Prediction of Failure in Scania Truck Due to Air Pressure System Failure

This paper addresses the prediction of failures in the air pressure system of Scania trucks to minimize associated operating costs. A custom ensemble model is proposed combining random forest, XGBoost, and multi-layer perceptron algorithms. By optimizing the classification threshold, false positives and false negatives are reduced, effectively minimizing costs. In comparison to previous studies, the model’s performance is evaluated using metrics like accuracy, AUC, precision, recall, etc., and it demonstrates superior performance across all these measures. This research significantly contributes to predictive maintenance in the automotive industry by offering valuable insights for effective failure management, cost reduction, and enhanced operational efficiency.