Research On Fault Diagnosis Of Motor Bearing Based On MEEMD

As for motor failures,there are about 40% failures caused by bearing among the overall failures,therefore,effective judgment on the failure of motor bearing has important theoretical research meanings and practical application values.There are more illusive component because the mode mixing of the EMD algorithm is easy to be appeared,and the incomplete decomposition and the residual noise caused by the EEMD algorithm introduced into noise.Therefore,this paper adopts MEEMD(modified ensemble empirical mode decomposition)algorithm to keep the advantages of above algorithms,which can well control the effect of the residual noise on the decomposition accuracy and guarantee the effectiveness of the decomposition results.What's more,the EMD(EEMD)decomposition of the cubic spline interpolation can be improved into MEEMD decomposition for the accuracy of the decomposition.The fast Spectral kurtosis figure is got by Spectral kurtosis algorithm can automatically provide the best center frequency and bandwidth for band-pass filter and overcome the disadvantage of judging the center frequency by the subjective experience.This paper makes effective judgments on the failures of motor bearing based on the combination of modified ensemble empirical mode decomposition(MEED)and the envelope detection of the spectral kurtosis.There are following main contents:First,from the vibration signal,this paper states the diagnostic mechanism on the bearing failure of the asynchronous motor and calculation on the fault character frequency of the antifriction bearing in theory,thus can provide a theoretical basis for showing the subsequent experiment results.Then,the collected motor vibration signal is decomposed by MEEMD to get some IMF components.The effective components are chose by analyzing the correlation coefficient method of the kurtosis to refactor the decomposition signal,thus can satisfy the noise reduction.The signal reconstruction is made spectral kurtosis analysis can get a set of best parameters for band pass filter,and the filter is used for the filtering of the signal reconstruction.The signal is made square envelope after filtering,and then it can get square envelope spectrum by the Fourier transform,the fault character frequency can be collected by the envelope spectrum analysis to diagnose the failure of motor bearing locations.Finally,the experimental platform for the failure diagnosis of the motor bearing is built to collect the data,which can verify the effectiveness and advantage of the experimental method.According to the collected vibration signal of the motor,the envelope detection combined by EEMD decomposition and spectral kurtosis is made comparative analysis with the envelope detection combined by MEEMD decomposition and spectral kurtosis,thus can draw a conclusion at last.The envelope detection combined by MEEMD decomposition and spectral kurtosis has better fault diagnosis accuracy and effect than that of the envelope detection combined by EEMD decomposition and spectral kurtosis.