Application of Genetic Algorithm in CANDU Reactor In-core Fuel Management

A strategy, using genetic algorithm （GA） to optimize and select the CANDU reactor refueling channels, is proposed in this paper. A back-propagation artificial neural network （BPANN） is incorporated to predict the core parameters, and the evaluation of the predicted results is taken as the fitness of the GA. GAREFUEL, a refueling channel selection code based on the GA and BPANN, is developed. The refueling simulation is performed using the GAREFUEL and FMPHWR for 360 days. The numerical results show that the selected refueling channels can meet the core parameter limits and GAREFUEL has superior efficiency and speed in selecting the refueling channels.