Physics-informed FEA–GPR–PSO optimization of perforated CARALL structures for simultaneous bending strength increase and lightweight design

Fiber metal laminates combine the advantages of metals and carbon fiber–reinforced polymers, among which carbon fiber–reinforced aluminum laminates (CARALLs) have attracted significant attention due to their excellent mechanical performance. However, insufficient bonding strength at the metal–adhesive interface often leads to interfacial delamination. Perforating the metal layer to form pin-like mechanical interlocking structures is an effective solution, while also enabling weight reduction. Nevertheless, perforation design involves multiple geometric parameters, including hole diameter, pitch-to-diameter ratio, hole type, and staggered arrangement, and improper combinations may induce stress concentration and stiffness degradation. Moreover, systematic optimization strategies suitable for high-end industrial applications remain limited. To address these challenges, this study proposes a physics-guided, data-driven optimization framework integrating finite element analysis (FEA), Gaussian process regression (GPR), and particle swarm optimization (PSO) for perforated CFRP–aluminum–CFRP laminates. Experimental tests were conducted to calibrate and validate the FEA model. Based on an experimental design, surrogate models were established for tensile and bending strength, achieving normalized root mean squared errors of 4.14 % and 3.35 %, respectively. Global optimization results demonstrate a 5.47 % reduction in structural weight and a 9.52 % improvement in bending strength, with only a 1.82 % decrease in tensile strength, highlighting the effectiveness of the proposed framework in balancing design.