Research Into The Fault Diagnosis Expert System Based On Neural Network

With the development of artificial intelligence, there are the diagnosis techniques in a new phase, which is intelligent diagnosis phase. Using the various methods in the fault diagnosis,especially,in connection with obscure diagnosis subjects, diagnosis features will be described beneficially. The inference processes fit in with the objective reality better. At the same time, the knowledge bottleneck problems of traditional fault diagnosis expert systems will be overcomed. Some kinds of AI methods integrated can improve the levels and efficiency of fault diagnosis greatly.In this thesis,the fault diagnosis problems of armored vehicles powertrain system are investigated using expert system and neural networks. On the base of the studying traditional expert system, a new expert system mode on the fault diagnose of gearshift box—the neural network expert system is put forward. it will conquer the limitation of the traditional expert system, for instance in not self-studying, self-adapting.In this thesis,A BP neural network expert system is also constructed. The network is trained using the sampled fault data, and the hybrid intelligent diagnosis to some fault is realized sequentially. During the training process of the fault samples, this training process can provide the scientific basis for the correctly chosen of the network parameter. The diagnosis results show that all of the fault samples can be convergent to the setting error value and the network has the generalizing ability.