Design of fault diagnosis system for polymerizer process based on neural network expert system

Based on the three-tier B/S network architecture, the remote monitoring and fault diagnosis system for polymerizer process is established. Then BP neural network expert system is adopted to diagnosis faults combined with specific polymerizer fault conditions and uses the obtained expert rules to train the parameters of BP neural network. With the polymerization reactor data processing system, a specific fault type is determined based the proposed strategy according to the fault information. Also diagnoses results and the related information are displayed in the browsers of the remote clients. The key technical used for the realization of remote monitoring and diagnosis include the usage of the dynamic JSP pages and Java tools access to the databases to realize the real-time on-line monitor on all operating parameters of the polymerizer equipment.