Hybrid Monte Carlo-deterministic method for 3D neutron transport simulation based on energy region division

摘要

Monte Carlo (MC) and deterministic methods are widely used for neutron transport simulation. The MC method computes with an accurate description of geometry and cross section but with a long execution time, while the deterministic method has higher computational efficiency but poor accuracy, especially in the resonance energy range. To achieve both high efficiency and accuracy, a hybrid MC-deterministic method was developed for three-dimensional neutron transport simulation based on SuperMC. Using a directly delivered scattering source, the simulation was conducted in the fast region by the deterministic method, resonance region by the MC method, and thermal region by the deterministic method. Detailed analysis was conducted to verify the performance of the hybrid method. Based on the fixed-source case and criticality case, higher accuracy of the hybrid method compared to the deterministic method and better acceleration performance compared to the MC method was confirmed. (C) 2019 Elsevier Ltd. All rights reserved.