# Nakahara and Sakino 2000b

In this paper, an analytical study to estimate the load-deflection response of square CFTs under eccentric axial load was presented. The theoretical results were compared with the load-deflection curves of specimens tested in Japan. In addition, a correlation was also investigated between the experimental peak moments and the ultimate strength moments computed by the AIJ (1997) and ACI (1989) design code provisions.

## Experimental Study

Twenty-one square CFT columns were tested under eccentric axial load as a part of the research program. The main parameters of the tests were yield strength of steel tube, compressive strength of concrete, magnitude of eccentricity, and D/t ratio. The ranges for concrete and steel strength were 3.86-11.64 ksi and 38.00-89.64 ksi, respectively. The D/t ratio was varying between 22.8 and 73.8 and the L/D ratio for all the specimens was equal to 3.

## Analytical Study

In the previous work of the authors (see Nakahara, H. et al., 1998), axially loaded square CFT columns were studied. The proposed steel and concrete compressive stress-strain relations worked well to predict the experimental response of axially loaded specimens. In the current study, similar compressive stress-strain curves were used. However, the slope of descending parts was milder to simulate the more ductile behavior seen in the tests with flexure due to eccentric loading. In addition, the strain gradient along the section was accounted for, with a linear strain distribution being assumed.

While calculating the load-deflection response of the specimens, the deflected shape was taken as a sine wave. Based on this assumption, the relation between mid-height curvature and mid-height lateral deflection was derived. Using the equilibrium condition, moment-curvature-thrust curves were then formed. It was found that ductility was affected primarily by the D/t ratio and compressive strength of concrete rather than the yield strength of the steel or the eccentricity. The analytical moment-curvature-thrust curves showed good correlation with the experimental results and predicted the post-buckling response of the specimens accurately. However, the initial flexural stiffness was overestimated, which was attributed to residual stresses existing in the specimens. The moment capacities of the specimens were calculated by the methods in AIJ (1997) and ACI (1989). The concrete compressive strength was multiplied by a reduction factor ranging from 0.855 to 0.960 in both methods to account for the scale effect. The experimental capacities were overestimated by AIJ (1997) and underestimated by ACI (1989).

## References

Nakahara, H. and Sakino, K. (2000b). “Practical Analysis for High Strength CFT Columns under Eccentric Compression,” Composite and Hybrid Structures, Proceedings of the Sixth ASCCS International Conference on Steel-Concrete Composite Structures, Xiao, Y. and Mahin, S. A. (eds.), Los Angeles, California, March 22-24, 2000, Association for International Cooperation and Research in Steel-Concrete Composite Structures, Los Angeles, California, pp. 441-448.