Coarse Mesh Iteration Approach for Analytical 1D Multigroup $S_N$ Eigenvalue Problems
Jilang Miao, Miaomiao Jin
TL;DR
This work tackles efficient eigenvalue solutions for analytical multigroup $S_N$ equations in 1D slab geometry by developing a coarse-mesh, analytically solvable fixed-source framework. It rewrites the transport problem as $rac{d}{dx}oldsymbol{ m }(x) = A_0 oldsymbol{ m }(x) + Q(k_{eff},x)$ with an external source $Q$ derived from an eigensystem expansion on the coarse mesh and stabilizes the iteration using a Wielandt shift. The method demonstrates that coarse-mesh solutions achieve accuracy comparable to a fine-mesh fixed-source solver while delivering substantial speedups (up to roughly an order of magnitude depending on the SN order) and remains extendable to heterogeneous regions with proper interface handling. Results on a 35 cm slab show $k_{eff}$ differences shrinking from a few hundred to near-zero pcm as $S_N$ increases, and flux accuracies improving dramatically from percent-level errors to sub-percent levels. Overall, the coarse-mesh analytical SN approach provides a computationally efficient path for eigenvalue problems in slab geometries with strong potential for broader applicability in reactor-scale analyses.
Abstract
This paper extends the fixed source capability of analytical 1D multigroup $S_N$ equations to solve eigenvalue problems on coarse mesh.
