Quality prediction of friction stir welded joint based on multiple regression: entropy generation analysis

Due to the outstanding advantages in light alloy material processing, friction stir welding technology is of great significance for obtaining high-quality welding products and accelerating aerospace lightweight. At present, most related studies predict joint quality before welding through computer technology. However, due to the complex thermo-mechanical coupling in the welding process, there may be a large deviation between the predicted results and the actual results, resulting in energy waste. In addition, these prediction models have poor real-time and versatility. Therefore, a new joint quality prediction method based on entropy production analysis is proposed in this paper. Firstly, based on non-equilibrium thermodynamics and extrusion theory, the entropy generation analysis model of the friction stir welding system is deduced. Using Liouvile’s Formula, the analytic solution of the entropy generation analysis model with unknown parameters is obtained. Secondly, combined with numerical simulation and multiple regression, the unknown parameters of the entropy generation analysis model are determined. Finally, multiple sets of welding experiments are designed to verify the effectiveness of the entropy generation analysis model. The welding process is analyzed by the proposed entropy generation analysis model to achieve the quality prediction of friction stir welded joint.