Rolling Bearing Fault Diagnosis Based on EMD-WVD Time-frequency Images and NRBO-SVM

To address the problem that the traditional time-frequency analysis method fails to capture the vibration signal dynamic changes in vibration signals of the rolling bearing with non-smooth, non-linear characteristics and can not accurately describe the fault characteristics, a rolling bearing fault diagnosis method based on EMD-WVD time-frequency images and NRBO-SVM is proposed. Firstly, the rolling bearing vibration signal is decomposed by employing the Empirical Mode Decomposition (EMD) method to obtain multiple Intrinsic Mode Functions (IMF) components with different frequencies to adapt to the non-stationarity of the signal. Then, the first few high-frequency IMF components are analyzed by Wigner-Ville distribution (WVD), to obtain multi-scale time-frequency images,which reveals the nonlinear characteristics of the signals with high resolution. Afterwards, the recognition of nonlinear features is enhanced by Local Binary Pattern (LBP) feature extraction. Finally, the Support Vector Machine (SVM) optimized by the Newton-Raphson-based optimizer (NRBO) is brought in for fault feature classification. The results show that the method improves the accuracy of fault diagnosis from 83.33% to 96.66%, and reduces the classification processing time from 78 s to 53 s.