Abstract
Friction stir welding (FSW) method becomes an effective technique for welding dissimilar alloys such as AA2024 and AA5083 as the conventional fusion welding methods are not very suitable for welding the dissimilar alloysand cause many defects during solidification process. Since the strength of FSWedjoints depends on the process parameters, a new method was used in this study to predict the tensile behavior of friction stir welded AA5083 and AA2024 aluminum alloy joints in terms of tensile elongation (TE) and ultimate tensile strength (UTS). A new metaheuristic algorithm called Marine Predators Algorithm (MPA) has been integrated with Random Vector Functional Link (RVFL) network to improve the prediction accuracy. Rotational speed, welding speed, tool axial force, and pin profile were used as input parameters while UTS and TE works as the corresponding outputs. The MPA-RVFL model was tested and validated,a great agreement was demonstrated between the experimental results and the predicted results indicating that the developed technique is accurate and reliable to predict the tensile behavior of welded aluminum joints.