Composite Connections in Steel and Concrete. I. Experimental Behaviour of Composite Beam-Column Connections

Transcription

1 ELSEVIER d. Construct. Steel Research 31 (1994~ Elsevier Science Limited Printed in Malta. All rights reserved X/94/$7.00 Composite Connections in Steel and Concrete. I. Experimental Behaviour of Composite Beam-Column Connections Y. Xiao, B. S. Choo & D. A. Nethercot* Department of Civil Engineering, University of Nottingham, Nottingham NG7 2RD, UK (Received 20 April 1993; accepted 19 July 1993) ABSTRACT A series of tests designed to investigate the interaction of a variety of different steel beam to column details with a composite metal deck floor is described. The main emphasis is on assessing the connections' moment capacity, rotational sh[lhess and rotation capacity. A wide range of variables was arranged to investi.qate these properties under the inltuenve of the composite action. Full details of the experimental hehaviour of all twenty specimens are reported. 1 INTRODUCTION Efficient and economical lightweight floor systems may be created by integrating the structural properties of concrete and cold-formed steel decking by ensuring composite action through the use of through-deck welded shear studs. This type of construction has now became a common feature in multi-storey steel frame buildings in several countries.1 3 Traditional methods of steel frame design disregard the actual behaviour of the beam to column joints, assuming one of the ideal models of perfectly rigid or perfectly pinned. In addition to the inherent stiffness and moment capacity of the steel detail, neglect of the contribution to joint properties from composite action with the floor slab may result in either an over-conservative or an unconservative assessment of the actual behaviour of the composite frame. Inclusion of more realistic joint effects does, however, require that a proper understanding of the structural behaviour of composite joints be available. Previous studies in this areas have been extensively reviewed. <5 However, a general basis for the design of *To w'hom correspondence should be addressed.

2 4 Y. Xiao, B. S. ('hoo, D.,4. Nethercot composite joints has not yet been established and the large number of variables present means that the existing database of test results provides very patchy coverage of the subject. Thus current codes do not readily cope with the design aspects of composite connections. 6'7 Recent tests conducted at Nottingham University are reported herein. The main emphasis is on assessing the three key indications of performance: moment resistance, rotational stiffness and rotation capacity. Knowledge of these fundamental properties is necessary before the true semi-rigid and partial strength nature of the connections can be properly incorporated into more realistic methods of composite connection and frame design. Specimens were therefore selected to include several currently used joint types as well as to investigate slab parameters expected to influence the connection behaviour directly. The study built on the pilot study of composite connections undertaken in Sheffield. s The full details of the behaviour of composite connections under monotonic loading will be described herein. Design implications and strength predication methods will be reported separately based on an appraisal of these results. 2.1 Specimen description 2 TEST SPECIMENS AND SET-UP The conventional metal deck flooring system comprises concrete topping on profiled metal decking on top of the steel beams, with shear connection provided by through-deck welded shear studs. The test specimens were configured as 'cruciform' type and 'cantilever' type, as shown in Figs 1 and 2. Four different types of steel joint have been used: seating cleat with double web cleats, flush endplates, partial depth endplates and tinplate, as shown in Fig. 3. The members used in the construction of the specimens Concrete slab Shear studs ~ Plate stiffening Reinforcement Bj~ate" / } " ~ lacking plate / " ~ Supportingcolumn Fig. I. Cruciform arrangement of composite connection.

3 Composite connections in steel and concrete 5 Transverse rebars J Major axis / - beam stub t.. Supporting column 203x203 UC52-5 / Additional trim bars Shear studs Rebar and mesh ~-~ T/-~ T/-~ /--~T/-~ T/-~ t u Partial depth endplate in the bottom position 305x165 IB40 Fig. 2. Cantilever arrangement of composite connection. / % i ui i i % %, m,% Type 1 : Seating cleat joint (a) Type 2 : Flush endplate joim (b) % -2, Type 3 : Partial depth endplate joint (c) q/ Fig. 3. Four types of steel joint Type 4 : Finplate joint (d) were 305 x 165 x 40, 457 x 191 x 98 UBs and 203 x 203 x 52 UC in grade 43 steel. Reinforcement was A142 mesh, supplemented in some cases by T10 or T12 rebars. PMF CF46 deformed metal decking was used as permanent shuttering with welded-through shear studs connected to the steel beam underneath. The bare steel joint was first assembled in position on the baseplate, the decking was added and then the slab was cast. All bolts (M20. Gr. 8.8j were tightened to 150N m torque by a torque wrench to maintain consistency and comparability between specimens.

