Prion and Boehme 1989
The authors presented results from tests of thin-walled steel tubes filled with high strength concrete. The program focused on the behavior up to the ultimate load and beyond, with an emphasis placed on ascertaining the amount of ductility that could be achieved with these members. Loading combinations included pure axial compression, pure bending, and various combinations of the two. Three cyclic load tests were performed along with the monotonic tests. The results were compared to design codes based on either strain compatibility or superposition (Japanese code). The description of the tests and the presentation of the results was extremely thorough and detailed and included a number of very illustrative graphs and figures.
Experimental Study, Results, and Discussion
Each of the tested specimens for all types of loading had the same cross-sectional properties, allowing a direct comparison of the results. In general, all of the specimens behaved in a ductile manner. The high strength concrete, however, behaved in a brittle manner.
Short Columns To observe the effect of confinement, the authors tested columns with a slenderness ratio (L/D) of less than 15. Along with the slenderness ratio, the compressive capacity of the section depends on the D/t ratio and the method of load application. To test the effect of load application, two types of tests were performed -- one loading the concrete alone, and one loading both the concrete and the steel equally. Preferably, only the concrete should be loaded to maximize the effectiveness of the steel tube (i.e., have it only subjected to tensile hoop stresses from the expanding concrete). However, in reality, friction and chemical bond between the two materials will cause axial stress in the steel due to a load transfer from the concrete to the steel. The short columns showed two stages of failure. The first stage involved a shear failure in the concrete causing an abrupt loss of capacity. The shear failure of the concrete caused the two wedges to slide past one another, effectively alleviating the axial compression in the steel and inducing additional circumferential stress in the steel tube. The failure mode at the secondary load level was of the desirable ductile type, although the tube was able to sustain only about 60% of the ultimate load. There was no noticeable difference in the results of the different methods of load application; the specimens with the concrete loaded only and both materials loaded uniformly performed quite similarly.
Beams Load was applied to the beam at two points such that there was a region of constant moment at the center of the beam. The length of this region was varied to study the effect of shear on the moment capacity of the section. No definite trend was recognized. A cyclic test was performed on one beam at a ductility level of two to three times the deformation at first yield until failure occurred. The beam specimens failed in a ductile manner at the points of load application. Shear fracture of the concrete initiated failure. This was followed by tensile yielding and subsequent rupture of the steel tube. Local buckling also occurred in the compression region of the beam near ultimate. Significant slippage between the materials was observed at the ends of the specimens. Although this did not seem to affect the moment capacity, it would have a substantial effect on the stiffness of the member. The specimen under cyclic loading showed a slight decrease in strength with each cycle but dissipated a large amount of energy.
Beam-Columns These tests were similar in fashion to the beam tests. The axial load was applied first and then transverse load was applied at two points creating a constant moment region. Two cyclic tests were performed, again with the load cycled at a ductility of two to three times the deformation at first yield. Failure of the beam-columns occurred at the center of the specimens. After cracking of the concrete and buckling of the steel tube in the compression region, the member failed due to a rupture in the tensile zone. These members showed a significant increase in moment capacity due to the presence of an axial load. The axial load causes a greater utilization of the concrete in the compression zone than occurred in the beam tests. The post-failure behavior mimicked the behaved observed in the beam tests; the moment capacity of the beam-column beyond ultimate decreased to a nearly equivalent level as the beam. The cyclic tests exhibited good ductile behavior. Failure occurred after 4 1/2 cycles due to tensile fracture. Slight pinching of the hysteresis loops was observed which was attributed to the opening and closing of concrete cracks combined with the yielding and buckling of the thin steel tube.
Eccentrically Loaded Columns The eccentric loading tests were performed using a high axial load. Unlike the previous tests, the eccentrically loaded columns failed in a non-ductile manner. The authors attributed this to the brittle nature of the high strength concrete and the lack of proper confinement. The authors warn that this behavior should be of concern since these members are often used in earthquake regions.
Comparison of Results
It was found that, with the exception of the beam-columns subjected to a high axial load, the CFTs could be modeled accurately with a strain-compatibility model. Behavior of CFTs beyond ultimate should be described by a superposition model treating the two materials as individual elements. The Japanese superposition model was generally unconservative. Due to the lack of confinement from the thin steel tube, the concrete in the tests was unable to reach its full moment capacity. Also, local buckling of the tube prevented the full steel plastic moment capacity to be achieved.
Reference
Prion, H. G. L. and Boehme, J. (1989). “Beam-Column Behavior of Steel Tubes Filled With High Strength Concrete,” Proceedings of the Fourth International Colloquium, Structural Stability Research Council, New York, pp. 439-450.