Application Study Of Empirical Wavelet Transform And Support Vector Machine In Fault Diagnosis Of Rolling Bearings

Rolling bearings are the core of rotating machinery.Due to their harsh working intensity and poor working condition,they are also the most vulnerable components.Therefore,identifying and judging the types of rolling bearings in a timely manner can effectively ensure the operation safety of the mechanical system,which is of great significance.This article takes the rolling bearing as the research object,combines the improved empirical wavelet transform method and the support vector machine,and applies them to the fault diagnosis of the rolling bearing.The contents of this article as follows:Firstly,the status quo of fault diagnosis at home and abroad and the processing method of vibration signal have been introduced and presented in this article.The failure mechanism of rolling bearing is analyzed,which prepared for theoretically deducing and analyzing the characteristic frequency of the fault and the vibration characteristics of the typical fault.Secondly,empirical wavelet transform(EWT)theory is introduced because it can't accurately segment the Fourier spectrum.Therefore,a method called Energy Scale Space Empirical Wavelet Transform(ESEWT)is proposed.And this method is applied to bearing fault diagnosis.Mainly,using the method of scale space,the Fourier spectrum is adaptively divided to obtain the demarcation point of each band.Then,then the band demarcation point is selected according to the energy level of each band;so that the band with energy greater than the mean is retained,and the adjacent band less than the mean value is merged.After getting the effective band demarcation point,the filter bank is designed and the component signal is obtained.Finally,the Hilbert transform is performed on each component signal to extract the fault feature frequency of the bearing.Through experimental verification,the ESEWT method can reduce the frequency band demarcation point and improve the frequency band breakup to some extent.At the same time,ESEWT method can not only accurately extract the bearing fault characteristic frequency,highlight the fault frequency and its harmonic components,but also can effectively identify the rolling bearing fault.Finally,the statistical theory and Support Vector Machine(SVM)are introduced.The fault diagnosis method of rolling bearing based on ESEWT and SVM is proposed.This method performs ESEWT decomposition of the vibration signal,and extracts the kurtosis,crest factor,pulse factor,margin factor of the vibration signal and the singular value entropy,energy entropy,permutation entropy,and sample entropy of the component signal after the ESEWT decomposition as the eigenvectors.Vectors above are input into the SVM model for training to obtain a classification model,and the test samples are classified and identified.The experimental verification shows that the method has high classification accuracy and proves its reliability and practicality in practical engineering applications.