Machine learning prediction and optimization of resistance spot welding tensile shear characteristics parameters in AISI 316L and DSS 2205 stainless steel alloys

This study investigated the tensile shear strength of resistance spot welds joining duplex stainless steel (DSS 2205) and austenitic stainless steel (AISI 316 L). Researchers varied key input parameters to assess their impact on joint strength and quality. Welding current was set at 5000 A, 6000 A, and 7000 A. Electrode tip diameter was tested at 4 mm and 8 mm. Welding pressure remained fixed at 4 MPa, while welding time ranged from 2 to 6 s. These parameter levels allowed systematic exploration of weld formation and mechanical behavior. Experimental results revealed that higher welding current and longer time generally increased tensile shear strength. Peak values reached 318.63 MPa under conditions of 7000 A and 5 s. To predict these outcomes, the XGBoost machine learning algorithm was applied. The model achieved excellent performance, with an R2 value of 0.99, mean squared error of 0.31, root mean squared error of 0.56, and mean absolute error of 0.42. Predicted and experimental results showed strong agreement, validating the model and demonstrating that optimal welding parameters enhance joint performance and support efficient automation of dissimilar stainless-steel resistance spot welding.