A Comparative Study of Machine Learning Models for Predicting Single Bead Geometry of SS316L Depositions by GTAW Wire Arc Additive Manufacturing Process

Machine learning (ML) has recently gained popularity as a computational method in the manufacturing sector. The current study compared three ML techniques, namely support vector machine (SVM), artificial neural network (ANN), and adaptive neuro-fuzzy inference system (ANFIS), for predicting bead geometry in wire arc additive manufacturing (WAAM). Single beads of stainless-steel 316L (SS316L) material were deposited in the WAAM process using a gas tungsten arc welding (GTAW) machine. Statistical metrics including the mean square error (MSE), mean absolute error (MAE), coefficient of determination (R² value), index of merit (IM), and root mean square error (RMSE) were used to evaluate the effectiveness of ML models. The result revealed that the ANFIS model showed the best results, having minimum RMSE and IM values of 0.28 and 0.60, respectively, for the prediction of bead height, while for predicting bead width, RMSE and IM values were found to be 0.11 and 0.60, respectively.