hybrid structures. In: 4th JTCC.1997. [2] Youssef NFG, Bonowitz D, Gross JL. A survey of steel moment-resisting frame buildings affected by the 1994 Northridge Earthquake. Report No. NISTIR 5625. Gaithersburg (MD): National Institute of Standards and Technology; 1995. [3] Popov EP, Yang TS, Chang SP. Design of MRF connections before and after 1994 Northridge Earthquake. Engineering Structure 1998;20: 1030 8. [4] Azizinamini A, Shekar Y. Design of through beam connection detail for circular composite columns. Engineering Structures 1995;17:209 31.

L.-Y. Wu et al. / Engineering Structures 29 (2007) 395 407 407 [5] Fujimoto T, Inai E, Tokinoya H, Kai M, Mori K, Mori O et al. Behavior of beam-to-column connection of CFT Column system under seismic force. In: Proceeding of the 6th international conference on steel concrete composite structure. [6] Kang CH, Shin KJ, Oh YS, Moon TS. Hysteresis behavior of CFT column to H-beam connections with external T-stiffeners and penetrated elements. Engineering Structure 2001;23(9):1194 201. [7] Elremaily A, Azizinamini A. Design provision for connections between steel beam and concrete filled tube column. Journal of Constructional Steel Research 2001;57:971 95. [8] Shin KJ, Kim YJ, Oh YS, Moon TS. Behavior of welded CFT column to H-beam connections with external stiffeners. Engineering Structure 2004; 26(11):1877 87. [9] Alostaz YM, Schneider SP. Analytical behavior of connections to concrete-filled steel tubes. Journal of Constructional Steel Research 1996; 40:95 127. [10] Alostaz YM, Schneider SP. Connections to concrete-filled steel tubes. In: Proceedings of the 11th world conference on earthquake engineering, Paper No.748. 1996. [11] Schneider SP. Summary of connections to concrete-filled steel tube columns. In: 4th joint technical coordinating committee meeting, US Japan cooperative earthquake research program on composite and hybrid structures. 1997. [12] Wu LY, Chung LL, Tsai SF, Shen TR, Huang GL. Seismic behavior of bolted beam-to-column connection for concrete filled steel tube. Journal of Constructional Steel Research 2005;61:1387 410. [13] Furlong RW. Design of steel-encased concrete beam-columns. Journal of Structural Division, ASCE 1968;94(ST1):267 81. [14] Krawinkler H. Shear in beam-column joints in seismic design of steel frames. Engineering Journal, AISC 1978;15(2):82 91. [15] Richart FE, et al. A Study of the failure of concrete under combined compressive stresses. Univ of Illinois Eng Expert Stat Bulletin 1928;185.

本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 (www.xuebalib.com) 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具