An intelligent system for interpreting the nuclear refuelling process within an advanced gas-cooled reactor

Abstract

Evaluation of the data produced during the refuelling process in a nuclear power plant is required to ensure proper ‘set-down’ of the fuel assembly, thereby allowing the continued and safe operation of the station. The process of evaluating the data can be time consuming owing to the large amounts of data requiring considerable domain experience and interpretation. This paper presents an intelligent system (IS) to automate the process of data analysis, thereby shortening the evaluation time and providing an explanation of the reasoning behind its conclusions. The intelligent system utilizes a knowledge-based system (KBS), neural network based classification, K-means clustering techniques and rule induction methods to evaluate the data and inform the operator of any errors encountered.