Machine learning tool for the prediction of electrode wear effect on the quality of resistance spot welds

Abstract The quality of resistance spot welding (RSW) joints is strongly affected by the electrode condition. Therefore, the capacity to automatically assess the effect of electrode deterioration on weld quality is crucial to ensure welds compliant with manufacturer quality requirements. This work develops a machine learning-based tool to automatically assess the influence of electrode wear on the quality of RSW welds. In this way, manufacturers can reduce the number of redundant spot welds due to electrode wear and increase competitiveness. Two different experimental campaigns were performed to evaluate the effect of electrode wear on the mechanical strength of spot welds. The resulting failure load of the joints have been used to build the weld quality classes of the machine learning tool, whereas data from electrode displacement and electrode force sensors, embedded in the welding machine, have been processed to identify the predictors of the tool. Five machine learning algorithms have been applied with a classification approach to predict whether the spot weld complies with the welding quality requirements. As a result, the artificial neural network is the most performing model with an accuracy of about 90%. This work paves the way for new maintenance strategies based on the real-time predictions provided by the machine learning algorithms and contributes to making the car-assembly process leaner and more sustainable.