Rolling bearing fault diagnosis based on HVD algorithm and sample entropy

Abstract

Currently, the methods for rolling bearing fault diagnosis using acquired signals still have certain deficiencies, such as mode mixing during signal decomposition and selection of the optimum fault feature. To address these problems, this paper used a method based on Hilbert vibration decomposition (HVD) and sample entropy to perform bearing fault diagnosis. This method firstly decomposed the acquired original signals of faulty bearings used the HVD algorithm, then extracted fault features from the processed signals, found the frequency and multiplied frequency of the bearing fault by substituting some experimental parameters and bearing parameters into the calculation formula. The results show that the method proposed in this paper can reduce other interfering signals during signal processing and achieve a fault diagnosis rate significantly higher than that of the original EMD algorithm.