Fault Diagnosis Of Rolling Bearing Based On Rough Set And Fuzzy Neural Network

Rolling bearing is widely used in mechanical equipment,its working state directly affects the production of mechanical equipment,so the fault diagnosis of rolling bearing has a certain practical significance.Rolling bearing can be diagnosed by temperature analysis or oil sample analysis.However,when the fault of rolling bearing is relatively slight,the temperature rise is not obvious,so it is difficult to diagnose.The oil sample analysis technology cannot diagnose the grease lubricated bearings,so the above diagnostic methods have some limitations.The vibration signal of rolling bearing reflects the working state of the rolling bearing,so the vibration signal analysis is the most practical method of bearing fault diagnosis.Rough set theory is an important method to deal with confusion and incomplete information in the field of artificial intelligence.It is widely used in the fields of data association mining,redundancy reduction and so on.The fuzzy neural network sets the advantages of neural network and fuzzy system,giving full play to the processing ability of the model to the uncertainty of the system,and the parameters of the model can be adjusted and optimized by self-learning.These two theoretical methods can be combined with the vibration signal analysis method for fault diagnosis of rolling bearings.In this paper,we extract the feature vectors from the vibration signals collected from the rolling bearing fault test rig,and study two methods for bearing fault diagnosis.First,based on the theory of rough set theory,the attribute vector of the extracted vibration signal is reduced by the rough set analysis tool Rosetta software,the rough set classifier is established,the fault diagnosis is realized by self-learning.Second,the rough set method is fused with the fuzzy neural network,in which the rough set method is only.As a tool for eigenvector reduction,the eigenvector after reduction is used as the input of fuzzy neural network,and the adaptive fuzzy neural network system is trained for the diagnosis of bearing fault.The results show that the diagnosis method is effective by combining rough set dimension reduction and adaptive fuzzy neural network.Finally,combining the different advantages of Lab VIEW and MATLAB,a fault diagnosis system is developed,which realizes the bearing fault diagnosis.