A discrete fracture model for two-phase flow in fractured porous media. (December 2017)

• Record Type: Journal Article
• Authors: Gläser, Dennis; Helmig, Rainer; Flemisch, Bernd; Class, Holger
• Is Part Of: Advances in water resources. Volume 110(2017)
• Page Start: 335
• Page End: 348
• Publication Date: 2017-12
• Subjects: Hydrology -- Periodicals; Hydrodynamics -- Periodicals; Hydraulic engineering -- Periodicals; 551.48
• DOI: 10.1016/j.advwatres.2017.10.031
• Languages: English
• ISSNs: 0309-1708

Abstract

Abstract: A discrete fracture model on the basis of a cell-centered finite volume scheme with multi-point flux approximation (MPFA) is presented. The fractures are included in a d -dimensional computational domain as ( d − 1 )-dimensional entities living on the element facets, which requires the grid to have the element facets aligned with the fracture geometries. However, the approach overcomes the problem of small cells inside the fractures when compared to equi-dimensional models. The system of equations considered is solved on both the matrix and the fracture domain, where on the prior the fractures are treated as interior boundaries and on the latter the exchange term between fracture and matrix appears as an additional source/sink. This exchange term is represented by the matrix-fracture fluxes, computed as functions of the unknowns in both domains by applying adequate modifications to the MPFA scheme. The method is applicable to both low-permeable as well as highly conductive fractures. The quality of the results obtained by the discrete fracture model is studied by comparison to an equi-dimensional discretization on a simple geometry for both single- and two-phase flow. For the case of two-phase flow in a highly conductive fracture, good agreement in the solution and in the matrix-fracture transfer fluxes could be observed, while for a low-permeable fracture the discrepancies were more pronounced. The method is then applied two-phase flow through a realistic fracture network in two and three dimensions.