Abstract
Models of porous media are often applied to relatively small systems, which leads not only to system-size-dependent results, but also to phenomena that would be absent in larger systems. Here we investigate one such finite-size effect: anisotropy of the permeability tensor. We show that a nonzero angle between the external body force and macroscopic flux vector exists in three-dimensional periodic models of sizes commonly used in computer simulations and propose a criterion, based on the ratio of the system size to the grain size, for this phenomenon to be relevant or negligible. The finite-size anisotropy of the porous matrix induces a pressure gradient perpendicular to the axis of a porous duct and we analyze how this effect scales with the system and grain sizes. We also analyze how the size of the representative elementary volume (REV) for anisotropy compares with the REV for permeability.
Introduction
In the field of computational fluid dynamics (CFD), simulating flow through porous media at the pore-scale is computationally expensive. Consequently, researchers often utilize small numerical models, sometimes containing as few as two to eight grains. However, these small samples frequently fall below the Representative Elementary Volume (REV)—the minimum volume required for properties like permeability to become independent of the specific sample chosen. In these mesoscopic models, a medium that should be isotropic can exhibit significant random anisotropy, causing the fluid flux vector to point in a different direction than the driving force, creating systematic errors and unexpected pressure gradients.
Methodology
The study employed a stochastic approach to generate porous matrices by randomly distributing identical, overlapping impenetrable cubes within a cuboidal domain.
Two primary simulation tools were used to analyze the flow:
- Lattice Boltzmann Method (LBM): Implemented via the GPU-accelerated Sailfish library, allowing for massive speedups compared to traditional CPU codes.
- Finite Volume Method (FVM): Using the
icoFoamsolver from the OpenFOAM toolkit for validation.
The researchers analyzed hundreds of independent samples across two target porosities ($\phi = 0.35$ and $0.70$) and various system sizes ($L$) relative to grain size ($a$).
Key Results and Discussion
The central finding of this research is a universal scaling law for the dispersion of the flow angle ($\sigma_\alpha$). The data indicates that $\sigma_\alpha \propto L^{-d/2}$, which in three dimensions ($d=3$) simplifies to $\sigma_\alpha \propto (L/a)^{-3/2}$.
This scaling holds even for systems where the size is only five times the grain size, suggesting that small-scale simulations can predict macroscopic behavior if the scaling is applied correctly. For the specific cube-based model, the authors determined that to keep the flow angle error below $2^\circ$, the system size $L$ must be at least 20 times the grain size $a$. Furthermore, the study demonstrated that in ducts with solid walls, the induced perpendicular pressure gradient $|\nabla_\perp P|$ follows the same scaling behavior as the flow angle.
Conclusion
This work quantifies the finite-size effects that lead to "fake" anisotropy in porous media simulations. The identified scaling law allows researchers to predict the necessary simulation size to reach the REV. While the study focused on identical cubes, the authors suggest the $L^{-3/2}$ relationship is a general consequence of the Central Limit Theorem and should apply to more complex, real-world geometries, though further verification with polydisperse media or CT-scan data is recommended.
Summary updated