With explosion of production capacity and improvement of complexity of schedule, the production safety is playing a much more important role in chemical engineering enterprises. Real-time monitoring for production equipment, which facilitates diagnosing and finding troubles and potential dangers, has an important research value and practical meanings. This thesis prefers a kind of intelligence expert system for safe production which is a combination of neural network and expert inference, applying to fault diagnosis of equipment, prediction and disposal of accident in chemical engineering process. The thesis analyzes performance of the system in practical application in chemical enterprises, and also investigates training times and accuracy of neural network, and inference efficiency of expert system.
展开▼