Pore-scale statistics of flow and transport through porous media.

摘要: Flow in porous media is known to be largely affected by pore morphology. In this work, we investigate the effects of geometry on transport and spatial correlations flow through two distinct structures arising from three-dimensional assemblies overlapping nonoverlapping spheres. Using high-resolution direct numerical simulations (DNS), perform Eulerian Lagrangian analysis characteristics media. We show that velocity distributions change nearly exponential Gaussian as porosity increases. A stretched distribution can used represent behavior for a wide range porosities. Evolution velocities studied uniform injection rule. Evaluation tortuosity trajectory length each medium shows model spheres results higher more skewed compared Wider give rise non-Fickian while found Fickian. Particularly, our first-passage time very similar those observed sandstone. Finally, using (3D) field obtained DNS at pore-scale, quantitatively despite randomness pore-space, spatially fluctuating 3D pore-space are strongly correlated relatively homogeneous monodisperse sphere packs Castlegate