Multiphase flow in porous media: Pore-scale to macroscopic modeling

The objective of this work is to develop a fundamental understanding of multiphase flow in oil reservoirs through pore-scale modeling. Pore-scale modeling is a powerful tool which may be used to calculate macroscopic flow properties such as capillary pressure and relative permeability. In addition, pore-scale models provide insights into the accuracy of assumptions in continuum-scale models used in reservoir simulation.; A two-phase, pore-scale model is implemented to determine the effect of pore-space correlations in rocks on oil-water flow characteristics. It is demonstrated that capillary pressure and relative permeability curves for uncorrelated and finite-correlation length media are deterministic, while those for fractal media are stochastic. The amount of oil that can be recovered from a reservoir by waterflooding decreases with increasing extent of spatial pore-space correlation.; A pore-scale network model is developed to describe three-phase flow during gas invasion in a medium saturated with water and oil. This model describes the effect of the spreading coefficient, which plays a key role in determining the amount of oil that can be recovered from a reservoir following gravity drainage or gas injection. Oil residuals close to zero are possible for systems with positive spreading coefficients while those for negative spreading coefficient systems can be considerably higher. Using the model, three-phase capillary pressure and relative permeability curves are calculated. Transport of isolated oil blobs is possible in three-phase systems, which cannot be described within the framework of existing models to describe multiphase flow. A modification to Darcy's law is proposed to describe this transport mechanism.; Saturation histories, other than those studied at the pore scale in this work, are possible during three-phase flow in reservoirs. A macroscopic model is developed to identify possible saturation histories during three-phase flow. A consistent formulation, linking pore-scale and macroscopic models, is proposed.