Research On Fault Diagnosis Method Of Rolling Bearing Based On Time-frequency Analysis

The failure of rolling bearing usually has the characteristics of nonstationarity, complexity and early fault, thus it is difficult to excavate the fault information of vibration signal by using traditional signal processing technology and single fault diagnosis method. In this paper, the hybrid fault diagnosis method and its engineering applications for diagnoiseing rolling bearing are studied.Aiming at the nonstationarity of rollling bearing failure, this paper proposes the fault diagnosis model of rolling bearing based on wavelet packet decomposition, which is based on the advantages of wavelet packet decomposition in dealing with nonstationary signals. The model uses the wavelet packet decomposition to decompose the vibration signal into several independent sub-bands, and calculates the energy value and time domain characteristics of each frequency band respectively. As a fault feature matrix input clustering algorithm, the fault type is diagnosed. The model achieves satisfactory results in the identification of the fault type of rolling bearing.Aiming at the complexity of the rolling bearing failure and the weakness of the early fault,the paper combines the empirical empirical modal decomposition and permutation entropy, and proposes a rolling bearing fault diagnosis model based on the empirical mode decomposition and multi - scale permutation entropy. The model uses the empirical empirical mode decomposition to decompose the vibration signal into several intrinsic modal components, and obtain the permutation entropy of different components on different scales, and then use the feature set input clustering algorithm to diagnose the fault. The results of the example show that the model can accurately identify the fault type and the severity of the fault.A fuzzy C-means clustering algorithm based on principal component analysis (PCA) is proposed to solve the problem that the clustering number needs to be set in advance and the clustering accuracy is low. The algorithm uses the principal component analysis to reduce the dimension of the feature set, extract the sensitive features, eliminate the redundant features, and improve the clustering accuracy. The clustering index is used to select the optimal number of clusters. The results show that the clustering accuracy of the improved algorithm is better than that of the unmodified algorithm. The results show that the clustering accuracy of the improved algorithm is better than that of the unmodified algorithm.