Fault Feature Extraction based on Cyclic Correntropy for Train Axle Bearings in Impulsive Noise

Train axle bearings are important components to support train operation. Fault diagnosis of train axle bearings is a necessary tool to prevent any unexpected accidents and ensure operation efficiency. However, impulsive background noise in railway fields brings a challenge to the fault diagnosis of train axle bearings. Therefore, this paper presents an intelligent fault identification method based on the cyclic correntropy spectrum (CCES) and least square support vector machine (LSSVM). To suppress the impulsive noise, the cyclic correntropy (CCE) function is introduced and the CCES is obtained subsequently. Then, cyclic frequency domain kurtosis indexes are extracted from the CCES. Finally, these extracted features are used to train the LSSVM for the fault pattern identification. The application of the proposed method is validated on train axle bearing data that are contaminated by impulsive noise.