ARC INRIA GEOFRAC
Large-scale computation of flow in complex 3D geological fractured porous media
Welcome to ARC GEOFRAC
In the last twenty years, interest in geological fractured rocks has been renewed by a variety of energy-related applications, such as carbonate oil reservoir exploitation, geothermal energy production, geological storage of high level nuclear waste, and geological sequestration of CO2. Fractures are highly permeable pathways within a less pervious but more porous medium generally called the rock matrix. Whatever the application or the details of the involved physical and chemical processes, the relative spatial organization of the fracture and matrix phases has a strong influence on exchanges between them. To date most models for flow in fractured media have relied on the separate homogenization of the fracture and matrix phases taken in a second step as two interacting continua. This so-called double-porosity approach has been motivated by the lack of extensive data on the fracture locations, by its conceptual simplicity and by the possibility of using existing developments on the more classical single porosity models. Important limitations of the double porosity approach have however recently come to light. An alternative approach is the discrete modeling of fractures. This approach encounters at least two challenging numerical issues. First, the fracture and matrix phases have very different hydraulic properties. Permeability is at least two orders of magnitude higher in the fractures than in the matrix. Specific numerical methods have to be used to cope efficiently with these highly localized discontinuities. Second, the complexity of the fracture structure creates intricate geometrical configurations which are difficult to mesh. We propose to address these issues by developing new numerical methods adapted to sharp discontinuities. The proposed methods should be highly robust and computationally efficient for handling upscaling problems in a stochastic modeling framework. They will be used in the last stage of this project for determining the upscaling laws in the most commonly encountered fracture formations. The numerical methods will be implemented in the H2OLAB development platform used for modeling hydraulic processes in heterogeneous fractured reservoirs.