Water Resources Research · 2008 · 29 citations · 27 references
EngineeringChemical EngineeringComputational Fluid DynamicsComputational Materials ScienceTransport PhenomenaFluid-flow In Porous MediaPorous MaterialsPorous Media FlowTransport In Porous MediaTransport Property SimulationDensityDispersion CoefficientHydrodynamic DispersionPackingPacking DensityPorous Media DispersionTransport Phenomena In Porous MaterialsTransport Properties In Porous Media
Pore‐scale simulations of flow and transport offer the possibility of modeling laboratory‐scale experiments without averaging the properties of the porous medium. Comparisons between simulation and experiment are complicated, however, by the difficulty of reproducing the exact geometry of the experimental porous medium in a simulation. Flow and dispersion are affected by differences in packing density, random packing variation, inhomogeneities, and confining walls. It is therefore important to understand the magnitude of such effects when comparing simulation and experiment. We quantify the sensitivity of pore‐scale simulations of flow and dispersion to variations in packing density, the random packing algorithm, and selective sphere removal for the case of randomly packed spheres. The results allow us to interpret differences in dispersion between pore‐scale simulations and NMR spectroscopy measurements in packed cylinders. Dispersion is analyzed in terms of the time‐dependent dispersion coefficient and by the propagator function (i.e., the density function of solute mass versus mass propagation distance). The simulation and experimental propagator functions share important qualitative features, and there are no obvious problems such as an incorrect center of mass. However, the simulation underestimates experimental dispersion, and the underestimate is greater than can be explained by possible differences in packing density and random packing between the NMR and simulated sphere packings. An examination of NMR images of the experimental packing reveals regularities in packing structure and raises the possibility that nonrandom ordering could account for the difference in dispersion.
27
Computer Simulation of Liquids
S. P, K. J · Journal of Molecular Liquids · 1988 · 15.9K citations
Understanding Molecular Simulation
Daan Frenkel, Berend Smit, Jan Tobochnik et al. · Computers in Physics · 1997 · 5.5K citations
3-D Pore-Scale Modelling of Sandstones and Flow Simulations in the Pore Networks
S. Bakke, Pål‐Eric Øren · SPE Journal · 1997 · 623 citations
Engineering, Earth Sciences, Geology +17