4 6 E Xiao, B. S. Choo, D. 4. Nethercot The steel decking was filled with ready-mixed normal weight concrete. This had a design cube strength of 30 N/ram 2, maximum aggregate size of 20 mm and a slump of 45 mm. In practice the slump values were found to be in the range ram, and the concrete strength on the day of testing an average of 35 N/mm 2. Concreting work was carried out inside the laboratory with the specimen in the test position. The metal decking was used as bottom formwork with non-absorbent plywood as side shuttering. Some curing oil was brushed on the surface of the side shuttering for easy demoulding. Casting of cubes and cylinders for strength tests was carried out at the same time. After the surface finishing, the specimen was covered by a polythene sheet and left to cure. The cylinders and cubes were cured in water at an average temperature of 20 C. These samples were tested at 7 days, 14 days and on the test day. Usually, the connections were tested when the concrete strength reached 30 N/ram 2 or after 28 days. 2.2 Test set-up and instrumentation The column of each specimen was fixed to the laboratory floor through a baseplate and supporting angles for symmetrical loading. This arrangement allowed for accurate adjustment of the specimen. The load was applied to the specimen using hand-operated hydraulic jacks mounted on the test rig, which was itself attached to the strong floor of the laboratory as shown in Fig. 4. Two loads were applied symmetrically at 1-5 m from the column flange through two hollow section load spreaders. The top and ? 25 Fig. 4. Test rig and set-up: 1, test rig; 2, specimen: 3. hydraulic jack; 4, load cell: 5, load spreader: 6, LVDT; 7, inclinometer: 8, strong floor.

5 Composite connections in steel and concrete 7 bottom of the column were both fixed when unsymmetrical loading was applied to the specimens of cantilever type. Measurement of the load was obtained through load cells positioned between the jack and a spherical seat. The arrangement allowed in-plane movement so as to ensure that the load was always applied to the same point on the slab. Strains in the reinforcement and on the steel beam were measured using strain gauges. The resistance and insulation were checked after the strain gauge coating to ensure proper operation after concreting. After assembly of the test specimen, all the instruments were mounted on the specimen. Inclinometers were used to measure the rotation of the column and the relative rotation of the beam so as to provide full information from which the moment-rotation relationship could be derived. Displacement transducers were used to measure the deflection of the specimen on the bottom flange of the beam and also to monitor slip of the seat cleat. The relative rotation of the beam to column can also be obtained from the deflection values within the calibrated horizontal length of the beam. Dial gauges were used to measure the deformation of the concrete slab and horizontal movement of the specimens. Load was applied to the specimen by a manual pump. The load increment was one-twentieth of the calculated capacity of the specimen, usually 5-10 kn each step, with half increments as failure was approached. Load cells were also connected to the load output chart recorder to ensure that the load history and maxima were graphed simultaneously. Loading was continued for several steps after the peak load value so as to obtain information on the ductility of the specimen and the post-failure situation. All the test data were monitored through a computer-controlled datalogging system using Axis software to display selected test information on line. After the test, all the specimens were carefully inspected, including taking the joint apart, removal of the bolts, checking the slip and the hole deformation. When significant large longitudinal cracking appeared in the slab, it was cut and the shear studs were checked. 3 TEST RESULTS AND ANALYSIS Test work was divided into three phases designed to complement one another. These separate phases of work have been partly reported in Refs 9, 10 and 11 respectively. All results for the full set of twenty specimens will be analysed and compared herein. The description of each specimen and the main test results are summarized in Table 1.

6 Ultimate rotation (mrad) ~ > 59'8 14"3 16" "5 26" Maximum rotation (mrad) b "6 41' ' "9 Failure mode b A B B B C,D G,E C,D G,D F F TABLE I Specimen Details and Test Results Specimen Joint type Web stiffeninq Reinforcement First crack and specimen ratio (%) moment shape (kn m) Ultimate moment (kn m) S J1 SCJ1 SCJ2 SCJ3 SCJ4 SCJ5 SCJ6 SCJ7 SCJ8 SCJ9 Seating cleat Seating cleat Finplate Flush endplate Flush enclplalc Flush endplate Flush endplate Flush endplate Partial depth endplate Partial depth endplate Cruciform None None Cruciform A142 mesh None (0.2%) Cruciform A 142 mesh l None (0.2%) Cruciform A142 mesh 30 85'7 None (0.2%) Cruciform TI2 rebar None (10%) Cruciform T12 rebar Web stiffeners (10%) Cruciform TI0 rebar and A "6 None mesh (1.0%1 Cruciform TI2 rebar and A Plate stiffening mesh (1-2%) Cruciform T10 rebar and A None mesh (0.8%) Cruciform T10 rebar and A '5 None mesh (0.8%)

7 SCJ10 Partial depth endplate Cruciform T10 rebar and A C, D None mesh (0.8%) SCJll Seating cleat Cruciform T10 rebar and A142 33' " C,D None mesh (0'7%) SCJ12 Finplate Cruciform T10 rebar and A H None mesh (0.7%) SCJ13 Partial depth endplate Cruciform TI0 rebar and A A (minor axis) None mesh (0"8%) SCJ14 Partial depth endplate Cruciform T10 rebar and A ' F (minor axis) None mesh (1.0%) SCJ15 Flush endplate Cruciform T10 rebar and A A Backing plates mesh (1.0%) SCJ16 Finplate Cruciform T12 rebar and A H Plate stiffening mesh (1.2%) SCJ17 Finplate Cruciform T10 rebar and A H mesh (0"5%) SCJ18 Partial depth endplate Cantilever T10 rebar and A ! (minor axis) mesh (0.8%) SCJ19 Partial depth endplate Cantilever T10 rebar and A ' ! (minor axis) mesh (0"8%) a I mrad = degree. ha, excessive deflection of beams; B, fracture of the mesh reinforcement; C, excessive deformation of column flange; D, buckling of column web; E, buckling of beam flange; F, buckling of beam web; G, shear studs failure; H, web side plate twisting; I, anchorage failure of reinforcement. e~ e~

