WAS (Welding Analytics System): Addressing Automotive Welding Challenges through Machine Learning

This paper investigates the integration of predictive maintenance strategies to enhance spot welding quality and extend the service life of welding robots in automotive manufacturing. By leveraging AI-driven systems, real-time data from welding robots are analyzed to detect anomalies, while machine learning models classify welding spots. Predictive maintenance models are utilized to identify anomalies in welding robot health, facilitating proactive maintenance interventions. The study showcases how predictive maintenance improves quality standards, minimizes downtime, and optimizes operational efficiency in large-scale automotive manufacturing environments, achieving a remarkable sensitivity of 95%. This research contributes to advancing understanding in predictive maintenance strategies tailored to spot-welding operations, thus driving improvements in product quality and reliability within the automotive manufacturing sector.