Beutel, Thambiratnam, and Perera 2001, 2002
Ten full scale connection tests were performed. Four were performed under monotonic loading and were presented in the first paper while the remaining six were preformed under cyclic loading and presented in a second, companion paper. The fully-restrained connections were between a circular CFT column and an I-shaped steel beam. The connections tested utilized reinforcing bars which were welded to the flanges of the beams and embedded into the concrete.
Experimental Study, Results, and Discussion
The tubes for the columns of all ten tests had a diameter of 16 in, thickness of 0.25 in, nominal strength of 50 ksi, and were cold formed from steel plate with a single full strength butt weld. The concrete has a compressive strength of 5.8 ksi. Nine of the ten tests were conducted with a 14 in deep beam weighing 30 lb/ft, while one of the cyclic tests was conducted with 16 in deep beam weighing 36 lb/ft. All of the beams were compact and all had a nominal yield strength of 44 ksi. Welded connection plates of the same thickness as the flanges were used to join the flanges to the column and welded cleat plates were used to join the beam web to the column. Two reinforcing bars were welded to each flange of the beams and embedded into the concrete core. The yield strength of the bars was 60 ksi and diameter of the bar varied with the different tests. The bars were straight for all four monotonic tests; however, various end details were used in the cyclic tests. The end details included: straight bar, bent bar, and welded to the opposite side of the steel tube.
Only T-shaped specimens, representing an external frame connection, were tested. The column height was 7.9 ft with the connection at mid-height and hemispherical bearing plates at the top and bottom to allow the column ends to rotate. The column was axially loaded with approximately 12.5% of the squash load. Load was applied to the beam tip at a distance of 8.0 ft from the column centerline.
The first four tests were performed under monotonic loading conditions. The bar sizes used for the test varied from a nominal diameter of 0.47 in to 0.94 in. Generally, as the bar size was increased, the ultimate strength increased, as well as the initial rotational stiffness and total energy absorbed by the specimen. However, these increases were not noted between the second largest and largest bars, which both had the characteristics to be described as full strength rigid connections. The proposed capacity prediction model was found to be accurate.
The remaining six tests were performed under cyclic loading conditions. As stated previously, the three different reinforcement bar end details were tested. Anchorage was a typical failure mode for the cyclic specimens. When the inelasticity occurred in the connection, as was the case for the straight bars there was a rapid deterioration of strength and stiffness. When a plastic hinge was formed in the beam, away from the connection, as was the case for the bent bar and welded bar, the performance improved. This led to the recommendation by the authors that, for seismic applications, as a minimum, bent bars should be used.
Reference
Beutel, J., Thambiratnam, D., and Perera, N. (2001). “Monotonic Behaviour of Composite Column to Beam Connections,” Engineering Structures, Vol. 23, pp. 1152-1161. Beutel, J., Thambiratnam, D., and Perera, N. (2002). “Cyclic Behaviour of Concrete Filled Steel Tubular Column to Steel Beam Connections,” Engineering Structures, Vol. 24, pp. 29-38.