8 10 Y. Xiao, B. S. Choo, D. A. Nethercot 3.1 Material tests Concreting work was carried out inside the laboratory with the specimen in the test position. Samples were cast at the same time to be used for monitoring the concrete strength. The cylinders (150 x 300 ram) and cubes (100x 100xl00mm) were cured in a water tank. These samples were tested at 7 days, 14 days and on the test day and/or at 28 days according to the BSI test standard. 12 The compressive and tensile strengths for all the specimens are summarized in Table 2. Material for tensile test coupons was gas-cut from the flanges and webs of the beams and columr~s after test, from areas where the stresses had been very low. Tensile tests were conducted in a 200kN Zwick 1484 computer-controlled universal test machine, with the test speed set at 20mm/min according to the BSI standard ~3 for materials testing. Values of static yield strength and ultimate strength for steel members are listed in Table 3. Test lengths were cut from the 10 mm and 12 mm diameter rebar and TABLE 2 Concrete Cube Strength and Splitting Strength Specimen Cube strenqth Splittin,q stren~tth ( N /mmz ) (N,"ram 2) 7 days 14 days Test day Test day SCJ1 (21 days) 26" SCJ2 (19 days) ~_,~"~ 3.59 SCJ3 (18 days) 31' SCJ4 (16 days) 39' SCJ5 (11 days) '21 SCJ6 (21 days) '56 SCJ7 (20 days) 25 30" SCJ8 (21 days) _~ ~;.~" 2.31 SCJ9 (21 days) SCJIO (21 days) SCJ 11 (28 days] 20" I I SCJI2 (28 days) _ 247 SCJ 13 (26 days) ' SCJ14 (21 days) SCJ15 (18 days) SCJI6 (20 days) SCJ 17 (28 days) I 303 SCJI8 (28 days] SCJ19 (40 days)

9 Composite connections in steel and concrete 11 TABLE 3 Tensile Strength of Steel Members Steel member Location Yield strength Tensile stren qth (N/mm 2) (N/mm 2) 203 x 203 UC52 Web 317" Flange x 165 UB40 Web '8 Flange x 198 UB98 Web Flange the A142 mesh (6 mm 4'). The test pieces prepared from the mesh included the weld point. Tests were conducted in the Zwick universal test machine under conditions defined by the BSI standard. 14 The test results are listed in Table Seating cleat with double web cleats connections (specimens S J1, SCJ1, SCJ11) Specimen S J1 was a bare steel joint, as shown in Fig. 3(a), with a seating cleat and double web stability cleats. The test result was required for a direct comparison with composite joint SCJ1 which had the same steel detail. The moment-rotation curve shown in Fig. 5 indicates the early onset of slip, with increasing growth of deflection and rotation as the load was increased. Horizontal slip occurred at the connection as the top of the beam moved out and the bottom of the beam pushed towards the column, causing opening of the web angles. The failure mode was excessive deflection of the beam. The last recorded beam rotation value was greater than 58 mrad. The measurements show that there was no vertical slip and only small deformations of the seating cleat. This test confirmed that bare TABLE 4 Tensile Strength of Rebar and Mesh ReinJorcement Diameter Yield stren qth Tensile stren#th (N/mm 2) (N/mm 2) TI2 rebar TI0 rebar 10" A 142 mesh 5" A 142 mesh (with joint)

10 12 ~: Xiao, B. S. Choo, D. A. Nethercot 2OO ~ 140 Z ~- 8O ~ SJ 1,g F1 r.' sc, Rotation(mRad) 6(3 Fig. 5. Moment rotation relations of cleated connections. steel joints of this type do possess some degree of moment resistance and rotational stiffness. Specimen SCJ1 had the same steel detail as S J1 and was cast with a t20mm deep concrete slab on to the profiled decking. Only the mesh reinforcement (A142), which is used to control shrinkage cracks in current construction practice, was included in the concrete floor. SCJll had the same detail as SCJ1 but with an increased reinforcement ratio to 0"7%. As expected, both specimens exhibited increased stiffness in the initial loading stage compared with S J1 as shown in Fig. 5. The first crack for SCJI was found parallel to the column flange at 15 kn m. This later formed into the main cracks of the slab. The stiffness of the joint decreased after the concrete cracked. This part of the path to failure is similar to a 'yield plateau'. The specimen regained stiffness later as the load was approaching its maximum value. Debonding of the concrete to the metal decking took place owing to the horizontal cracking formed at the surface between concrete and decking, and mesh fracture occurred eventually. The load fell rapidly after this stage. The ultimate bending strength of the specimen was increased by about 40% as compared with bare steel joint SJ1. But the ultimate rotation of the joint was only 14-3 mrad because of the typical brittle failure mode, mesh fracture. A mixture of rebar and mesh was introduced in specimen SCJ11. The decrease in the initial stiffness caused by the onset of slip was not severe because of the reinforcement preventing the large cracks from forming in the concrete. The cracks were smeared over the whole length of the specimen with the width of the main cracks being much less than for SCJ 1.

