Transformer Vibration Forecasting for Advancing Rail Safety and Maintenance 4.0
Authors:
Darío C. Larese,
Almudena Bravo Cerrada,
Gabriel Dambrosio Tomei,
Alejandro Guerrero-López,
Pablo M. Olmos,
María Jesús Gómez García
Abstract:
Maintaining railway axles is critical to preventing severe accidents and financial losses. The railway industry is increasingly interested in advanced condition monitoring techniques to enhance safety and efficiency, moving beyond traditional periodic inspections toward Maintenance 4.0.
This study introduces a robust Deep Autoregressive solution that integrates seamlessly with existing systems t…
▽ More
Maintaining railway axles is critical to preventing severe accidents and financial losses. The railway industry is increasingly interested in advanced condition monitoring techniques to enhance safety and efficiency, moving beyond traditional periodic inspections toward Maintenance 4.0.
This study introduces a robust Deep Autoregressive solution that integrates seamlessly with existing systems to avert mechanical failures. Our approach simulates and predicts vibration signals under various conditions and fault scenarios, improving dataset robustness for more effective detection systems. These systems can alert maintenance needs, preventing accidents preemptively. We use experimental vibration signals from accelerometers on train axles.
Our primary contributions include a transformer model, ShaftFormer, designed for processing time series data, and an alternative model incorporating spectral methods and enhanced observation models. Simulating vibration signals under diverse conditions mitigates the high cost of obtaining experimental signals for all scenarios. Given the non-stationary nature of railway vibration signals, influenced by speed and load changes, our models address these complexities, offering a powerful tool for predictive maintenance in the rail industry.
△ Less
Submitted 20 January, 2025;
originally announced January 2025.
