Research on wheel out-of-round fault diagnosis based on vibration data images

Abstract

The wheel out-of-round fault of urban rail vehicles has a very important impact on the safe operation of urban rail trains. Therefore, it is of great significance to achieve an accurate diagnosis of the wheel out-of-round fault of trains. The purpose of this paper is to summarize the diagnosis methods of the wheel out-of-round fault, and propose a new diagnosis method based on vibration data images, which can effectively identify the wheel out-of-round fault. The one-dimensional vibration signal is converted into a two-dimensional texture matrix. The Statistical Geometrical Features (SGF) method extracts the feature information of the two-dimensional gray image and combines it with a support vector machine for training and recognition to achieve the fault diagnosis of the wheel out-of-roundness. The feasibility and accuracy of the method are verified by simulation and experimental signal analysis, respectively. The experimental results show that the overall recognition accuracy of the model simulation data and the two-wheel experimental bench data exceeds 91%, exhibiting significantly high fault identification accuracy. In this paper, a wheel out-of-round fault diagnosis model based on vibration data images has been established by analyzing the vertical dynamic signal of the axle box, which has the advantages of fast recognition in combination with two-dimensional grey-scale images, no signal pre-processing, and high recognition accuracy. It provides a new method for monitoring and diagnosing wheel out-of-round faults in urban rail vehicles.