Modern techniques for designing high-strength circular concrete-filled tube columns: Knowledge-guided data approach

The recent updates in the AISC Specification for high-strength composite columns apply exclusively to rectangular shapes. This research is a response to the increasing demand to include high-strength circular concrete-filled tube (HS-CCFT) members in design considerations, defined as concrete compressive strength exceeding 70 MPa and steel yield strength exceeding 517 MPa. The paper introduces a unique methodology that combines 320 experimental test results with knowledge-guided advanced data-driven models, namely Gene Expression Programming (GEP) and Artificial Neural Networks (ANN), to develop simplified, partially mechanical-based design equations. The proposed equations include material modulus ratios alongside the traditional confinement factors, providing more reliable estimates of concrete confinement. The analysis also indicates that the limiting slenderness ratio for local buckling in HS-CCFTs is higher than in conventional CCFTs, extending up to D/t = 220. The proposed plastic stress distribution method, which accounts for steel local buckling and concrete confinement, eliminates the need for section classification, thus simplifying the design process while improving prediction accuracy. The statistical metrics CoV of 0.1, mean of 1.05, and R² of 0.99 demonstrate the exceptional reliability of the developed equations, presenting a strong case for inclusion in future design codes and offering practical tools for engineers to optimize structural design and safety.