Yuanzineng kexue jishu (Mar 2024)
Generating Multi-group Homogenized Cross-sections Using Continuous-energy Monte Carlo Method for Fast Reactor Analysis
Abstract
The deterministic two-step method for the fast reactor neutronics calculation, composed of cross-section homogenization and diffusion or transport core calculations, was widely applied in the fast reactor engineering design and analysis field. The homogenized cross-section calculation method based on Monte Carlo with continuous-energy and fine geometry can provide high-precision cross-sections for advanced fast reactors. The current development status and trends of coupling Monte Carlo-generated homogenized cross-sections with diffusion and transport core calculations were briefly reviewed in this paper. The methods discussed include the Monte Carlo flux-volume homogenization method, the superhomogenization equivalence technique (SPH), and the Monte Carlo flux-moment homogenization method (MHT). The MET-1000 metal fuel fast reactor is used as a benchmark. The SPH equivalent techniques are widely used to preserve the reaction rates of a reference heterogeneous model and a homogenous model. In this paper, the SPH was applied to the control rods' cross-section address to improve the diffusion core calculations. This equivalence technique reduces the overestimation of the control rod worth using the diffusion core solver from 13.5% to 0.35% and improves power distribution prediction accuracy. With SPH correction, the MC/diffusion in this work exhibits about <±4% error as the insertion of control rods in power distribution. For the transport core calculations, the reasons for core reactivity overestimation were quantitatively analyzed, and the MHT method was developed. The basic principle of the MHT homogenization method is to incorporate the anisotropy of the total cross-section concerning the incident angle into the scattering matrix. This allows for the consideration of cross-section anisotropy while maintaining the generality of the generated total cross-section within the core transport solver. The MHT reduces the error of the transport core solving of MET-1000 by 698 pcm. The factors that cause the residual bias were discussed, but there is only about 130 pcm unsolved bias. The flux-volume homogenization method exhibits uneven error distribution, tending to underestimate the power at the inner core top and overestimate the power at the outer core bottom, with errors ranging from -3.63% to +4.02%. The MHT homogenization method reduces power prediction errors, with errors ranging between -2.39% and +2.76%, and achieves a more uniform error distribution. This work presented Monte Carlo homogenized cross-section generation methods applicable to diffusion and transport core calculations for fast reactor neutronics analysis. The MHT homogenization method provides a novel approach for generating cross-sections suitable for core transport calculations in Monte Carlo simulations. However, further validation is needed with different core solvers and fast reactors such as small fast reactors and more heterogenous fast reactors. The Monte Carlo homogenization method still requires extensive research in various aspects, including the generation of discontinuous factors, the BN leakage model, and methods for handling historical effects.
Keywords
