Automated analysis of AGR reactor core condition monitoring data

This paper describes the design, development and deployment of the British Energy Trace Analysis (BETA) condition monitoring system that performs automated analyses on reactor refueling data to help support the continued safe operation of Advanced Gas Cooled Reactors (AGR) in the UK. The system performs automated analyses of a particular set of refueling data to detect the presence of damage in the graphite bricks, which compose the core. By detecting anomalous changes in the load of fuel during refueling, the system can identify likely cracks in the fuel channel bricks. Though the example presented here describes the development of a condition monitoring system for AGR stations, the methodology used to extract knowledge from condition monitoring analyses could potentially be adapted to other applications. Based on feedback from the plant operators, future versions of the software will include the storage of previous analyses and more advanced analyses in a centralized system that can be accessed remotely. Constructing a database of previous analyses will also allow the generation of envelopes of normal behavior, which will increase the ability of the system to detect anomalous behavior. Future work for the BETA system includes the application of data fusion techniques to combine the existing analyses with other reactor data sources to further enhance the ability of BETA to detect fuel channel distortions.