Prediction of tensile behavior of FS welded AA7039 using machine learning

Abstract

The present study deals with the application of machine learning approaches such as Gaussian process regression (GPR), support vector machine (SVM), artificial neural network (ANN), and linear regression (LR) in analyzing and predicting the tensile behavior of friction stir welded AA7039. For this, FSW experiments are conducted at different values of rotational speed, welding speed, and tilt angle whereas the ultimate tensile strength (UTS) is considered as an output parameter. Besides this, the coefficient of correlation (CC) and root mean square error (RMSE) are considered as performance measurement parameters. It is observed that for a combination of process parameters, the ANN model is most useful for predicting the tensile behavior of FS welded AA7039 with minimum values of RMSE and maximum value of CC for test data set. Conversely, the LR model is found insignificant with maximum RMSE for test data set. In addition, electron backscattered diffraction (EBSD) is used to determined the grain size of base metal and nugget zone (NZ). The minimum grain size is obtained for higher strength specimen i.e. 4 microns. © 2020 Elsevier Ltd