Incipient Fault Diagnosis Method of Railway Vehicle Door System Based on Random Forest

Reliability of the Rail Vehicle Door System (RVDS) has strong negative impact on the safe operation of rail vehicles. In order to diagnose the Incipient fault state of the RVDS more accurately, a incipient fault diagnosis method based on feature assessment and Random Forest (RF) is proposed in this paper. Firstly, the time-frequency indexes are extracted from door's position, motor's speed and current as the eigenvector to characterize the health status of the door system. Then a feature evaluation method is applied to calculate evaluation factors of all features, and the sensitive ones are selected as the input of the RF according to the size of the evaluation factors to realize the automatic recognition of different incipient fault states of the door. The simulation results show that this method combines feature extraction, feature evaluation and Random Forest enables to precisely extract incipient fault information and improve the classification accuracy, which has important research significance in the field of incipient fault diagnosis of rail vehicles.