Neural Network Approach to Model Mixed Oxide Fuel Cycles in Cyclus, a Nuclear Fuel Cycle Simulator

摘要

Nuclear fuel cycle simulators (NFCSs) are fundamental in guiding policy and economic decisions regarding nuclear fuel cycle (NFC) options. This paper introduces a new method to predict the mixed oxide fuel (MOX) criticality value using an artificial neural network (ANN) model, while most current NFCSs use simple MOX fabrication estimations that do not account for burnup effects. The authors generated over one million depletion simulation results of MOX fuel with varying plutonium vectors and plutonium content to train an ANN network to predict the fuel's Beginning of Cycle (BOC) and End of Cycle (EOC) criticality. Results show that the trained ANN can predict criticality of MOX fuel within 1% error compared with the test data.