Application of Modified Morphological Pattern Spectrum and LSSVM for Fault Diagnosis of Train Wheeltset Bearings

The diagnosis of faults in train wheelset bearings is crucial for railway infrastructure manager as it contributes to the safety of railway operations. This paper aims to develop a novel fault diagnosis method based on modified morphological pattern spectrum (MMPS) and least square support vector machine (LSSVM) to identify the different health conditions of wheelset bearings. The opening minus closing gradient is proposed to replace the erosion and opening operator to calculate the morphological pattern spectrum (MPS) in consideration of its advantage for fault feature extraction. The proposed method is experimentally demonstrated to be able to recognize the different fault types of wheelset bearings.