Fault Detection And Recognition Of Rolling Bearing Based On Adaptive Morphological Filtering

Rolling bearing is one of the most important mechanical components in rotating equipment,but it is easy to be worn,which will greatly affect the operation safety of the whole equipment.This paper mainly studies how to detect whether there is a fault in the running process of the rolling bearing and identify the fault type,it has important significance in safety maintenance of mechanical equipment.The experimental data is from the QPZZ-II rolling bearing fault simulation platform,through which the four operating conditions of the bearing,respectively,the normal,the rolling body fault,the outer ring and the inner fault,are simulated.The rolling bearing acoustic emission signal is collected by the SAEU2 S acoustic emission system.Firstly,the adaptive morphological filtering method is studied to reduce the noise of acoustic emission signals.Design a weighted cascaded morphological filter,using the open operation to suppress positive pulse peak noise of the acoustic emission,and using the closed operation to suppress negative pulse valley noise of the acoustic emission.Compared with the different denoising indexes,the correlation indexes between the noisy and the denoising signals are introduced as the denoising evaluation index of the measured acoustic emission signal,the paper achieves adaptive open and close operations' weight and completes the design of morphological filters.Simulation results from a series of acoustic emission signal show that this method can effectively remove impulse noise in the signal.The comparison of the SNR and the correlation index shows that both trend of them are consistent with the change of the weight,so as to proves that the correlation index can replace the SNR.It is able to be used to evaluate the effect of denoising even the noise intensity of signal is unknown.Secondly,in view of the poor performance of adaptive morphological filtering method in white noise suppression,combination of adaptive morphological filtering and Ensemble Empirical Mode Decomposition(EEMD)has been done for denoising.Using EEMD to decompose the signal processed by morphological filtering,the signal is decomposed into a finite number of intrinsic mode functions(IMF)which can reflect different vibration modesfrom high frequency to low frequency,so as to isolate the high frequency interference signal.The simulation and comparison of the simulated acoustic emission signals show that the combination method is feasible and effective.Finally,the measured acoustic emission signals of the rolling bearing fault simulation platform are processed by the combination method,and the frequency domain characteristics of different bearing fault condition are analyzed and compared to identify the fault type.The experimental results verify the effectiveness of the presented method in the field of bearing signal processing.