Research On Feature Extraction And Recognition Of Weak Faults Of Rolling Bearings Based On Improved ENEMD

Rolling bearings are one of the most widely used basic components in rotating machinery and equipment.Because they are often in harsh working environments,they cause frequent failures,which affects the operation of the entire equipment,which in turn causes the equipment's work efficiency to decline or even stop production.It will cause certain economic loss,and it will cause catastrophic accidents,causing casualties and serious social harm.Therefore,the research on rolling bearing fault diagnosis technology,especially the early weak fault diagnosis,has important scientific and theoretical significance and engineering application value.However,in actual engineering,the characteristics of the early weak faults of rolling bearings are relatively weak.In addition to the influence of many factors such as the environment,it is necessary to obtain a strong noise pollution signal with extremely low signal-to-noise ratio in fault diagnosis.Come with some difficulty.At present,EMD(Empirical Mode Decomposition)and EEMD(Ensemble Empirical Mode Decomposition)as an adaptive time-frequency analysis method is often used to diagnose weak faults in bearings.However,due to modal aliasing and excessive decomposition in EMD,it is difficult to extract weak fault features and weak fault diagnosis is not ideal.Aiming at the above problems,this paper studies from two aspects of feature extraction and fault identification.The main research work is as follows:(1)Aiming at the problems of modal aliasing and residual noise caused by the EMD and EEMD methods,the early weak feature extraction effect of rolling bearings is not ideal.Based on the EMD theory,a method based on the combination of ENEMD(Ensemble Noise-reconstructed Empirical Mode Decomposition)and Teager energy operators is proposed bearing weak fault feature extraction method uses the resampling of the inherent noise component in the original signal to change the local extreme point characteristics of the signal,thereby improving the modal aliasing phenomenon.At the same time,it replaces the artificially added white noise in the EEMD by the inherent noise component.The phenomenon of noise residue is avoided,and the early weak fault characteristics of rolling bearings are better extracted.(2)Aiming at the problem of excessive decomposition in ENEMD,which makes modal selection difficult and incomplete extraction of effective information,which leads to difficulty in extracting weak bearing features of rolling bearings,a method for enhancing weak bearing feature of rolling bearings based on improved ENEMD is proposed.By introducing the idea of ??clustering,the decomposed IMF(Intrinsic Mode Function)components are combined into an optimal number of CMF(Combined Mode Functions),which overcomes the problem of excessive decomposition of fault signals in ENEMD and realizes the autonomy of decomposition modes.Selection reduces the number of modes,improves the integrity of effective information,and enhances the characteristics of weak bearings.(3)Automatic identify the type of weak bearing failure.The sensitive CMF components with improved ENEMD decomposition are screened out,the approximate entropy is calculated as the feature vector of the signal,the feature vector matrix is??constructed,and it is input to the fuzzy C-mean classifier for fault identification.In this paper,aiming at the problem that the early weak fault features of bearings are not obvious and easy to be submerged in strong background noise,which makes the extraction of weak fault features difficult or not ideal,we trace back to the source.Through the improved enemd,we effectively overcome the phenomenon of modal aliasing and noise residue in EMD and EEMD,solve the problem of excessive decomposition in enemd,enhance the weak fault features,and then make the weak fault characteristics of bearings The feature extraction is more ideal.Finally,the fuzzy c-means classifier is used to realize the automatic recognition of weak faults,and the clustering effect is more excellent.Through the combination of theoretical analysis and engineering data verification,the effectiveness and superiority of the proposed method are proved,which has certain engineering practical significance.