A comparative study of multiple regression analysis and back propagation neural network approaches on plain carbon steel in submerged-arc welding

Abstract

Weld bead plays an important role in determining the quality of welding particularly in high heat input processes. This research paper presents the development of multiple regression analysis (MRA) and artificial neural network (ANN) models to predict weld bead geometry and HAZ width in submerged arc welding process. Design of experiments is based on Taguchi’s L16 orthogonal array by varying wire feed rate, transverse speed and stick out to develop a multiple regression model, which has been checked for adequacy and significance. Also, ANN model was accomplished with the back propagation approach in MATLAB program to predict bead geometry and HAZ width. Finally, the results of two prediction models were compared and analyzed. It is found that the error related to the prediction of bead geometry and HAZ width is smaller in ANN than MRA.