Towards an estimation aid for nuclear power plant refuelling operations

Analysis of monitored refuelling data is required to confirm that refuelling operations have been correctly completed and that the reactor plant is in a safe condition for continued operation, This paper describes a methodology for identifying key points in the refuelling process thereby providing decision support for post-refuelling analysis. A feature identification technique is described which provides reliable input to Artificial Neural Networks (ANNs) and regression estimation techniques. This technique is shown to be robust against variations in the input data. The analysis in this paper shows that the regression models and ANNs can also provide similarly accurate predictions of a key refuelling event.