Machine Learning Implementation in Automotive Windshield Bending Furnace Failures Prediction

Abstract

Conveyor motors in tempering furnaces endure constant mechanical stress in windshield manufacturing, making them prone to failure. Predictive maintenance is crucial to prevent unplanned downtimes and reduce production costs. This work focuses on the establishment of a predictive model to identify failures in tempering furnace conveyor motors in the automotive glass manufacturing line. Seven machine-learning (ML) models were tested: Decision Tree (DT), Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting (GB), K-Means, and Naive Bayes (NB). Their performance is compared based on metrics such as accuracy, recall, F1-score, loss, and using also the confusion matrix analysis. Among the evaluated models, XGBoost demonstrated the highest performance in predicting failures. It achieved an accuracy of 98.43% on the validation set and 97.97% on the test set, with a precision of 97%, recall of 98%, and an F1-score of 97%. These results confirm the model’s effectiveness in accurately detecting faults in tempering furnace conveyor motors. Consequently, we find that XGBoost holds significant potential for integration into automated predictive maintenance systems, aiming to reduce unplanned downtime and prolong the operational lifespan of industrial equipment.