Combining models of behaviour with operational data to provide enhanced condition monitoring of AGR cores

Installation of new monitoring equipment in Nuclear Power Plants (NPPs) is often difficult and expensive and therefore maximizing the information that can be extracted from existing monitoring equipment is highly desirable. This paper describes the process of combining models derived from laboratory experimentation with current operational plant data to infer an underlying measure of health. A demonstration of this process is provided where the fuel channel bore profile, a measure of core health, is inferred from data gathered during the refuelling process of an Advanced Gas-cooled Reactor (AGR) nuclear power plant core. Laboratory simulation was used to generate a model of an interaction between the fuel assembly and the core. This model is used to isolate a single frictional component from a noisy input signal and use this friction component as a measure of health to assess the current condition of the graphite bricks that comprise the core. In addition, the model is used to generate an expected refuelling response (the noisy input signal) for a given set of channel bore diameter measurements for either insertion of new fuel or removal of spent fuel, providing validation of the model. This benefit of this work is that it provides a greater understanding of the health of the graphite core, which is important for continued and extended operation of the AGR plants in the UK.