Fault Feature Extraction Technology Of Rolling Bearing Based On Acoustic Emission

Roller bearing is most widely used and most prone to be damaged in mechanicalequipments. Whether the roller bearing can run normally is related to the whole equipmentsperformance directly. Fault diagnosis of roller bearing has great significance to safetyproduction and the machine running smoothly. Therefore, it is necessary to find the mosteffective method to the fault diagnosis of roller bearing.Vibration signal of early roller bearingfault is very weak and the impact energy is very low, the vibration signal is often submergedby strong background noise. Due to the acoustic emission signal has high frequencycharacteristics, it can eliminate the effect of low-frequency noise and acoustic emission sensormaterial is acceptable for the internal release of strain energy, can effectively detect earlyfaults of roller bearing, Therefore, in this paper, acoustic emission technology is used to faultdiagnosis of roller bearing.The N204model roller bearing was used in the technology research of roller bearingfault diagnosis, the roller bearing fault simulation test bench was set. The state monitor andfault diagnosing for outer ring fault, inner ring fault, roller element fault and outer ring rollerelement coupling fault, and collect the corresponding acoustic emission signal.This paper put forward an improved wavelet threshold denoising method, this methodcombines the advantages of traditional hard threshold, soft threshold denoising, And thesimulation signal analysis indicates that the improved wavelet threshold denoising methodwas superior to the traditional hard threshold, soft threshold, Combine this method toEmpirical Mode Decomposition method, and applied them to fault diagnosis of rollingbearing. Provide a new method for rolling bearing fault diagnosis.This paper combines markov distance and empirical mode decomposition, First of all,the acoustic emission signal was pretreated by wavelet denoising. Secondly, extract the largestcomponent by the means of markov distance and original acoustic emission signal correlationfor subsequent analysis, eliminate the smallest of the false correlation with original signal componen. Eliminate the smallest of the false components. Spectrum analysis was used forthe intrinsic mode components, and compared it with the original signal envelope spectrumanalysis directly. The effectiveness of this method was validated by the bearing acousticemission experiments.