Multi-Objective Optimization of MIG Welding and Preheat Parameters for 6061-T6 Al Alloy T-Joints Using Artificial Neural Networks Based on FEM

To control the welding residual stress and deformation of metal inert gas (MIG) welding, influence process parameters preheat (welding speed, heat input, temperature, area) is discussed, a prediction model established to select optimal combination parameters. Thermomechanical numerical analysis was performed obtain according 100 × 150 50 4 mm aluminum alloy 6061-T6 T-joint. Owing complexity process, an Latin hypercube sampling (OLHS) method adopted for with uniformity stratification. Analysis variance (ANOVA) used find degree speed (7.5–9 mm/s), input (1500–1700 W), temperature (80–125 °C), area (12–36 mm). The range research are material, method, thickness plate, procedure specification. Artificial neural network (ANN) multi-objective particle swarm optimization (MOPSO) combined effective minimize stress. results showed that had greatest effect on minimization stress, followed by area, respectively. Pareto front obtained using MOPSO algorithm ?-dominance. minimum at same time, when selected as preheating 85 °C 12 mm, 8.8 mm/s 1535 W, were validated finite element (FE) method. error between FE compromise solutions less than 12.5%. optimum in can be chosen designers actual demand.