11 Composite connections in steel and concrete 13 The specimen failed in a more ductile way after the maximum load was reached and still possessed quite high residual strength. The rotational capacity and moment capacity were both increased as compared with SCJ I. The ultimate moment of resistance was almost four times that of the bare steel joint and three times that of the connection with only mesh reinforcement. Prior calculation had identified the column web as the weakest element. Buckling of the column web in compression as shown in Fig. 6 appeared after the reinforcement had reached yield as indicated by the strain gauge readings. The presence of additional reinforcement changed the form of the eventual failure to one in which the steel members were the controlling factor, leading to a substantial increase in the moment resistance as shown in Fig. 5. Direct comparison of the behaviour of these three different types of cleated joint shows that the decrease in stiffness caused by cracking in the slab and the onset of slip are important for cleated connections. Failure was initiated by mesh fracture when this was the only reinforcement available. Tensile tests on samples of the mesh revealed a low elongation of 3 5%. With additional reinforcement, the stiffness and moment capacity were greatly increased owing to the direct strengthening of the tensile zones, bolt slip was minimized and increased rotation capacities achieved. Fig. 6. Column web buckling and flange excessive deformation of composite cleated connection.

12 14 Y.. Xiao, B. S. Choo, D. A. Nethercot 3.3 Composite flush endplate connection (specimens SCJ4, SCJ5, SCJ6 and SCJ7) All flush endplate connections were generally as shown in Fig. 3(b). Specimen SCJ3 used only A142 mesh in the concrete and was designed to investigate the effect of a more rigid steel detail on the strength and stiffness of the composite connection. Compared with cleated and tinplate specimens SCJ1 and SCJ2, the specimen had a relatively high initial stiffness. The final strength of the joint was 84 kn m, much stronger than SCJ1 and SCJ2. The failure mode was mesh fracture. A comparison of the behaviour of three different joints each with the same type of mesh reinforcement is shown in Fig. 7, from which it can be seen that the flush endplate connection has the highest moment capacity. Decrease in stiffness caused by cracking in the slab and the onset of bolt slip occurred comparatively early for both cleated and tinplate connections. The flush endplate connection was not affected by bolt slip owing to its bolt orientation. The ratios of reinforcement of SCJ4 and SCJ5 (with column web stiffeners) were designed as 1"0% rebar in the concrete. Flexural cracking in the slab started from the tips of the column's flange. Compared with specimen SCJ3, a more uniform pattern of cracking was obtained. The moment rotation responses of SCJ4 and SCJ5 are shown in Fig. 8. Unlike cleated joints, these joints were able to maintain the applied loading long after the onset of cracking. The loss of stiffness was not very large compared with specimen SCJ3. Failure of SCJ4 was associated with E ~ 5o SCJ 1 SCJ2 SCJ Rotation (mrad) 9 o Fig. 7. Moment-rotation relations of different steel details with A 142 mesh (0./,,).

13 Composite connections in steel and concrete % reinforcement and web stiffener c..; * - - ~ 160 Z ~ 120 ~ SCJ3 O,I SCJ4 80 ""--0"--- SCJ ~ ' I L At", I Rotation(mRad) Fig. 8. Effect of reinforcement ratio and column stiffening on performance of composite flush endplate connections. excessive deformation of the column flange and buckling of the web of the column. Strain gauge readings indicated yielding of the reinforcement. The maximum moment achieved was 203 kn m for SCJ4. Specimen SCJ5 had the same reinforcement ratio as SCJ4. However, at the level of the bottom flange of the beam, its column web was stiffened by two horizontal steel plates as indicated in Fig. 9. The cracking pattern was similar to that of specimen SCJ4. Both strength and rotational stiffness were clearly enhanced by the presence of the web stiffeners. Ultimate moment for the joint was 241 knm. Instead of excessive column flange deformation, failure of the joint was due to local buckling of the bottom flange of the beam. No buckling of the column web or stiffeners was detected. The connection's ability to resist compressive force in the lower part was increased by the web stiffening which resulted in the moment capacity increasing. This arrangement permits full utilization of the steel sections and is also in line with the design philosophy of restricting failure to beams rather than columns. SCJ6, with the same ratio of reinforcement as SCJ4, was designed to check the effect of a different shear/moment ratio on the moment-rotation behaviour. The load position was moved closer to the column at 0"8 m to produce the high shear. Compared with SCJ4, the general behaviour was quite similar but the resistance moment of SCJ6 was decreased to 157"6 kn m as indicated in Fig. 10. The eventual failure was by buckling of the column web and excessive flange deformation. The presence of the high shear force decreased the moment capacity by 22%. SCJ7, with the ratio of 1"2% reinforcement, was designed with two

14 16 Y Xiao, B. S. Choo, D. A. Nethercot Fig. 9. Web stiffener for column stiffening. 24( O ~13 ~ SCJ4 "t * SCJ6 ] (I Rotation(mRad) (1 Fig. i0. High shear force and plate stiffening on the flush endplate connection. framed-in beam stubs connected to the minor axis of the column rather than two horizontal web stiffeners. By doing so, it was expected that the column web would be stiffened by the beam stubs, acting as de facto stiffeners. The test was terminated owing to excessive column flange deformation. This arrangement without the usual web stiffeners only changed the failure mode from web buckling to excessive flange deforma-

