Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. On upscaling certain flows in deformable porous media
 Multiscale modeling & simulation 7 (2008), No.1, pp.93123 ISSN: 15403459 

 English 
 Journal Article 
 Fraunhofer ITWM () 
Abstract
We consider certain computational aspects of upscaling fluid flows through deformable porous media. We start with pore level models and discuss upscaled (homogenized) equations and respective cell problems. Analytical solution of a cell problem in certain geometry, as well as an accurate numerical procedure for the general case, are presented. A microscale (pore level) fluidstructure interaction (FSI) problem is formulated in terms of incompressible Newtonian fluid and a linearized elastic solid. At least three different macroscopic models can be derived from this microscale formulation, depending on the assumptions on the fluidstructure interface. The first two are the wellknown linear poroelasticity model [M. Biot, J. Appl. Phys., 12 (1941), pp. 155164] and its nonlinear extension [C. K. Lee and C. C. Mei, Int. J. Eng. Sci., 35 (1997), pp. 329352]. Both are derived under the assumption of small (at pore level) displacements of the interface with the difference that the first model excludes pore scale rigid body motions, while the second one accounts for them. A third macroscopic model is explained below. It concerns a particular case, namely, a porous medium formed by long parallel channels with thick elastic walls. An asymptotic solution to the FSI problem is derived for such a geometry, allowing finite (at pore scale) displacements for the interfaces. A nonlinear Darcytype upscaled equation for the averaged pressure is obtained. The cell problems for each of the above cases, as well as a numerical algorithm for solving these cell problems, are discussed. The microscale cell FSI problems are treated numerically by an iterative procedure which solves sequentially fluid and solid subproblems and couples them via the interface conditions. Numerical and asymptotic solutions are found to converge to each other, thus validating both the numerical solver and the analytical derivation.