Masuo et al. 1991
The buckling behavior of CFT columns was studied both experimentally and analytically, using both lightweight and normal weight concrete subjected to concentric axial load. Ultimate loads were discussed with regard to three parameters: concrete weight, size of the steel tube, and effective column length.
Experimental Study, Discussion, and Results
Twenty-six columns were tested, 18 lightweight and 8 normal weight. Initial out-of-straightness was accounted for by assuming a mid-height deflection of L/2000 for all columns except those with a slenderness factor
of 0.8-1.0, where Leff is the effective column length and Po is the squash load, given by
The initial deflection at mid-height for columns in this range of slenderness ratios was computed as the deflection of the column before the test divided by the effective length. In the tests, an initial deflection of L/4000 or L/8000 was used, the higher number for more slender columns. The authors found that both weights of concrete with slenderness factors around 0.3 were definitely affected by confinement, the normal weight concrete showing a somewhat larger effect. The load-deflection relations were also significantly affected by the confining effect in this range of slenderness factors. Several detailed graphs elucidate this point. Varying the D/t ratio from 30-40 and holding the other test parameters constant did not seem to affect the squash load. Finally, the ultimate loads of the CFTs for both weights at a slenderness factor of 0.6 were somewhat larger than the European column curve, to which they were compared to.
Analytical Study
Load-deflection curves were developed using a mechanics-based analytically method incorporating the following assumptions. A sinusoidal shaped deflection curve with an initial out-of-straightness was used with equilibrium taken at mid-height on the basis of assuming plane sections remain plane. The stress-strain relations for the steel were assumed to follow the Ramberg-Osgood function. Two different constitutive equations were used for the concrete based on Popovics's formula, one assuming no confinement and the second assuming confinement. Load-deflection relations that ignored confinement were less than the experimental results, especially when slenderness factors are small, suggesting significant strength augmentation due to confinement. The analytical results made accurate predictions of the load deflection curves when confinement was assumed in the model.
Reference
Masuo, K., Adachi, M., Kawabata, K., Kobayashi, M., and Konishi, M. (1991). “Buckling Behavior of Concrete Filled Circular Steel Tubular Columns Using Light-Weight Concrete,” Proceedings of the Third International Conference on Steel-Concrete Composite Structures, Wakabayashi, M. (ed.), Fukuoka, Japan, September 26-29, 1991, Association for International Cooperation and Research in Steel-Concrete Composite Structures, pp. 95-100.