15 Composite connections in steel and concrete 17 tion. No clear moment enhancement was found. Longitudinal cracking was visible on the surface of the slab. After the test, the concrete floor was cut into two pieces to expose the shear studs with the concrete removed; the longitudinal cracks were continuous through the shear connector area. Shear connectors were found that had been bent severely as indicated in Fig. 11. Although the reinforcement had been increased to 1.2% in SCJ7, the moment capacity was still determined by the shear connector strength. The tensile regions of bare steel joints are relatively weak compared with those of composite connections because of the presence of the reinforced concrete floor. If the shear resistance of the shear connector is sufficient, the main variable will be the reinforcement ratio in the slab because the influence of the concrete will be negligible after cracking. Reinforcement ratios from 0-2 to 1"2% were covered for all three phases of the work. Test results showed that the initial stiffness and moment resistance were increased as the reinforcement ratio increased. Even with mesh reinforcement only, moment resistance of the composite connection was improved compared with an equivalent bare steel joint. The rotation capacity was improved as the eventual failure changed from a brittle type to a ductile one. Use of an appropriate ratio of reinforcement in the connection clearly improves the joint properties of moment-resistance, initial stiffness and rotational capacity. Fig. 11. Shear studs excessive bending.

16 18 Y Xiao, B. S. Choo, D. A. Nethercot Three different eventual failure modes (in column, beam and reinforcement respectively) were detected in the flush endplate connections with different ratios of reinforcement and web stiffening present. The tests have shown that such composite connections (i.e. with thin endplates, 10 mm thick) possessed a good initial stiffness and suffered comparatively little loss of stiffness after cracking of the concrete floor and yielding of the reinforcement (because of the relatively high rigidity and rotational capacity of this steel detail). Although the column stiffening should be avoided whenever possible for ease of construction, its use is clearly advantageous in terms of beam-column connection strength enhancement and failure mode control. Framed-in beams in the orthogonal direction did not increase the moment resistance, but effectively prevented column web buckling. 3.4 Flush endplate connection with backing plates (specimen SCJ15) Usually, the tension zone of a bare steel joint design is very weak owing to bending of the column flange. Its tensile strength then controls the design for the moment capacity of the joint. This is particularly true for the endplate joint when the beam is bolted to the flange of the column. Bolting plates to the back of the column flange in the tensile region can increase the stiffness and moment capacity of the joint. 15 When a concrete floor is present on top of the steel beam, the controlling area of the composite connection is usually the compression zone of the column owing to the restraint provided by the reinforcement to the tensile column area. Modes of failure for such composite connections will normally involve yielding of the reinforcement and compressive buckling of the lower portion of beam and/or column flange and web. The column web and flange can be stiffened by using welded web stiffeners. Flush endplate joints with orthogonal framed-in beams were also found to function as a kind of web stiffener in test SCJ7, when failure was caused by excessive compressive deformation of the column flange. One way of providing a stiffener to the ~zolumn flange in the compressive zone is to use the concept of the backing plate, as indicated in Fig. i, to enhance the bending resistance so that the need for a welded web stiffener can be waived, Specimen SCJ15 was a flush endplate connection with backing plates on the column flanges. The ratio of reinforcement was decreased to 1'0% corn pared with SCJ7. The... its qndieate " ' ~..~..~ ~h;o.~,o...~-~ e.f impro~'ement for flush endplate joints is quite beneficial, Excessive deformation of the column flange was prevented as shown in Fig. 12. When used with plate stiffening of the column web, backing plates can successfully improve the connection behaviour without the need for fitted and welded stiffening,

17 Composite connections in steel and concrete 19 Fig. 12. Backing plate to prevent excessive column flange deformation. 3.5 Composite partial depth endplate connection (specimens SCJ8, SCJ9 and SCJIO) The partial depth endplate as shown in Fig. 3(c) is normally considered sufficiently flexible to be considered as a simple connection in its bare steel state. It is current practice for the endplate to be welded to the top portion of the beam so as to ensure 'simple' behaviour. 16 However, this was thought to be incompatible with composite connection design principles owing to the availability of the stronger tensile zone. Three different positions of the partial depth endplate connection with the same ratio of reinforcement (0-8%) were chosen to study this influence. Specimens SCJ8, SCJ9 and SCJ10 represented different locations of endplate welded to the top, middle and bottom portion of the beam. The moment-rotation curves of all three specimens are shown in Fig. 13. The

