Generating multi-group cross-sections using continuous-energy Monte Carlo method for fast reactor analysis

Abstract

The deterministic two-step method, comprising multigroup cross-section generation and core calculation, is widely applied in fast reactor design and analysis. Monte Carlo (MC) methods with continuous energy and fine geometry provide high-precision cross-sections essential for advanced fast reactor neutronics analysis. This paper presents an analysis of integrating MC-generated homogenized cross-sections with various core solvers, demonstrating their effectiveness and potential improvements in fast reactor simulations. For diffusion core calculations, the superhomogénéisation (SPH) technique reduces control rod worth overestimation from 13.5% to 0.35% in the MET-1000 benchmark, improving power distribution predictions. In transport core calculations, the flux-moment homogenization technique (MHT) addresses reactivity overestimation by incorporating cross-section anisotropy, reducing error by 698 pcm. For Method of Characteristics (MOC) core calculations, transport-corrected multigroup cross-sections yield high precision in pin-by-pin power distribution for a 100 MWe lead-based fast reactor benchmark. While MC methods require significant computational resources, such as 62 CPU-hours for the MET-1000 core and 85.5 CPU-hours for the 100 MWe lead-based fast reactor core, they offer flexibility in geometry modeling. This work highlights MC multigroup cross-section generation methods applicable to diffusion, MOC, and transport core calculations for fast reactor analysis, suggesting further exploration into their performance in various reactor parameters and computational efficiency.