Abstract | ||
---|---|---|
In this paper we study the flux formulation of unsteady diffusion equations with highly heterogeneous permeability coefficients and their discretization. In the proposed approach first an equation governing the flux of the unknown scalar quantity is solved, and then the scalar is recovered from its flux. The problem for the flux is further discretized by splitting schemes that yield locally one-dimensional problems, and therefore, the resulting linear systems are tridiagonal if the spatial discretization uses Cartesian grids. A first and a formally second order time discretization splitting scheme have been implemented in both two and three dimensions, and we present results for a few model problems using a challenging benchmark dataset. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1016/j.cam.2017.12.003 | Journal of Computational and Applied Mathematics |
Keywords | Field | DocType |
Direction-splitting,Multi-scale methods,Parabolic equations,Flux-splitting | Tridiagonal matrix,Parabolic partial differential equation,Discretization,Linear system,Mathematical analysis,Scalar (physics),Flux,Mathematics,Cartesian coordinate system | Journal |
Volume | ISSN | Citations |
340 | 0377-0427 | 0 |
PageRank | References | Authors |
0.34 | 6 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
P. D. Minev | 1 | 22 | 6.23 |
S. Srinivasan | 2 | 2 | 2.40 |
Petr N. Vabishchevich | 3 | 37 | 27.46 |