Multi-response optimization of friction stir welding parameters for AA6063-T6 using hybrid RSM–ANN–GRA–TOPSIS

Friction Stir Welding (FSW) is a solid-state joining technique capable of producing high-quality aluminum joints without fusion-related defects. This study presents a hybrid multi-objective optimization framework that integrates Response Surface Methodology (RSM), Artificial Neural Network (ANN), Grey Relational Analysis (GRA), and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to optimize FSW parameters for AA6063-T6 aluminum alloy. The effects of tool rotational speed, travel speed, and shoulder diameter on ultimate tensile strength (UTS), Vickers hardness (VHN), and peak temperature were systematically evaluated. The optimal parameters—3000 rpm rotational speed, 3 mm/min travel speed, and 20 mm shoulder diameter—produced a GRA–TOPSIS index of 0.9987, yielding a UTS of 185 MPa (97.37 % joint efficiency) and a maximum hardness of 81.97 HV. The hybrid RSM–ANN–GRA–TOPSIS model improved the optimization index by 3.15 % compared with the traditional RSM–ANN approach, while the ANN prediction accuracy achieved R² values of 0.97–0.9992 with RMSE between 1.2 and 6.8. The proposed hybrid framework demonstrates strong predictive capability and offers a reliable methodology for enhancing mechanical performance and process stability in FSW applications.