Prediction of ultimate load capacities of CFST columns with debonding by EPR

Concrete filled steel tubular (CFST) structures have become a viable alternative to reinforced concrete or steel structures due to several advantages. One of the most important is the confinement effect of the concrete core provided by the steel tube. However, this beneficial composite action will be probably weakened by debonding, so reliable formulations to predict the reduced ultimate resistance are needed. In this paper, a review of existing specifications and experimental tests carried out on compressed circular CFST columns with and without debonding, is given. Accordingly, more circular CFST long specimens with debonding should be necessary to understand the combined influence of slenderness ratio, load eccentricity ratio and confinement factor on the reduction coefficient (KD) of ultimate load capacities (Nu). The combined influence of arc-length ratio and thickness of circumferential debonding gap on KD should be further studied by experimental tests. Moreover, the existing formulae for KD and Nu show a low accuracy in predicting the test results and should be improved. To this aim, an evolutionary polynomial regression (EPR) MOGA-based methodology was performed to obtain more accurate formulations for Nu and KD of circular CFST columns with debonding. The formulae extracted from the Pareto front of non-dominated solutions, demonstrate good accuracy, higher than the ones in literature. The proposed models are consistent with the physical interpretation of the studied phenomenon according to which Nu and KD decrease as debonding parameters increase and can be used to calculate the real resistance of CFST structures.