Fault diagnosis method of rolling bearing based on EMD-Hilbert envelope spectrum and BPNN

Rolling bearings are an important component of rotating machinery, and accurate diagnosis of their faults is also very important. This paper proposes a rolling bearing fault diagnosis method combining empirical mode decomposition (EMD)-Hilbert envelope spectrum analysis and BP neural network (BPNN). First, EMD is used to decompose the intrinsic modal function (IMF) component containing the bearing fault feature information from the original vibration signal data of the rolling bearing, and then the IMF component is processed in combination with the Hilbert envelope analysis method to obtain a clear fault feature frequency. The processed fault characteristic frequency is subjected to dimensionality reduction by principal component analysis (PCA) to propose redundant data. Finally, the dimensionality-reduced feature data is input into BPNN to establish a fault diagnosis model. The results show that the rolling bearing fault diagnosis method based on EMD-Hilbert envelope spectrum analysis and BPNN can effectively identify different fault states of rolling bearings.