The results of a theoretical and experimental investigation of square, pin-ended, eccentrically loaded CFT columns were presented. The principal variables of the eight tests were the eccentricity of loading, the slenderness ratio, and the inclination of the loading axis (i.e., the angle of the applied load relative to the principal axis of the member). The author presented a very detailed description of the results.
Experimental Study, Discussion, and Results
Influence of the Concrete Core. Under axial loading, the concrete core provided about 30% of the load carrying capacity while under pure bending it only provided 7.5%. If no local buckling occurs, the moment capacity of the hollow tube is adequate for calculating the design load for pure bending (as was proposed by Furlong, 1967). The contribution of the concrete is enhanced by axial force. For increasing axial load, the concrete provides an increasing proportion of the moment capacity. Since the steel (which has a higher elastic modulus than the core) is located on the outer perimeter of a CFT, it represents a larger portion of the member stiffness. Therefore, for low axial loads, the hollow tube undergoes only 25% more deflection due to bending than the equivalent CFT. While it is widely recognized that the triaxial effect plays a role in strengthening short circular columns, it was not evident for the tested square columns, echoing earlier investigator's conclusions.
Bond. The author noticed no evidence of slip in the tested specimens, but he drew no definitive conclusions in this regard. The analytical method assumed perfect bond. Since the results were very close to the experimental, this suggested that bond was maintained in the member. If bond did not exist, the load-carrying capacity of the section was not significantly affected.
Ductility. The author's tests reinforced earlier investigative reports of the CFT's ability to withstand large deformations. In the unloading portion of the experiment the CFTs were able to maintain a high proportion of their maximum loads even in a state of large deformation. This was largely due in part to the stabilizing role of the concrete in preventing early local buckling of the tube.
Biaxial Bending. Bending tests were performed with the moment applied at three axes of inclination with respect to the perpendicular faces of the square cross section: 0°, 30°, and 45°. The orientation of the tube's axes with respect to the bending direction had little effect on the amount of moment the member could withstand. The columns deflected in the plane of the applied moment and displayed no twisting due to their very high torsional rigidity.
Stability. For slenderness ratios above 45, the maximum axial load was reached before the member realized its full moment capacity. Columns with such large slenderness ratios do not reach their full cross-sectional strength before the member fails by overall flexural buckling.
Using moment-thrust-curvature relationships, an elastic-plastic column stability analysis was performed. Simplified stress-strain relationships were used for the steel and concrete. The procedure was based on the method developed by Roderick and Rogers. Limited details about the procedure were given. The analytical predictions matched the experimental deflections exactly up to the yield point of the specimens. After yielding there was some discrepancy. The author attributed these variations to a number of possibilities: errors in determining the steel tube yield stress, variation of yield across the cross section, a different concrete strength than assumed, or a residual stress pattern that was not accurately simulated by the stress-strain curve.
The results of the bond tests were inconclusive and the author recommended further study of this phenomenon. He also questioned the applicability of pin-ended, single column results to an entire building frame where entirely different loading conditions could exist. Experimental tests regarding the behavior of connections and the interaction between structural members was a high recommendation of the author. At the paper's writing, investigation into the latter problems was being conducted at the University of Sydney.
Bridge, R. Q. (1976). “Concrete Filled Steel Tubular Columns,” Report No. R283, School of Civil Engineering, University of Sydney, Sydney, Australia, 1976.