Fault Diagnosis Of Rolling Acoustic Emission Signal Based On Information Entropy

Rolling is one of the most important parts of the machinery and equipment. Itsoperational status of a direct relationship to the entire system’s efficiency andproduction safety.Therefore,condition monitoring and fault diagnosis of rollingbearing has a wide range of economic and social benefits.This article is committed touse acoustic emission signals generated in the rolling bearing to detect the fault.First,this article provides an overview of the traditional bearing fault diagnosismethod. Typical vibration signal analysis are described in detail, and the authorpointed out its limitations, namely: the low-speed rolling bearing fault diagnosis, theextraction and denoising effect of the vibration signal is not very satisfactory,therefore, put forward a method for bearing fault diagnosis by the acoustic emissionsignal.Second, introducing the basic concepts of information entropy,the authordescribes the four kinds of information entropy of acoustic emission signals of rollingbearings in detail: singular spectrum entropy of the time domain,power spectralentropy of the frequency domain, characteristic entropy of the wavelet space andwavelet energy spectrum entropy of the wavelet space-frequency domain.Calculating the information entropy of the fault bearing, the author obtaine theinformation entropy band of each information entropy of bearing fault and initiallymake the fault diagnosis.Finally, based on the idea of information fusion, the author propose the conceptof integration information entropy distance as a method and basis to bearing faultdiagnosis. The distinguish to the bearing fault is effect because the integrationinformation entropy distance take into account the failure characteristics of thevarious symptom domains and verify the feasibility of the theory by simulation.