A UNIFIED PREDICTIVE DESIGN METHOD FOR SEISMIC BEARING CAPACITY OF DSCSW–BEAM CONNECTIONS INCORPORATING CORRECTION FACTORS Α, Β, Γ — VALIDATION AND RECOMMENDATIONS

Abstract

A predictive analytical method for estimating the bearing capacity of double-skin composite shear wall (DSCSW) connections with embedded H-section beams is proposed and validated. The model incorporates three correction factors: α (embedment length), β (embedment height), and γ (plate thickness). Using 13 nonlinear FE simulations and experimental benchmarks, predicted capacities were compared against numerical results. The baseline formula underestimated capacity by up to 18.7% without correction factors. After calibration, the modified model reduced prediction error to ±6.2% across all cases. Sensitivity analysis revealed that α contributed up to 14.5% variation in strength, β contributed 11.3%, and γ contributed 9.8%. The corrected model successfully captured the nonlinear interaction between embedment geometry and shear transfer mechanisms. A design example demonstrated that for an embedment length of 600 mm, height of 350 mm, and plate thickness of 20 mm, the proposed model predicted a capacity of 875 kN, closely matching the FE-obtained 892 kN (error 1.9%). Statistical validation (R² = 0.96, RMSE = 21.4 kN) confirmed the reliability of the formula for engineering practice. This study contributes a unified predictive framework with quantifiable accuracy, offering practical guidance for structural designers and future code recommendations.