Matsui and Tsuda 1987
The authors presented an experimental study of CFT beam-columns subjected to vertical axial load and lateral load. The main test parameters were the presence of concrete in the tube, the D/t ratio (concetrating on CFTs with larger D/t ratios), and the method of loading--monotonic or cyclic. The authors proposed an even larger value for the limiting width-thickness ratio of the tube to 2.0 times that of a hollow steel tube. The first author had earlier proposed a value of 1.5 (Matsui, 1986).
Experimental Study, Results, and Discussion
The test program consisted of twenty-six cantilever specimens ranging in D/t from 47 to 94--twelve hollow specimens and fourteen CFT specimens. All of the hollow tubes were tested monotonically; six CFTs were tested cyclically and the remainder monotonically. Each cantilever specimen was oriented vertically to simulate a column in a frame structure. A constant axial load ranging from 0.1*Po to 0.4*Po was applied vertically to the top of the column and a lateral load was applied to the top of the column at a distance of 29.5 inches from the base, which approximates the location of the inflection point in a frame structure.
The experimental results were compared to the capacity predicted by a plastic collapse mechanism line, which assumes a plastic hinge forms at the base of the beam-column. Similar to the results of Matsui (1986), the hollow tube was unable to reach the mechanism line due to local flange buckling and then local web buckling of the tube at the beam-column base. The CFT specimens, on the other hand, invariably reached the mechanism line, even specimens having a D/t ratio of 2.0 times the recommended Japanese design specification to prevent local buckling. The cyclic hysteresis curves for the CFTs exhibit large ductility and large energy-absorption capacity. As the D/t ratio or the axial load ratio of the beam-column increases, the effect of the concrete becomes much more apparent. Hollow tubes exhibit a noticable loss of capacity and ductility as these quantities increase, while CFTs do not.
The maximum strength was compared to the predicted capacity of the section which was calculated by using a fully plastic stress distribution--all of the steel and the concrete in compression contribute moment resistance. The ratio of the maximum experimental moment to the calculated moment was conservative for most cases, ranging from 0.91 to 1.11. The authors propose this method of calculation over the Japanese Standard, which uses 0.85*f'c and is more conservative. The ratio of experimental to the author's calculated moment for the steel tubes greatly underestimated the capacity, ranging from 0.27 to 0.72.
Reference
Matsui, C. and Tsuda, K. (1987). “Strength and Behavior of Concrete-Filled Steel Square Tubular Columns with Large Width-Thickness Ratio,” Proceedings of Pacific Conference on Earthquake Engineering, Vol. 2, Wellington, New Zealand, pp. 1-9.