# Kawano and Matsui 1997 II

An experimental study was conducted to examine the axial behavior of circular CFTs. The specimens were also analyzed using finite element analysis. From the experimental and computational results, equations were derived to estimate the energy dissipation capacity of CFTs and the number of cycles up to fracture of the steel tube.

## Experimental Study Results and Discussions

Both HTs and CFTs were tested. The specimens were subjected to cyclic axial loading until fracture. Pin-ended support conditions were provided. Values for the D/t ratio, L/D ratio, and loading pattern were the main parameters of the experimental program. The specimens were specified as short, medium, and long according to having L/D ratios of 5, 10, and 20, respectively. Values for the D/t ratios were 20, 30, and 50. The specimens were manufactured from steel tubes having diameters of 2.38 in. or 4 in. The ranges for the measured tensile yield strength of steel and measured compressive strength of concrete were 44.8 to 61.2 ksi and 4.37 to 6.69 ksi, respectively. Displacement-controlled loading was applied. For short columns, the loading pattern was controlled by measuring axial displacement at the mid-length. In the case of medium and long columns, the loading pattern was controlled by measuring the axial displacement between the two ends and the axial displacement at the mid-length was still recorded although it was not used as a control parameter. The HT specimens exhibited fewer cycles prior to fracture, and they showed a rapid decrease in strength after local buckling. The concrete infill improved the cyclic behavior, and the CFT specimens experienced stable hysteresis loops and a larger number cycles until fracture. An increase in the D/t ratio and an increase in the loading amplitude both reduced the number of cycles prior to fracture. A more dramatic decrease in strength was observed for the specimens with high L/D ratios. The diameter of the steel tubes did not affect the response of the columns significantly. The number of cycles until fracture was consistently larger than the number of cycles until local buckling. Consequently, it was concluded that the capacity of the CFTs should be based on the energy absorption capacity until fracture rather than the energy absorption capacity until local buckling. The energy absorption capacities of the specimens were affected by the D/t ratio and the amplitude of the axial displacement. It was found to decrease with an increase in D/t ratio and an increase in axial displacement amplitude.

## Analytical Study

The tested columns were analyzed using the finite element method. A fiber-based cantilever beam-column model was employed in the analysis and axial loads were applied with an eccentricity of 5% of the tube diameter to account for imperfections. The stress-strain relationships for steel and concrete were adopted from the available literature. Local buckling and slip along the concrete-steel interfacebetween steel and concrete were not taken into account. Good correlation was achieved between the experimental and computational peak axial loads. However, the hyteresis loops were not estimated with high accuracy. This was attributed to local buckling, which was not considered in the analysis. An equation was proposed using the experimental results in order to estimate the number cycles until fracture. It was dependent on the D/t ratio and amplitude of the longitudinal strain at the mid-length of the specimens. The specimens having medium L/D ratio were shown to have the smallest number of cycles until fracture because both compression and bending were effective to increase the amplitude of longitudinal strain. In the case of long or short columns, either bending or compression governed the behavior alone. Consequently, their amplitude of longitudinal strain was calculated to be smaller and they had a higher number of cycles until fracture. Multiplying the number of cycles until fracture with the energy absorption at each cycle, both of which were obtained analytically, the energy absorption until fracture was calculated. Good correlation was achieved with the experimental results and the trend for the effect of the L/D ratio on the CFT energy absorption was similar to the one observed while calculating the number of cycles until fracture.

## References

Kawano, A. and Matsui, C. (1997). “Buckling Behavior and Aseismic Properties of Concrete-Filled Tubular Members under Cyclic Axial Loading,” Composite Construction in Steel and Concrete III, Buckner, C. D. and Shahrooz, B. M. (eds.), Proceedings of the Engineering Foundation Conference, Irsee, Germany, June 9-14, 1996, American Society of Civil Engineers, New York, New York, pp. 602-615.