Intelligent Fault Diagnosis Technology Research And Application

In recent years, diagnosis technology has already come into the new stage-intelligent fault diagnosis to satisfy diagnosis request in some complex system. This is a diagnosis method based on expert knowledge and artificial intelligence technology.In this paper, research on fault diagnosis is based on expert system and neural network .A method to built expert system with fault tree is presented for the cooling system of the high-temperature nitric acid. The method distills principle knowledge on the basis of fault tree, so that the building process of expert system is to be simplified and developed. The reliability of expert system is improved because of introducing the certainty of symptom and integrating the experience reasoning and principle reasoning.For large-scale complex diagnosis objects, there are many questions in applying single neural network to construct fault diagnosis system such as the problem of managing teacher data, lowering the precision of studying and affecting the efficiency of studying and so on. In this paper, hybrid fuzzy neural networks architecture is proposed. A series of fuzzy neural subnetworks are integrated in this hybrid fuzzy neural networks to perform the task of fault classification. Making use of the fault tree of the system to construct the subnetworks and building the hybrid neural network that includes several subnetworks to solve the questions of single neural network.In the process of fault diagnose, all kinds of values of symptom and fault should be fuzzy values. Fuzzy logic is imported in these two methods of fault diagnosis. The simulations show the feasibility and validity of these two methods.