Mesh Generation for Fusion Codes
Mesh generation builds the computational grids fusion codes solve on, aligning them to magnetic geometry while handling singular features like the X-point.
The grid comes first
Every field-solving fusion code needs a mesh: the discrete grid on which the equations are approximated. Mesh generation for fusion is specialized because the physics is strongly anisotropic, transport along field lines is enormously faster than across them, so grids are usually aligned to the magnetic field to represent this anisotropy accurately without numerical pollution.
A poor mesh corrupts the solution regardless of the physics model. Mesh quality, cell aspect ratio, smoothness, and alignment, is a first-order determinant of accuracy and stability.
Flux-aligned grids
Core and edge codes build grids from flux surfaces, so that one grid direction follows the field and another crosses it. This lets high parallel conductivity be resolved cheaply. The challenge is the region near the separatrix and X-point, where flux surfaces pinch together and flux-aligned coordinates become singular.
Handling the X-point
At the X-point the poloidal field vanishes and a purely flux-aligned mesh degenerates. Practical edge grids split the domain into blocks, core, SOL, and private-flux regions, meshed separately and stitched, or use flexible unstructured meshes that tolerate the singularity. This is one of the more intricate tasks in fusion computing.
Design relevance
For the tightly shaped Hyperion breeder, with its low aspect ratio and negative triangularity, generating good field-aligned meshes for equilibrium, edge, and MHD codes is a prerequisite for every downstream simulation. Mesh quality is checked as part of the verification record before construction.
- Builds the grid codes solve on
- Field-aligned to capture strong anisotropy
- X-point breaks flux-aligned coordinates
- Block-structured or unstructured meshes handle it