Fault Diagnosis of Train Rotating Parts Based on Multi-objective VMD Optimization and Ensemble Learning

Abstract Rotating machinery is widely used in various systems of trains, and its health status is directly related to the reliability of train operation. Therefore, the fault diagnosis of train rotating machinery plays a promoting role in the safe operation and maintenance of trains. Due to the influence of track excitation and other equipment vibration, the vibration signal of rotating machinery is non-stationary and nonlinear, which brings challenges to the fault diagnosis of rotating machinery. To solve this problem, an intelligent fault diagnosis method for rotating machinery based on multi-objective variational mode decomposition (VMD) optimization and ensemble learning is proposed. Firstly, an improved multi-verse optimizer (MVO) algorithm is proposed by introducing Tent chaos theory and Levy flight theory, and a multi-objective VMD parameter optimization model is established. The improved MVO algorithm is utilized to optimize the parameters and reconstruct the signal. Then, a multi-domain feature extraction method is proposed for multi-domain fault feature extraction and feature selection. Finally, an adaptive mutation particle swarm optimization random forest (AMPSO-RF) pattern recognition algorithm is proposed for fault pattern recognition. Through the verification of bearing and gear fault data, the results showed that the raised method can accurately realize fault diagnosis of train rotating parts, and the accuracy rate reaches 100%, which is better than the compared methods.