18 20 E Xiao, B. S. Choo, D. A. Nethercot 160 SCJI8 Bottom SCJI9 ~ SCJ 10 Z 120 E i, t, ~ i ~, I1 ", t, Rotation(mRad) Fig. 13. Influence of the partial depth endplate location. 50 test curves of the two specimens SCJ8 and SCJ9 are quite similar. The cracking pattern was limited to several large cracks. Moment capacity dropped down after the peak load and increased again as the web of the beam came into contact with the flange of the column. But the initial stiffness and moment capacity were increased for specimen SCJ9 compared with SCJ8. The eventual failures for SCJ8 and SCJ9 were both beam web buckling as indicated in Fig. 14. SCJ10 exhibited a different behaviour compared with the above two specimens. The stiffness and moment resistance further increased. The eventual failure changed from the beam to the column web buckling as indicated in Fig. 15. More fine cracks appeared in the concrete floor. The test result indicated that the position of the partial depth endplate is crucial for the stiffness and moment capacity of the composite connection. When the concrete floor is in place, the current practice of placing the partial depth endplate near the top of the beam will cause a smaller resistance moment and stiffness. A more appropriate arrangement would be to weld the partial depth endplate towards the bottom portion of the beam. 3.6 Minor axis connections of the partial depth endplate (specimens SCJ13 and SCJ14) Four specimens have been tested with beams provided with partial depth endplates being connected to the minor axis of the supporting column. Specimens SCJ13 and SCJ14 were tested as cruciforms. Both connections had the same reinforcement ratio of A142 mesh plus 6T10 and 2T12 rebars (ratio of 0-8%). The results showed once again (as for previous tests

19 Composite connections in steel and concrete Fig. 14. Beam web buckling of partial depth endplate in top position. Fig. 15. Column web buckling of partial depth endplate in the bottom position. 21

20 22 Y. Xiao, B. S. Choo, D. A. Nethercot SCJ8 and SCJ10) that suitable positioning of a partial depth endplate is crucial to the achievement of high stiffness and moment capacity of the connection. The most appropriate position is to locate the plate on the lower portion of the beam. Figure 16 shows that SCJ13 with the plate on the bottom of the beam has a larger initial stiffness and almost twice the resistance moment as that of SCJ14 with the partial depth plate at the top. SCJ13 exhibited no significant deformation in the joint area and failed owing to excessive deformation of the slab and debonding of the concrete. Compared with the equivalent major axis connection (SCJ10), the failure mode of the minor axis connection was transferred from column web buckling to the slab. The stiffness and moment capacity were actually increased as the opportunity for compression zone deformation was effectively eliminated. SCJ14 failed owing to buckling of the beam web (in the same way as SCJ8) as shown in Fig. 14. The strength and stiffness of connections employing symmetrical partial depth endplates or flush endplates to the minor axis of the supporting column are usually increased compared with equivalent major axis connections because the compressive force transfer effectively bypasses the column web. The connection will resist an appreciable moment and possess a high degree of rotation capacity if there is no significant moment transferred into the column. If, however, the joint must resist an unbalanced moment, then the minor axis connection is likely to be less efficient because of the low out-of-plane compressive strength of the column web. This point will be further discussed when presenting the test results for the edge joint specimens SCJ18 and SCJ Z '- SCJ 13 ~tr- // / Minor axis / ~ (bottom) t i 40 ~mtula / // // / / J SCJ8 20 ~f / / / / / / jr'major axis r'.'* V / / O-- n I i I Rotation(mRad) Fig. 16. Partial depth endplate connected to the major and minor axis.

21 Composite connections in steel and concrete Composite edge joints (specimens SCJ18 and SCJ19) Two specimens, SCJ18 and SCJ19, were designed as cantilevers with the object of investigating the anchorage situation of the reinforcement in the edge joint and the interaction with out-of-plane compressive strength of the column web. For SCJ18 the plate was placed at the bottom of the beam and for SCJ19 at the top. The experimental results of Ref. 17 have already indicated that poor anchorage of the reinforcement was likely to prevent edge joints from developing high moment resistances. All five cantilever tests of Ref. 17 failed as a result of the concrete being pushed out from the slab behind the column, thus causing loss of anchorage of the reinforcement. Much stronger details in terms of anchoring of the reinforcement in the column area were used this time. All the longitudinal bars were bent into the concrete and extra transverse reinforcement bars were added to the top and bottom of the slab at the back of the column. Trim bars (45 '~) were placed round the back of the column. This was found to give better control of the diagonal cracking. During the loading SCJ18 was very stiff until it exhibited a sudden failure. When the maximum load was reached, the specimen immediately lost stiffness and the load dropped down very fast. The failure initiated from cracking of the bottom of the slab at the back of the column and eventually tensile splitting of the concrete occurred. Anchorage of the reinforcement was lost owing to failure of the slab and the failure was very brittle with a low residual strength. Yielding and large deformation of the column web were also detected, although the load achieved was not very high. The final moment strength was only 86kNm. A similar failure occurred for SCJ19 but with an even lower initial stiffness. The ultimate strength reached was 61 kn m as a result of the failure of the concrete slab. During loading the beam bottom flange of SCJ19 eventually came into contact with the web of the column. The strength of the connection was increased and then failure of the slab occurred. Figure 17 presents the stiffness and strength characteristics for these two specimens. From these further tests on the edge connection, although suitable reinforcement design did increase the moment resistance somewhat, it was still difficult to prevent slab failure and loss of anchorage for the reinforcement as shown in Fig. 18. The mode of failure was a very sudden and brittle one without prior warning and is clearly undesirable. The excessive deformation of the column web should also be noted. Although the failure load was just half that of the same joint detail when tested as a cruciform type, early yielding occurred as unbalanced moment was transferred into the column web. Decreases in the moment resistance are inevitable under the influence of unbalanced moment to the minor axis connections.

