Research On Fault Diagnosis Of Mine Ventilator Bearing Based On EMD And SVM

National "13th Five-Year Plan" emphasizes that coal mines should ensure safe and efficient production.The National Security Administration also requires: coal mine gas control and ventilation management should be strictly implemented.The coal mine ventilation is the key equipment in ventilation system,it is also responsible for the safety of underground mine safety production and personnel safety.Therefore,the investigation about detecting the state of the coal ventilator and fault diagnosis is of important significance.This paper takes the rolling bearing as the research object,the current analysis method of vibration signal and the development status of fault diagnosis are summarized.According to the structure characteristics of coal mine ventilator bearing,the failure mode,fault mechanism and fault type are analyzed.Paper introduces and analyses the method of Empirical Mode Decomposition(EMD).By analyzing,the EMD is demonstrated to have superior performance in the non-stationary signal processing than the traditional time-frequency analysis methods(such as: Fourier Transform,Wigner-Ville distribution and Wavelet Transform).However,EMD method also has some shortcomings such as endpoint effect and mode aliasing.The endpoint effect and modal aliasing of EMD are improved in this paper,respectively,and the termination frequency-based improved EEMD approach is introduced,such that the method has advantages over the EEMD approach in the aspect of modal aliasing,false component and running speed.The improved EEMD method is applied to analysis signals under different status of the rolling bearing,the energy entropy is extracted as fault feature.In order to make the fault feature data more concise and effective.the Principal Component Analysis is used for secondary processing of the fault feature.Finally,the Support Vector Machine(SVM)is used to diagnose the fault of the ventilator bearing.As for the problem of SVM parameters is difficult to determine,the Particle Swarm Optimization(PSO)algorithm is used to optimize parameters of the Support Vector Machine(SVM).In addition for the existing problems of PSO,the multiple methods fusion approach is adopted to improve the PSO.Through the analysis of the bearing data,the results illustrate that the fault diagnosis model used in this paper has good performance.