Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.03605 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914969895829504 |
|---|---|
| author | Wang, Dean |
| author_facet | Wang, Dean |
| contents | The quasidiffusion (QD) method, also known as the Variable Eddington Factor (VEF) method in the astrophysical community, is an established iterative method for accelerating source iterations in SN calculations. A great advantage of the QD method is that the diffusion equation that accelerates the SN source iterations can be discretized in any valid discretization without concern for consistency with the transport discretization. QD has comparable effectiveness with diffusion synthetic acceleration (DSA), but the converged scalar flux of the diffusion equation will differ from the transport solution by the spatial truncation errors. Larsen et al. introduced a new consistent QD method (CQD), which includes a straightforwardly defined transport consistency factor closely related to the well-known coarse mesh finite difference (CMFD) and DSA methods. The CQD method preserves the discretized scalar flux solution of the SN equations, and it is stable for problems with optically thin spatial cells, but just like nonlinear diffusion acceleration (NDA), it degrades in performance and eventually becomes unstable when the spatial cells become greater than about one mean free path thick. In this paper, we performed a formal Fourier analysis of the CQD method to show that its theoretical spectral radius is essentially the same as that of the NDA method. To improve the stability of CQD, we introduce the lpCQD method, which adopts the idea of the linear prolongation CMFD (lpCMFD) method. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_03605 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Stabilizing the Consistent Quasidiffusion Method with Linear Prolongation Wang, Dean Numerical Analysis The quasidiffusion (QD) method, also known as the Variable Eddington Factor (VEF) method in the astrophysical community, is an established iterative method for accelerating source iterations in SN calculations. A great advantage of the QD method is that the diffusion equation that accelerates the SN source iterations can be discretized in any valid discretization without concern for consistency with the transport discretization. QD has comparable effectiveness with diffusion synthetic acceleration (DSA), but the converged scalar flux of the diffusion equation will differ from the transport solution by the spatial truncation errors. Larsen et al. introduced a new consistent QD method (CQD), which includes a straightforwardly defined transport consistency factor closely related to the well-known coarse mesh finite difference (CMFD) and DSA methods. The CQD method preserves the discretized scalar flux solution of the SN equations, and it is stable for problems with optically thin spatial cells, but just like nonlinear diffusion acceleration (NDA), it degrades in performance and eventually becomes unstable when the spatial cells become greater than about one mean free path thick. In this paper, we performed a formal Fourier analysis of the CQD method to show that its theoretical spectral radius is essentially the same as that of the NDA method. To improve the stability of CQD, we introduce the lpCQD method, which adopts the idea of the linear prolongation CMFD (lpCMFD) method. |
| title | Stabilizing the Consistent Quasidiffusion Method with Linear Prolongation |
| topic | Numerical Analysis |
| url | https://arxiv.org/abs/2410.03605 |