It is difficult to use a single diagnosis theory or method to monitor and diagnose the whole condition of a large complicated system. To realize real-time condition monitoring and fault diagnosis of a large automatic production line in a steel factory, according to the structure and fault diagnosis characteristics of the production line system, a hybrid fault diagnosis expert system based on knowledge and neural network has been researched and built. This paper introduces the basic composition of the hybrid expert system, and gives the rule examples of the expert system based on knowledge, and emphatically introduces knowledge expression and knowledge acquisition and fault diagnosis inference of the fault diagnosis expert system based on neural network. Finally, fault diagnosis examples based on neural network are given. Research results show that a hybrid expert system is very effective to monitor and diagnose a large complicated modern production process.
展开▼