The Design And Application Of Aero-engine Failure Mode Analysis And Diagnosis System

The faults analysis and diagnosis of aero-engine is an important research subject in areo-engine health management area. With the development of modern aircraft engine, taking consideration into the complex non-linear relationship between faults and symptoms, it becomes more and more difficult to analyze and diagnose the faults. The article introduces the basic theory of fuzzy logic and artificial neural network, and combines them to enhance the advantages, then studies the intelligent faults diagnosis technology based on fuzzy neural network and the improved network, finally develops an aero-engine failure mode analysis and diagnosis system in order to implying the engineering application.To begin with, the article has introduced the theory and principle of aero-engine fault diagnosis, discussed the advantages and characteristics of diagnosis methods based on fuzzy logic and neural network, and combined fuzzy logic with ANN and learn from others strong points to offset one's weakness. Then, the article proposes a new intelligent fault diagnosis based on fuzzy neural network. The diagnosis theory and arithmetic are described in detail.And the model of aero-engine failure is set up by using measured data at vibration faults as training sample. The experimental results demonstrate that, compared with the traditional methods such as BP neural network and fuzzy logic, the fuzzy neural network proposed can not only effectively improve the accuracy of fault diagnosis, but also evaluate the possibility or severity of various of faults and make the diagnosis results more practical.Aero-engine failure mode analysis and diagnosis system has been designed and developed by combining with failure mode analysis method and fuzzy neural network algorithm. The system mainly includes failure mode analysis, failure data management and faults diagnosis which is the key part of whole system. And faults diagnosis is divided into two steps: the network model training and practical work processing. With plenty of training samples, the system based on well-trained fuzzy neural network has achieved the desired effect, and has been applied to the practical project of aero-engine faults diagnosis.Meanwhile, the research of the article is significantly valuable to further development of fault diagnosis technology and the aero-engine monitoring and faults diagnosis system.