Journal : Journal of Theoretical and Applied Information Technology, 17 mars 2024
The prevention of recurring failures in modern manufacturing systems is of paramount importance for minimizing costs and downtime. Despite the potential for real-time data analysis using sensors offered by Industry 4.0 technologies, their widespread adoption, particularly among smaller manufacturing systems, remains a challenge. In response, this paper introduces an alternative approach to predictive maintenance planning, utilizing historical maintenance intervention data in the absence of sensor-based information. The study investigates the pivotal role of Artificial Intelligence (AI), specifically Machine Learning (ML) and Prognostics and Health Management (PHM), in augmenting the efficiency of predictive maintenance. Utilizing a comprehensive dataset from Schleuniger cutting machines spanning June 2021 to June 2023, our research evaluates two predictive maintenance approaches: Precision-Based Maintenance Prediction (PBMP) and Occurrence-Driven Maintenance Prediction (ODMP). The objective is to extract valuable insights from historical maintenance data, enabling proactive decision-making and preventing future failures. The deployment results presented in this study demonstrate the effectiveness of predicting the number of failures, providing valuable information that can enhance maintenance planning and reduce total downtime. By addressing the practical challenges faced by smaller companies in adopting sensor technologies, this research contributes valuable insights to the broader landscape of predictive maintenance in manufacturing.