22 24 ~: Xiao, B. S. Choo, D.,4. Nethercot 100 8O ~" o /) Rotation(mRad) Fig. 17. Influence of anchorage failure on moment-rotation response of composite joint. edge Fig. 18. Anchorage failure of reinforcement in the edge joint. It is theoretically possible to construct even stronger detailing of the reinforcement at the back of the column to strengthen the concrete stab and to take additional measures to stiffen the column web. The moment capacity will be increased and the brittle failure will possibly be eliminated. For simplicity, however, design of edge joints and corner joints is more

23 Composite connections in steel and concrete 25 realistic if they are treated as simple connections owing to the complexity of these special measures. 3.8 Finplate composite connection (specimens SCJ2, SCJ12, SCJ16 and SCJ17) The tinplate joint shown in Fig. 3(d) was also included in this study. The steel detail derives its flexibility from bolt deformation in shear, bolt hole distortions in bearing in the tinplate and/or the beam web and outof-plane bending of the tinplate. SCJ2 with mesh reinforcement was tested to check the behaviour of such joints when functioning in a composite fashion. Figure 7 shows how the stiffness decreased after the first crack appeared in a similar way to SCJ1. After the peak load, the moment resistance dropped sharply as the cracks became enlarged and the mesh fractured. The ultimate rotation of 16.4 mrad (see defination in EC3) was higher than that of SCJ1. This type of joint was more flexible than the cleated joint. An increased ratio of reinforcement was used in SCJ12 to improve the moment resistance. The moment rotation relation for SCJ12 given in Fig. 19 shows significantly greater ductility in the region of maximum load. The beam bottom flange moved towards the column flange when the loading continued. After contact of the flanges, the connection stiffness increased and the moment resistance was further increased. The ultimate rotation of the beam reached 39 mrad, with eventual failure being caused by the bearing strength of the beam web and excessive twisting of the web plate as shown in Fig. 20. Permanent 240 ~ SCA ~ 1.2% reinforcement and bigger steel section.~, x,:,,-~,,~,, i, Rotation(mRad) Fig. 19. Influence of steel section and metal decking on composite tinplate connection.

24 26 Y Xiao, B. S. Choo, D. A. Nethercot Fig. 20. Bearing failure of beam web and further web twisting. deformation had taken place in the web bolt holes as revealed in the check after the test. The moment resistance of this type of connection was clearly governed by the tensile force of the reinforcement and the bearing in the beam web or tinplate. The performance of both types of connections has clearly been enhanced by the additional reinforcement. The change of the failure mode away from the connection into the members makes it possible to utilize fully the strength of the surrounding members. A tinplate connection with a much deeper steel beam section and increased slab depth was designed for test SCJ16. It was intended to increase further the moment capacity for this type of very flexible connection. Tests on bare steel finplates is have indicated both low stiffness and moment capacity with failure sometimes occurring prematurely by lateral torsional buckling. The shallow beam composite section with a light ratio of reinforcement tested as SCJ12 had already achieved an appreciable moment resistance; the composite slab for SCJI6 was then strongly reinforced with the aim of achieving higher stiffness and strength. The moment rotation curve for the test is shown in Fig. 19. Slip of the bolts in the horizontal direction at a fairly early stage is clear, causing loss of stiffness. The connection regained stiffness when the bolts came into bearing with the edges of holes and exhibited a larger ultimate rotation compared with equivalent endplate connections. The moment capacity was 225 kn m, which is a high hogging region resistance. Debonding of

25 Composite connections in steel and concrete 27 the concrete from the metal decking was visible owing to the large deflection of the beam. Eventual failure was by bearing of the bolts as the beam web moved towards the weld toe of the plate causing torsion of the tinplate and beam web. The bottom flange of the beam finally touched the column flange. SCJ17 was a tinplate composite connection with the metal decking orientation parallel to the beam direction. The aim was to check the contribution from the metal decking to the composite connection. Compared with the equivalent previous specimen SCJ12, the reinforcement ratio was decreased to 0'5% according to the pre-calculation of the potential contribution to the moment capacity of the metal decking. From the moment-rotation relation shown in Fig. 19, it is clear that the stiffness of the specimen was less owing to the lower position of the metal decking compared with that of the mesh and rebar reinforcement. The final ultimate strength was similar to that of SCJ12, which means that metal decking can be regarded as equivalent reinforcement. Its influence on the initial stiffness of the connection should also be considered. The difference between a composite tinplate connection and a bare steel tinplate is that failure of the composite version starts from the bearing capacity of the last one or two bolt rows which must balance all the tensile forces. The premature horizontal torsional buckling caused by the eccentricity of the plate to the centreline of the beam web is avoided owing to the restraint supplied by the concrete floor. Flexible joints like finplates can therefore produce high moment resistances when properly designed to function as composite connections. However, it should be noted that the contribution from the metal decking cannot always be relied upon, especially when it is arranged by butting sheets against a column rather than being continuous past the column. 4 CONCLUSIONS A test programme designed to study systematically the moment resistance, initial stiffness and rotational capacity of composite connections has been described. Several different types of joint were chosen as a way of investigating changes to the major parameters which affect joint behaviour. The following conclusions can be drawn from the analyses of the tests: Different failure modes were observed for the different steel details and the different reinforcement arrangements. Failure was usually through a combination of yielding of the reinforcement and failure of the steel joint part.

