An Explainable Machine Learning Framework for Railway Predictive Maintenance using Data Streams from the Metro Operator of Portugal
2025-08-08
Summary
The article discusses a machine learning framework designed to enhance railway predictive maintenance using real-time data streams from sensors in the Metro of Porto, Portugal. This framework includes a data processing pipeline that preprocesses data, classifies potential failures using machine learning models, and explains outcomes in natural language and visual formats. The system demonstrated high accuracy and F-measure scores, indicating its effectiveness in predicting and explaining potential failures in railway systems.
Why This Matters
Predictive maintenance is critical in the railway sector to anticipate failures before they occur, reducing downtime and operational costs while improving safety. The novel framework presented in this study addresses significant challenges in predictive maintenance, such as model interpretability and real-time data processing. By providing detailed explanations of predictions, the framework helps maintenance teams understand and act on potential issues promptly, enhancing service reliability and safety.
How You Can Use This Info
Professionals in the transportation sector can leverage this framework to improve maintenance operations and decision-making processes. By adopting such advanced predictive maintenance systems, operators can minimize unexpected service interruptions and maintenance costs. Additionally, the explainability feature aids in better understanding system behaviors, thereby facilitating more informed and quicker decisions in operational contexts.