Finding the Dominant Properties of Porous Media: An Example of Fluid Flow

Elucidating the complex impact of the geometry and the topology of a porous environment on its performance is important in both studying and computational modeling of such media. This work complements these efforts by putting forth a novel approach capable of identifying media characteristics likely to dominate the underlying physical phenomena. We specifically focus on the flow of a fluid through the medium, due to its significance in nature and emerging engineering applications. Our approach uses statistical analysis to infer flow preferences from small perturbations of the porous environment. We show that the features identified by our methodology are uniquely bound to the type of examined media and that their choice seems uncorrelated among different material classes. We believe that the proposed approach can provide a first estimate of media characteristics that can be tuned in their real applications to achieve better performance. Additionally, it can be used to improve computational approximations, such as two-dimensional or network models.