Mitigating Spatial Error in the iterative-Quasi-Monte Carlo (iQMC) Method for Neutron Transport Simulations with Linear Discontinuous Source Tilting and Effective Scattering and Fission Rate Tallies

• DOI: 10.48550/arxiv.2401.04029
• Publication Date: 2024-01-01
• Authors: Samuel Pasmann; Ilham Variansyah; C. T. Kelley; Ryan G. McClarren
• Keywords: Neutron Transport

The iterative Quasi-Monte Carlo (iQMC) method is a recently proposed for multigroup neutron transport simulations. iQMC can be viewed as hybrid between deterministic techniques, Monte simulation, and techniques. holds several algorithmic characteristics that make it desirable high performance computing environments including $O(N^{-1})$ convergence scheme, ray tracing sweep, highly parallelizable nature similar to analog Carlo. While there are many potential advantages of using also inherent disadvantages, namely the spatial discretization error introduced from use mesh across domain. This work introduces two significant modifications help reduce error. first an effective source whereby strength updated on-the-fly via additional tally. version sweep essentially agnostic mesh, material, geometry. second addition history-based linear discontinuous tilting method. Traditionally, utilizes piecewise-constant in each cell mesh. However, through technique utilize piecewise-linear without refining Numerical results presented 2D C5G7 Takeda-1 k-eigenvalue benchmark problems. Results show significantly reduces global tallies eigenvalue solution both benchmarks. Through was able converge problem less than $0.04\%$ on uniform Cartesian with only $204\times204$ cells.