Abstract Details
Abstracts
Author: Ben Zhu
Requested Type: Poster
Submitted: 2024-04-14 10:48:38
Co-authors: B. Dudson
Contact Info:
Lawrence Livermore National Laboratory
7000 East Avenue, L-440
Livermore, CA 94550
USA
Abstract Text:
While flux-coordinate independent (FCI) method has been adopted by a few recently developed boundary plasma models for both tokamaks (e.g., GDB, GRILLIX and GENE-X) and stellarators (e.g., BSTING), its validity and performance have not yet been systematically studied and compared to the conventional field-line aligned (FA) approach and the less common direct approach (DA). In this work, we try to address two main concerns often raised for FCI method, perpendicular pollution and projection error. The former one is done by comparing the accuracy of simulation results with three different approaches in an analytically solvable thermal diffusion test where a field-line aligned Gaussian-shape filament is initialized on a rational surface. We find that FA field-line approach is robust, reliable, and computationally efficient. For FCI, perpendicular pollution can be largely eliminated once the interpolation error is reduced. Computationally, it can be as efficient (i.e., similar time-step) as the FA approach if a FA-equivalent mesh is used. In this test, FCI and FA approach yield the same accuracy with the same number of grid points. Direct approach may need a higher resolution, thus a smaller dt and more expensive computationally to obtain a solution without substantial perpendicular pollution. To quantify the projection error of FCI approach (i.e., FCI often assumes the poloidal cross-section is the “perpendicular” plane to the equilibrium magnetic field), 3D Hasegawa-Wakatani simulations are performed, and turbulence statistics are compared between the FCI and DA simulations. These results will be presented at the conference.
Comments: