 |
| Titel |
Simulation of two-phase flow in porous media using mimetic finite difference methods |
| VerfasserIn |
Peter Bastian, Olaf Ippisch, Sven Marnach, I. Sorin Pop |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250040985
|
|
|
|
|
|
| Zusammenfassung |
| The simulation of two-phase flow in porous media has a wide variety of applications. The
equations governing these flows are inherently nonlinear. In applications like CO2
sequestration in geological formations and petroleum engineering, we often have to cope with
unstructured geometries and highly heterogeneous media, which raise additional difficulties
for the simulation. In our work, we try to improve the numerical methods used in these
simulations.
The numerical methods most commonly used for multiphase flow in porous media – the
cell-centered finite volume (FV) method, the vertex-centered finite volume method and the
finite element (FE) method – aren’t suited particularly well for simulation on complex
geometries and heterogeneous media. Cell-centered FV methods only work on a very
restricted class of grids and do not allow for local refinement. Vertex-centered FV methods
perform poorly on highly heterogeneous materials. Standard FE methods aren’t locally mass
conservative.
For diffusion type problems, mimetic finite difference (MFD) methods remedy these
shortcomings of the commonly used methods. MFD methods can be considered as a
generalization of cell-centered finite volume methods to unstructured grids and have proven
highly robust and accurate in applications. Moreover, they are suitable for even more general
grids then vertex-centered FV methods.
In our work, mimetic finite difference methods are adapted to the case of multiphase flow
in porous media. We present a way to apply this method to the fully coupled, fully implicit
solution of the two-phase Darcy flow equations. We also cover our quite general
and flexible simulation framework based on DUNE and DUNE-PDElab and show
some numerical examples computed on different kinds of computational grids. |
|
|
|
|
|
|