FAULT DIAGNOSIS EXPERT SYSTEM OF MARINE DIESEL ENGINE VIA ANN

摘要

This paper describee the discrimination of the causes of faults in a marine diesel engine using expert systems. The expert system is composed of some artificial neural networks (ANN). Each neural network is a three-layer neural network of Perceptron type with a back propagation (BP) algorithm. The structure of the system and the methods of knowledge representation, knowledge acquisition aivi knowledge reasoning are discussed in this paper. And the optimal parameters of the BP model are selected by practical training of the models. An example of fault diagnosis of the system is given.