Fault diagnosis of rolling bearing based on fuzzy neural network and chaos theory

摘要

Awareness of the importance to make system reliable has been raised from engineering practice, and fault diagnosis of rolling bearing must be taken seriously.Although numerous studies on fault diagnosis have been carried out, there are still a number of key technical issues.Uncertain problem is one of them.Fault diagnosis based on fuzzy neural network and chaos theory can solve uncertain problem essentially, moreover it is easy to understand because of it is based on human language, the system features is easy to maintain.Therefore it is an effective method to diagnosis complex system.The input nodes of fuzzy neural network is designed by using the minimum embedding dimension of phase space reconstruction, constructing the residual generator based on fuzzy neural network and chaos theory.We can effectively detect the signal which has chaotic and fuzzy property through a reasonable evaluation of the prediction error.And it is applied to the fault diagnosis of rolling bearing, to some extent, solving the problems of complex system modeling and fault feature extraction based on fuzzy theory.