High-Speed Train Axle Bearing Fault Detection Using Signal Decomposition and Weighted Statistical Features: Application to Test Bench Bearings

The evolution of transportation has Considerably improved the efficiency of logistics and travel, with high-speed trains (HSTs) emerging as a transformative solution for rapid, long-distance mobility. Axle bearings are critical components of HSTs that require regular maintenance to prevent accidents, derailments, and other operational failures. This study utilizes signal decomposition techniques, particularly Empirical Mode Decomposition (EMD), on collected signals from the Complex Systems and Interactions (CSI) laboratory test rig at Ecole Centrale Casablanca. Relevant IMFs are identified by calculating weighted statistical features, and their validity is further assessed through cross-correlation and the energy-to-signal ratio. This methodology enables the accurate identification and analysis of various types of bearing faults, representing a crucial step toward enhancing the safety and reliability of HSTs.