Research On Bearing Fault Classification Based On Attribute Partial-ordered Srtucture Diagram

Rolling bearings as the key components of rotating machine,are widely used in manufacturing,energy,defense and other important equipment in various fields,while rolling bearings are also the most likely damaged as the components of rotating machine,Related research shows that nearly one-third of the rotating machinery system fault caused by bearing,It has a great significance to analyze and diagnose the running status of rolling bearing in real time.Therefore,Research in this field has always been the focus of research at home and abroad.In this paper,In view of the shortcomings of traditional diagnosis methods in recognition accuracy and stability,Feature extraction and fault classification as the starting point of fault diagnosis of rolling bearing,A classification method based on attribute partial ordered structure theory is proposed.The main research work of this paper is as follows:The data source of this paper is the intelligent instruments and monitoring diagnosis of Mechanical Engineering Institute of Xi'an Jiao Tong University,which has been used by many researchers working on bearing fault research.Based on the formal background of formal concept analysis theory,this paper uses the method of equal width discretization to obtain the background,supplemented by wavelet analysis and principal component analysis as the method of data processing and feature extraction,and the attribute partial order structure classification method is the core.Through the structure of bearing fault data attribute partial ordered structure diagram,to achieve bearing fault,including rolling failure,inner ring failure,outer ring fault,and the normal operation of the bearing classification of data.As we all know,Vibration characteristic is the most fundamental expression of mechanical equipment running state,this paper carries on the data processing of bearing vibration signal under different fault condition during operation,to generate the properties of partial order structure diagram as the basis,from the cluster structure in the partial order relation attribute summed up the fault of the same type,and distinguish between different types of faults from branch in the model,The distance between nodes in the partialdiagram is proportional to the similarity of the objects,So we can obtain properties of the object from the relationship between nodes,,with the attributes of objects of different types of fault classification of bearing fault in different realization.Through the experimental classification to achieve a higher accuracy,the establishment of stable,reliable and practical requirements of the classification model for rolling bearing fault diagnosis provides a new way of thinking and new methods.