Double-sided friction stir welding of AA 2024-T6 joints: Mathematical modeling and optimization

Abstract

The heat-treatable AA 2024-T6 has been used widely in the aerospace and defense industry. In this paper, the effective parameters of the double-sided friction stir welding (DFSW) process of AA 2024-T6 were investigated by using Response Surface Methodology (RSM). Also, the optimal parameters have been obtained by using the Artificial Neural Network (ANN) and optimization algorithms. The most significant process parameters, i.e. rotational speed, welding speed, tilt angle, and tool pin length, were considered to predict the behavior of the response, i.e. ultimate tensile strength (UTS) and percentage elongation (% El). For this purpose, a central composite design, as well as analysis of variance, was used to investigate effective parameters. Also, an intelligent connection between the above parameters and responses was developed by using ANN. The optimal neural network was integrated with the whale optimization algorithm (WOA) to find optimal parameters and achieve desirable responses. The obtained result was verified by experiment. Results indicated that the tool pin length (h) factor has the highest contribution percentage between the parameters (14.95% and 9.59% for UTS and %El, respectively). Besides, the parameters of welding speed (v) and tilt angle (α) have a significant effect on the UTS and %El of DFSW joints, respectively. Using a rotational speed of 1250 rpm, welding speed of 80 mm/min, the tilt angle of 4°, and a tool pin length of 2.3 mm, as optimal parameters of the DFSW process, can be achieved the maximum of ultimate tensile strength and elongation.