Domain Decomposition Methods Applied to a System of ConvectionDiffusion-Reaction Equations
|
Abstract:
Obtaining efficient and robust solutions to realistic fluid flow problems with chemical reactions is a challenging computational task. This is due to the potential wide disparity in time and space scales and the severe nonlinearities. However, there is significant industrial application for such a capability in areas such as energy production, propulsion, and materials processing. These applications are often characterized by regions of low Mach number flow. This study investigates Newton-Krylov algorithms with domain-based preconditioning for the solution of these types of problems. Domain-based preconditioners may offer improvements in robustness and parallelism over more standard ILU type preconditioners. Results are presented comparing additive and multiplicative Schwarz preconditioners to more standard global ILU preconditioners on a low Mach number combustion problem. 1.
Citations
| 180 | Inexact Newton Methods – Dembo, Eisenstat, et al. - 1982 |
| 70 | Domain decomposition algorithms with small overlap – Dryja, Widlund - 1994 |
| 41 | An iterative solution method for linear systems of which the coe cient matrix is a symmetric M-matrix – Meijerink, Vorst - 1977 |
| 37 | Schema Theory – Williams - 1994 |
| 33 | Overlapping domain decomposition algorithms for general sparse matrices – Cai, Saad - 1996 |
| 28 | Matrix-free methods for stiff systems of ODE’s – Brown, Hindmarsh - 1986 |
| 13 | Newton-Krylov methods for low Mach number compressible combustion – Knoll, McHugh, et al. - 1996 |
| 6 | Domain decomposition methods for the parallel computation of reacting flows – Keyes - 1989 |
| 4 | Aconjugate gradient-truncated direct method for the iterative solution of the reservoir simulation equation, Soc.ofPetroleum Eng – Watts - 1981 |