26 28 E Xiao, B. S. Choo, D. A. Nethercot 2. The initial stiffness, moment resistance and rotational capacity were dramatically affected by changes to the reinforcement ratio in the slab, metal decking, steel joint type, column web stiffening and moment shear ratio. 3 The provision of adequate anchorage to reinforcement in the edge connection is very difficult owing to the slab around the column being 'pushed off'. To avoid the need for complicated reinforcement detailing, a simple zero moment capacity joint design is realistically suggested for the edge and corner joint. 4. Variation of the position of the partial depth endplate can also change the strength and stiffness of the connection. Following the current practice for bare steel joints does not provide the best behaviour for the composite connection. Placing the endplate at the level of the lower flange is suggested for composite beam-column connection design. 5. The presence of the column web stiffener not onty prevents failure in the column but also increases the moment capacity of the connection without having much influence on the rotation capacity of the beam-column connection. Beam stubs with endplates on the column web can take over the role of web stiffeners to prevent local buckling of the column web. 6. Loose plates on the back of the compressive column flange zone of the flush endplate joint can effectively prevent excessive deformation of the column flange. Working together with the perpendicular plates connected to the column web, backing plates can stiffen the compressive zone of the connection which is usually very weak. 7. Minor axis connection of the partial depth endptate composite connection gives behaviour comparable to the major axis connections. Failure in the column area will be eliminated if there is no significant unbalanced moment transferred into the column. However, when significant moment is transferred into the column web the joint will fail at an early loading stage because of the weak outof-plane compressive strength of the column web. 8. Flexible tinplate composite connections can produce appreciable resistance moment if properly designed. The bearing capacity of the bolts in the lower portion of the tinplate will control the final moment capacity of the connection. ACKNOWLEDGEMENT The project was funded through a grant from the Building Research Establishment as part of their role in calibrating the forthcoming Euro-

27 Composite connections in steel and concrete 29 code 4 for use in the UK. This financial support is gratefully acknowledged. REFERENCES 1. Patrick, M., Hogan, T. J. & Firkins, A., Composite floor systems in commercial buildings. Proceedings of 3rd Conference on Steel Development, AISC, Melbourne, Australia, May 1988, pp Mathys, J. H., Multistory steel buildings--a new generation: the current scene. Struct. Engr, 65A(2) (1987) Wright, H. D., Evans, H. R. & Harding, P. W., The use of profiled steel sheeting in floor construction. J. Construct. Steel Res., 7(4) (1987) Zandonini, R., Semi-rigid composite joints. In Structural Connections: Stability and Strength, ed. R. Narayanan. Elsevier Applied Science, London, 1989, pp Nethercot, D. A., Tests on composite connections. Proceedings of 2nd International Workshop: Connections in Steel Structures, Behaviour, Strength and Desiyn, Pittsburgh, PA, Apr. 1991, pp BS 5950, Structural Use of Steelwork in Building, Part 3.1: Code of Practice for Design of Simple and Continuous Composite Beams. British Standards Institution, London, Eurocode 4, Design of Composite Steel and Concrete Structures. Prepared for the Commission of the European Communities, Draft, Mar Davison, J. B., Lain, D. & Nethercot, D. A., Semi-rigid action of composite joints. Struct. Engr, 68(24) (1990) Xiao, Y., Nethercot, D. A. & Choo, B. S., The influence of composite metal deck floorings on beam-column connection performance. Proceedin,qs ~f 1 Ith International Speciality Conference on Cold Formed Steel Structures, Rolla~ MO, USA, Oct pp Xiao, Y., Nethercot, D. A. & Choo, B. S., Moment resistance of composite connections in steel and concrete. Proceedinqs ~[" the First World Cotference on Constructional Steel Design, Acapulco, Mexico, Nov. 1992, pp Xiao, Y., Choo, B. S. & Nethercot, D. A., Design of semi-rigid composite beam column connection. Proceedinys ~71 CoJ![brence on lnnovatit~e Des#in h~r Structures, Brighton, UK, Apr BS 1881, Testin 9 Concrete, Part 116: Method for Determination of Compressive Strength: Part 117: Method fi~r Determination ~f Tensile Splittin.q Strength. British Standards Institution, London, BS EN , Tensile Testing ~['Metallic Materials. Part I: Method?[Test at Ambient Temperature. British Standards Institution, London, BS 4449, Carbon Steel Bars for the Reinforcement ~[" Concrete. British Standards Institution. London, Moore, D. B. & Gibbons, C., The design of flush and extended endplate connections with backing plates. In Constructional Steel Desiqn: World Developments, ed. P. J. Dowling, J. E. Harding, R. Bjorhovde &

28 30 E Xiao, B. S. Choo. D. A. Nethercot E. Martinez-Romero. Elsevier Applied Science, London, 1992, pp BCSA/SCI, Joints in Simple Construction, Vol. 1: Design Methods, 2nd edn. Steel Construction Institute, Ascot, UK, Lam, D., Influence of composite flooring on steel beam-column connections. MPhil thesis, University of Sheffield, UK, Jan Owens, G. W. & Moore, D. B., Steelwork connections. Struct. Enyr, 70(3/4) (1992)

29 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 ( 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