Multi-Level iterative methods in computational plasma physics
Plasma physics phenomena occur on a wide range of spatial scales and on a wide range of time scales. When attempting to model plasma physics problems numerically the authors are inevitably faced with the need for both fine spatial resolution (fine grids) and implicit time integration methods. Fine grids can tax the efficiency of iterative methods and large time steps can challenge the robustness of iterative methods. To meet these challenges they are developing a hybrid approach where multigrid methods are used as preconditioners to Krylov subspace based iterative methods such as conjugate gradients or GMRES. For nonlinear problems they apply multigrid preconditioning to a matrix-few Newton-GMRES method. Results are presented for application of these multilevel iterative methods to the field solves in implicit moment method PIC, multidimensional nonlinear Fokker-Planck problems, and their initial efforts in particle MHD.
- Research Organization:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 757012
- Report Number(s):
- LA-UR-99-1128; TRN: US0003755
- Resource Relation:
- Conference: APS, Atlanta, GA (US), No date supplied; Other Information: Conference held during 03/1999; PBD: 1 Mar 1999
- Country of Publication:
- United States
- Language:
- English
Similar Records
Terascale Optimal PDE Simulations (TOPS) Center
Enhanced nonlinear iterative techniques applied to a nonequilibrium plasma flow