Fault Diagnosis Method of Complex Supply and Transportation Mechanism Based on EEMD Multi-scale Sample Entropy and PSO-SVM

Abstract

The complex supply and transportation mechanism has the characteristics of difficult feature extraction. However, traditional fault recognition methods have disadvantages such as poor noise robustness and low fault recognition accuracy. In order to improve the accuracy of recognition, through combining Ensemble Empirical Mode Decomposition (EEMD) with Multi-scale Sample Entropy (MSE), a fault diagnosis method of complex supply and transportation mechanism based on EEMD-MSE and Particle Swarm Optimization Support Vector Machine (PSO-SVM) is proposed. First, the method performs cross-correlation information analysis on a series of stable Intrinsic Mode Functions (IMF) of different time scales obtained by EEMD decomposition of the original vibration signal to reconstruct the signal; Then, the MSE of the reconstructed signal is calculated to construct a feature vector, which is divided into a training set and a test set; The penalty parameters and kernel function parameters are optimized through the PSO algorithm. The optimized SVM is used as a multi-fault classifier, and the training set is used as the input vector to train the vector machine to obtain the optimal parameters. The trained model is detected by using the test set to verify the accuracy of effective fault diagnosis. The bench test of the ammunition supply mechanism has proved that the proposed method can effectively identify its health status and failure type.