MOAS

A real-time knowledge-based operator advisory system (Maryland Operator Advisory System; MOAS) is being developed at the University of Maryland using PICON (real-time expert system shell for process control). MOAS is an expert system designed to assist the plant operators during transient conditions initiated by the Condensate and Feed Water System (CFWS) of a pressurized water reactor. The MOAS performs (1) intelligent prealarming and alarming, (2) real-time sensor validation and sensor conflict resolution, (3) real-time hardware failure diagnosis, and (4) real-time corrective measure synthesis. The core of the MOAS knowledge base is constructed using an integrated deep-knowledge approach called Coal Tree-Success Tree (GTST). The process domain knowledge of the CFWS is completely and rigorously modeled in the GTST format, which is then used to build up the knowledge base of the expert system. This paper briefly describes the GTST model, design of the current MOAS and ongoing and future efforts