DATA–DRIVEN FINITE ELEMENT MODELLING OF STUB RECYCLED AGGREGATE CONCRETE-FILLED STEEL TUBE COLUMNS
Abstract
<p>Concrete-filled steel tubes incorporating recycled aggregate concrete represent a sustainable composite structural members in which the steel tube provides effective confinement and enhances column performance. Although the experimental database for such members has grown in recent years, robust high–fidelity numerical models remain limited. This study presents a computational framework that combines Bayesian optimisation with nonlinear FEM analysis in ABAQUS, automated through MATLAB–Python scripting. The approach optimises key parameters of the confined concrete stress–strain relationship while reproducing experimental axial load–displacement responses of stub columns in compression. The optimised parameters are then used in combination with the regression analysis to develop reliable nonlinear constitutive models for the concrete within steel tube. Additional comparison with two established confined concrete models shows good agreement in load–displacement behaviour and improved performance. The proposed framework is efficient and transferable.</p>