Shakir-Khalil and Hassan 1994
Push-out tests were conducted on rectangular CFT columns in this research. Several factors that could affect the bond strength, including the shear connector pattern, concrete grade, concrete age, and water-cement ratio were studied
A total of five series of specimens (A, B, C, D, and E) were tested. The axial load was applied to the concrete core at the top, while there was a 1.97 in. gap below the concrete at the bottom of the CFT. Thus, the steel tube alone supported the columns at the bottom. All of the specimens had the same steel tube size with a D/t ratio of 30. The cubic compressive strength of the concrete varied from 3.19 to 9.86 ksi. The columns had a total length of 17.72 in.
All CFTs in test series A included shear connectors. Either black bolts or threaded bars were attached to all of the steel tubes, excluding the control specimen, which had no shear connectors. The threaded bars were inserted either straight or at a 45° inclined position. The direction of inclination and position of these bars also varied. A bond strength of 36.3 psi was obtained for the control specimen. The specimen that used threaded bars as shear connectors exhibited a lower failure load and a more dramatic decrease in strength when it was compared to the specimen with bolts for shear connectors. This was because the threaded bars could not undergo enough rotation, while the black bolts resisted the slip of the concrete core with a significant amount of rotation. However, the initial stiffness of the specimen with the threaded bars was higher. The specimens having inclined threaded bars were simultaneously subjected to shear force and compressive or tensile force as the slip proceeded. It was found that the specimen with shear connectors having inclination in the opposite direction to the slip, in which the shear connectors were subjected to shear and compression, had 30% higher failure load than the specimens having shear connectors inclined in the same direction with the slip. The former specimen also showed the best slip response in this test series in terms of both strength and ductility. The position of the inclined bars that were in the same direction with the slip was found to have no effect on the failure load. However, more ductile slip response was obtained when the threaded bars were placed at the same level on the two sides of the steel tube.
The specimens in tests series B were divided into two subgroups having different concrete strengths as the concrete was casted at different times. Both subgroups had black bolts as shear connectors, although there was also control specimens with no shear connectors. The number and spacing of the shear connectors varied. In addition, for some of the specimens, the steel tube was reinforced with steel plates at the bolt locations. The bond strength of the specimen with no shear connectors was 30.5 psi for the subgroup having relatively higher concrete strength, and it was 60.9 psi for the subgroup with lower concrete strength. The steel plates reinforcing the steel tube caused a 60% increase in failure load, as it caused the bearing capacity of the bolts to increase. On the other hand, this resulted in a more rapid decay in strength as the bolts were forced to fail in shear, which was a more sudden type of failure. The specimens with no stiffening plate behaved in a ductile manner, as the bolts underwent large amounts of rotation. In addition, as the bolts deformed, the shear area increased and the bolts resisted the push-out force both in shear and in tension. The spacing of the bolts did not have any significant affect on the response of the specimens. Test series C was divided into three subgroups with different concrete strengths. All of the specimens had no shear connectors. The average bond strength of the subgroup with the highest concrete strength was lowest. This was attributed to the large shrinkage as a result of high cement content. The average bond strength for all of the groups was less than 58 psi.
Test series D, E and, F had subgroups with concrete ages of 28, 56, and 85 days. Test series D had the highest cement content and its bond strength was the lowest. For each subgroup, when the age of the concrete increased, it was found that the bond strength decreased. The amount of reduction in bond strength was lowest for the test series F, which had the lowest cement content. Although the load-slip response for the specimens with low strength concrete had a high failure load, the specimens exhibited a dramatic drop in strength after the peak load. On the contrary, the response of the specimens with high cement content showed a ductile response and lower failure load. When the post-peak response of the specimens were compared, the ones having low and medium strength concrete had higher strength and this might be due to low shrinkage of concrete in these specimens.
The authors concluded that the amount of shrinkage is important for bond strength. It is possible to limit the shrinkage by using a low cement content and high aggregate-cement ratio.
Shakir-Khalil, H. and Hassan, N. K. A. (1994). “Push-Out Resistance of Concrete-Filled Tubes,” Tubular Structures VI, Proceedings of the Sixth International Symposium on Tubular Structures, Grundy, P., Holgate, A., and Wong, W. (eds.), Melbourne, Australia, 14-16 December 1994, A. A. Balkema, Rotterdam, The Netherlands, pp. 285-291.