Anomaly detection of high-speed train braking system based on wireless sensors network
• Temperature of brake discs in High-Speed trains represents a crucial parameter for maintenance and safety aspects. • Brake discs temperature monitoring is actually performed wayside and not on-board train, limiting the information availability. • A new wireless sensor system was developed for retrofitting of in-service trains to monitor brake temperatures for condition monitoring. • The custom sensors are energy-independent and operate as stand-alone units, transmitting data via Bluetooth Low Energy (BLE) to a Central Control Unit. • An anomaly detection algorithm, relying on a data fusion strategy with multiple measurement types (e.g., temperature, acceleration, pressure) provides real-time malfunction alerts. To date, high-speed trains are equipped with a dense network of sensors, which allows for monitoring many operational parameters of the vehicle. The effort made to develop such monitoring and diagnostic systems aims to continuously increase the level of safety, thus ensuring timely intervention in case of anomaly of one or more components. However, some features are still not directly monitored on board but are tracked by systems installed on the infrastructure. One of these features is represented by the detection of brakes temperature, which is significant for condition monitoring and for guaranteeing that the vehicle travels safely. In this work, a custom wireless sensors system was developed to be installed on in-service trains, with embedded temperature sensors for detecting anomalies in the brake system. These sensors are energy-independent and thus completely stand-alone. The data collected by each sensor is then transmitted via Bluetooth Low Energy (BLE) to a Central Control Unit, in which an anomaly detection algorithm has been implemented. This algorithm employs a data fusion strategy, relying on different types of measurements (temperature, acceleration, pressure, etc.) and merging the data of all sensor nodes, to provide real-time alarm notification in case of malfunctioning of the brakes.
Authors
- Nicola Giulietti (ORCID: https://orcid.org/0000-0001-9922-3201)
- Federico Zanelli (ORCID: https://orcid.org/0000-0001-9720-7793)
- Nicola Debattisti (ORCID: https://orcid.org/0000-0002-3759-7563)
- Marco Mauri (ORCID: https://orcid.org/0000-0002-3996-1904)
- Francesco Castelli Dezza (ORCID: https://orcid.org/0000-0003-1567-0077)
Institutions
- University of Pavia (IT)
- Politecnico di Milano (IT)
Publication Details
- Journal
- Measurement
- Published
- 2025-05-05
- DOI
- https://doi.org/10.1016/j.measurement.2025.117774
- Citations
- 1
- Primary Topic
- Traffic Prediction and Management Techniques
- Type
- article
- Field-Weighted Citation Impact
- 1.01