Machine learning approach for predicting the peak temperature of dissimilar AA7050-AA2014A friction stir welding butt joint using various regression models

Abstract

• Employing machine learning models, the temperature was predicted. • Both the DTR and RFR models achieved high prediction accuracy. • For temperature prediction in FSW processes, the RFR model should be prioritised. Friction stir welding (FSW) is a solid-state joining process that produces joints without melting and recasting. This process gains attention in the aerospace sectors for joining similar and dissimilar metals like the Aluminium alloy 7050 and 2014A. FSW provides very low peak temperature and heat distribution compared to the conventional joining techniques. In the FSW technique, the normal working temperatures range from 200 to 550 °C based on its process parameters. The aluminium alloys 7050 and 2014A are precipitant hardened alloys, and their precipitates dissolve at temperatures above 350 °C. It leads to a degradation of the mechanical properties of the weldment. Predicting peak temperature (PT) is the major phenomenon in getting good quality weld joints with respect to the FSW process parameters. This study deals with the prediction of peak temperature using machine learning (ML) approaches in various regression analysis methods like linear regression (LR), polynomial regression (PR), support vector regression (SVR), decision tree regression (DTR), and random forest regression (RFR). As a result, the RFR analysis is strongly suitable to predict the peak temperature in the FSW process. A tool rotation speed of 1000 rpm ensured the peak temperature of less than 300 °C with good appearance, proper material mixing, and the absence of defects.