Abstract
The timely recognition and accurate classification of emergency situations is an important task for nuclear plant operators. The Reactor Emergency Action Level Monitor (REALM) expert system was developed to provide advice to operators during emergency situations. A technique for filtering plant sensor data was developed for REALM and is broadly applicable to power plant monitoring tasks. Innovative computer architecture and verification and validation methods were developed during a research program to implement REALM at Con Edison’s Indian Point 2 nuclear plant.
POS Data Analysis System based on Association Rule Analysis