DEVELOPMENT OF WATER CHEMISTRY DIAGNOSTIC SYSTEM FOR BWRS USING FUZZY-REASONING

A diagnostic system for BWR water chemistry has been developed incorporating high-speed fuzzy reasoning, holding in view preventive maintenance as principal objective. The system combines two expert systems-one for detecting an anomaly in the plant operating condition and quickly inferring its cause in a preliminary analysis, and the other performing a detailed diagnosis of the detected anomaly. The system for anomaly detection and quick preliminary diagnosis infers the occurrence of various anomalies and their causes from the fuzzy relation existing between the fuzzy causes and symptoms, applying a high-speed method for solving the inverse problem of fuzzy correspondence. The system can be run on a small size computer, and is thus conveniently adapted to practical application at machine side. The system for detailed diagnosis incorporates the expertise of water chemistry engineers, and outputs pertinent information and data to guide operators. Trial application of the system using data from actual plant operation conclusively demonstrated its effectiveness as a practical preventive maintenance system for actual plants.