Research On Fault Diagnosis Method Of Rolling Bearing Based On Improved Empirical Wavelet Transform

Rolling bearing is one of the most widely used key components in rotating machinery.Because it often works in the harsh environment of high speed,high temperature and heavy load,it also makes it the most easily damaged part in mechanical equipment.To carry out real-time condition monitoring and fault diagnosis of rolling bearing is of great significance to avoid safety accidents and improve the safety of mechanical equipment.The early fault characteristic information of rolling bearings is weak,and it is easily affected by strong interference noise.It is difficult to diagnose and identify them using common fault diagnosis methods.In this paper,an early weak fault diagnosis method for rolling bearing based on improved empirical wavelet transform is proposed.The main research contents are as follows:(1)Aiming at the problem of redundant components in spectrum division of traditional empirical wavelet transform,an improved empirical wavelet transform method based on mutual information is proposed,which can eliminate redundant cut-off points,repartition the spectrum and obtain more valuable components.The results show that the improved empirical wavelet transform can effectively reduce redundant components and improve the accuracy of signal decomposition.(2)Combining improved empirical wavelet transform with minimum entropy deconvolution for rolling bearing fault diagnosis.After improving the components obtained from the empirical wavelet transform,the best component selection criterion based on mutual information and kurtosis is used to optimize the component containing the most abundant fault information for reconstruction.The minimum entropy deconvolution effectively reduces the influence of interference noise.Research shows that this method can effectively diagnose early weak faults of the wheel on the bearing and has certain practical application value.Compared with empirical mode decomposition,it has a more obvious diagnostic effect.(3)The improved empirical wavelet transform and support vector machineoptimized by the quantum genetic algorithm are used for rolling bearing fault diagnosis.The bearing data of Case Western Reserve University and high-speed trains were used for verification.Research shows that the support vector machine rolling bearing fault diagnosis method based on improved empirical wavelet transform and quantum genetic algorithm can effectively identify bearing types with different health states and fault sizes,which is more effective than traditional empirical wavelet transform and variational modal decomposition methods,and has a better classification and identification effect than the traditional empirical wavelet transform and variational mode decomposition method.