Yao, Goldsworthy, and Gad 2007
Numerical and experimental studies of T-stub connections are presented in this paper. The experiments consisted of one full scale T-stub connection using blind-bolts and a series of pullout tests on bent reinforcing bar. Two numerical studies were also conducted, namely, a three dimensional finite element model of the T-stub connection and a one dimensional anchorage model.
Experimental Study, Results, and Discussion
Blind bolts are mechanical connectors with which access to only one side of the connection is necessary. Several proprietary bolts are available commercially. In this experiment, a reinforcing bar with a 90° bend is welded to the interior end of these bolts. The weld was designed to exceed the yield strength of the bar.
A T-stub connection using these blind bolts with reinforcing bar extensions was tested under monotonically increasing tension. Two T-stubs were attached on either side in the middle of a 40 in. long circular CFT. This connection models the tension flanges of a double-tee joint between a steel beam and circular CFT column. The diameter of the tube was 12.75 in., the thickness was 0.24 in. and was made of 50 ksi steel. The end plate was 0.79 in. thick, made of 44 ksi steel, and curved to fit the circular tube using brake press. The flanges were 0.63 in. thick, made from 36 ksi steel, and welded to the curved end plate. A total of six 0.63 in. diameter blind bolts with a yield strength of 93 ksi and ultimate strength of 116 ksi were used in each end plate. The reinforcing bar extensions were 0.63 in. diameter and had a yield strength of 73 ksi. Concrete with a compressive strength of 6.5 ksi was used to fill the tube.
The specimen reached a maximum capacity of 155 kips when failure occurred in the weld between the blind bolt and the reinforcing bar. The outward displacement was assumed to be made up of two components, the slip of the bolts and the end plate deformation. At low loads it is shown that the outward displacement is primarily due to end plate deformation, whereas, at high loads, the outward displacement is primarily due to slip of the bolts. Through examining the strain measurements, it was determined that nearly 65% of the transferred load was taken by the reinforcing bar while the remaining 35% was taken by hoop stress in the steel tube.
In addition to the connection test, eighteen pullout tests were conducted to evaluate the anchorage capacity of the reinforcing bar extensions. The eighteen tests were divided into three groups, the first group was tested monotonically, the second and third groups were tested cyclically at different load levels. Within each group the test variable was thickness of the steel tube, either 0.24 in., 0.31 in., or 0.39 in, leaving duplicates of each test. The other dimensions and material properties were identical to the connection test. All of the tested bars failed by fracture away from the tube rather than a bond failure. Under the imposed cyclic load, little deterioration of strength was noted.
Analytical Study
A finite element model of the T-stub connection was created in ANSYS for comparison to the experimental results. Eight-node brick elements were used to model the tube, plates, and bolt, while, nonlinear springs were used to model the reinforcing bar. The concrete was modeled as rigid and contact constraints were employed. Pretension was modeled in the bolts. The results of this model compared well to the experimental results. An anchorage model of the reinforcing bar in the concrete was created using one dimensional nonlinear springs. The results of this model also compared well to the experimental results.
Reference
Yao, H., Goldsworthy, H., and Gad, E. (2008). “Experimental and Numerical Investigation of the Tensile Behavior of Blind-Bolted T-Stub Connections to Concrete-Filled Circular Columns.” Journal of Structural Engineering, Vol. 134, No. 2, pp. 198-208.