The Study Of The Fault Diagnosis Methods And Applications Based On Fuzzy Neural Network

This paper briefly introduced the current situation of the methods and applications of equipment diagnosing through modern intelligent technology and outlined the advantages by applying fuzzy neural network in fault diagnosis. Based upon the description of the principles of fuzzy theory and neural network, this paper analyzed the defects and merits of both technology and stated the importance of its integration. It also recommended a new integration method to establish a FNN model for fault diagnosing after a thoroughly research of the most popular combining modes in fuzzy theory and neural network. On the theory of fuzzy BP network, this model composed a fuzzy reasoning method which could realize information translating knowledge thorough distilling, prioritization and selection of fuzzy rules. This model also take an important role in both selecting knowledge quickly and improving diagnosis validity. through weights of FNN transferring into diagnosis guiding operator based on case-reasoning. Finally, I applied the model in a case study and reached a significant result, which helped to establish a fault diagnosis system of different ply temperature in freeze-dry machine in the goal of keeping the smooth operation of the equipments.