Search

In porous media flow, conventional methods struggle to determine a reliable representative elementary volume (REV), particularly in highly heterogeneous samples or dynamic scenarios. Previous approaches, often based on simplified geometries, lack accuracy, and overlook uncertainty quantification. Significant cross-coupling between phases can distort dynamic properties, leading to inaccurate REVs. This complexity is further compounded when considering the multiscale nature of rock geometry. To address these challenges, this thesis proposes a multiscale approach to upscale multiphase fluid flow in complex porous materials, providing a robust understanding of REV determination and effective material properties through insights from four research papers. The first paper introduces a new upscaling technique combining asymptotic homogenization with hydrodynamic and thermodynamic bounds. A novel variational thermodynamic approach yields upper and lower bounds of entropy production, deriving effective material properties with quantified uncertainties. The second paper focuses on the role of wettability and fluid saturations in determining the REV for multiphase flows. Finite volume simulations using Volume of Fluid (VoF) method highlight the influence of wetting conditions and initial saturations on sample size requirements. The third paper proposes a holistic approach for upscaling dynamic properties of multiscale microporous media. Hydrodynamic and thermodynamic bounds, multi-scale domain discretisation, and an open-source numerical solver achieve a complete representation of the system’s behaviour. Application to distinct carbonate samples emphasizes the impact of multiscale nature on REV estimations. The fourth paper provides macroscopic validation for microscopic data and insights on fractures within the research. Laboratory-scale models and numerical simulations visualize flow processes in fractured cores, revealing the complex interplay of fracture orientations and aging effects on oil recovery. This validation consolidates the findings of the third paper and sets the stage for future research by providing a comprehensive framework for fractured media. In conclusion, this thesis delivers a novel framework for upscaling multiphase flow in complex porous materials. By linking microscopic dynamics to macroscopic behaviour, it advances our understanding of fluid flow in heterogeneous media, with profound implications for engineering and environmental applications.