Multi-objective optimization of welding-induced residual stress and deflection in 6082-T6 aluminum alloy using validated thermo-mechanical modeling

• A validated simulation predicts weld temperature, deflection, and residual stress. • Effects and interactions of welding factors on mechanical response were identified. • Optimization balances deflection and residual stress, supporting welding design. Welding of aluminum alloys often introduces residual stress and deflection, compromising dimensional precision and structural performance. This study investigates the influence of key process parameters of gas metal arc welding on the thermo-mechanical response of 6082-T6 aluminum alloy butt joints. A numerical method was developed and validated using experimental measurements of temperature distribution (thermocouples), deflection (3D laser scanning), and residual stress(X-ray diffraction). A full-factorial design of experiments (DOE) was conducted, varying clamping configuration, plate thickness, welding sequence, and cooling conditions. Analysis of variance (ANOVA) quantified main and interaction effects. The study identified a trade-off between deflection and residual stress, which was addressed through multi-objective optimization using a desirability function approach. Deflection was reduced from 1.44 mm (measured experimentally) to 0.6 mm under optimized conditions, while the minimum residual stress was 171 MPa, representing a decrease of approximately 12%. The optimum condition corresponded to a partially restrained clamping configuration, a plate thickness of 4 mm, a continuous single pass welding sequence, and natural air cooling. Predictive models based on ensemble regression techniques were constructed using the 72 DOE-based FEM cases and validated with experimental measurements to estimate responses and rank influential parameters. The models achieved an R² values of 0.93 for deflection and an R² value of 0.94 for residual stress. Consistency between statistical and predictive analyses confirmed the dominant factors. The optimization framework offers a data-driven approach to improve welded structural integrity and highlights the potential of integrated simulation and data analysis in materials processing and design.