Fault feature analysis of high-speed train suspension system based on multivariate multi-scale sample entropy

Abstract

In monitoring high-speed train suspension system working state, this paper proposes fault feature extraction method based on multivariate multi-scale sample entropy (MMSE) due to high-speed train's characteristics of large number of freedom of motion and strong correlation between different monitored data points. After using multivariate empirical mode decomposition (MEMD) in different working conditions of multi-channel synchronous conjoint analysis of vibration signals, access to the common pattern between different data channels. Choose the main intrinsic mode functions (IMFs) which can reflect the fault feature to reconstruct the original fault signal, and calculate the multivariate multi-scale sample entropy of the reconstructed signal as the fault feature. Finally, the support vector machine (SVM) is used to identify the fault state classification. Various experimental results show that the recognition rate can reach more than 90% of the classification results at various speeds, verifying the effectiveness of the proposed method.