Fault Diagnosis Based On Tree Heuristic Feature Selection And FS-DFV For Rolling Element Bearings

In modern industry,rolling bearing is one of the most basic components of rotating machinery.it is widely used in aerospace,engineering manufacturing and other industries.the long-term operation of the machine will cause wear,peeling and other conditions on the rolling bearing part.the result will reduce the service life of the machine,even threaten the life of people,and will cause irreparable consequences.therefore,the accurate condition monitoring and fault diagnosis of rolling bearing play an important role in ensuring the reliable operation of the machine,and the effective diagnosis of it has become a research hotspot.based on empirical mode decomposition,dependent feature vector,fuzzy set theory and probabilistic neural network,this paper proposes a new method based on the analysis and processing of rolling bearing vibration signal,feature extraction,feature reduction and fault diagnosis the diagnosis method of rolling bearing based on tree search and fuzzy dependent feature vector.the main research work and innovations are as follows:For the non-linearity and non-stationary of bearing signal,empirical mode decomposition is introduced into vibration signal processing in this paper.it is a new time-frequency analysis method and an adaptive time-frequency localization analysis method:(1)Intrinsic Mode Function(IMF)is related to sampling frequency;(2)it is based on the change of data itself.This is where Empirical Mode Decomposition(EDM)is superior to Fourier transform,which gets rid of the limitations of Fourier transform.through the numerical simulation analysis,the validity of the method in dealing with vibration signal is verified.lay a solid foundation for the subsequent feature selection.Because the fault feature sample set extracted from the initial bearing signal by EMD is a high vit collection,there is a lot of redundant information in the high feature dimension,which leads to the calculation burden and complexity of fault diagnosis,and affects the effectiveness of diagnosis.In this paper,in order to make the extracted fault features more significant,the theory of dependent feature vector is introduced.the dependent feature vector can further mine the essential difference of fault,improve the accuracy of fault,and have obvious advantages in sample representation.in view of the uncertainty of the dependent feature vector in the representation of the overlapping mode,based on this,the fuzzy set theory is introduced,and fault diagnosis based on tree heuristic feature selection and FS-DFV for rolling element bearings is proposed.through the establishment of heuristic tree model,the design of tree heuristic feature search strategy,the selection of appropriate feature selection criteria,improve the conventional feature selection mode.In addition,the fuzzy set is used to deal with the problem of overlapping pattern in the extraction of dependent feature vector,and the fuzzy membership degree is used to guide the subsequent feature extraction of overlapping pattern.Based on the above research,the probability neural network is used to design the fault classifier,and the test results of rolling bearing fault diagnosis are compared on the same equipment.the data from the bearing data center of Case Western Reserve University in the United States are selected for data analysis.the experimental results show that the proposed method can effectively improve the diagnosis efficiency of rolling bearing.