Abstract
SphiNx is an SN-DFEM-based neutron transport code built on the MOOSE framework, designed to solve both eigenvalue and fixed-source problems with support for anisotropic scattering. Two eigenvalue-solving algorithms, the traditional power iteration (PI) accelerated by the coarse mesh finite difference (CMFD) method and the Jacobian-Free Newton-Krylov (JFNK) method, are implemented in SphiNx with the support of solvers provided by MOOSE. The code's accuracy is validated through benchmarks, including a 1D anisotropic scattering problem, the IAEA 2D problem, and the Takeda 3D benchmark, showing eigenvalue discrepancies below 100 pcm and flux errors under 5 %. In the conducted performance comparisons, the CMFD-accelerated PI method demonstrates reduced computational time relative to the JFNK method. SphiNx also demonstrates strong parallel efficiency, maintaining a 46.11 % efficiency with a 1008-fold increase in processor count.
| Original language | English |
|---|---|
| Article number | 103641 |
| Journal | Nuclear Engineering and Technology |
| Volume | 57 |
| Issue number | 9 |
| DOIs | |
| State | Published - Sep 2025 |
| Externally published | Yes |
Keywords
- Coarse mesh finite difference method
- Discontinue finite element method
- Discrete ordinates
- JFNK
- MOOSE
- Neutron transport
Fingerprint
Dive into the research topics of 'SphiNx: A SN-DFEM neutron transport code based on the MOOSE framework'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver