Investigation of the porous drag and permeability at the porous-fluid interface: Influence of the filtering parameters on Darcy closure

• Published in: Particuology 2023-07, Vol.78, p.122-135
• Main Authors: Sadowski, Wojciech, di Mare, Francesca
• Format: Article
• Language: English
• Subjects: Explicit filtering; Permeability; Porous drag; Porous-fluid interface
• ISSN: 1674-2001; 2210-4291
• DOI: 10.1016/j.partic.2022.09.010

The porous-fluid interface encompasses a region bridging the flow inside a porous medium and a free-flowing fluid. In the context of volume-averaged simulations, it can be described by a set of gradually changing parameters defining the porous medium, mainly porosity and permeability. In this paper, both the permeability and the porous-induced drag force are evaluated a-priori, by explicitly filtering a set of Particle-Resolved Simulations (PRS) of the flow in the channel partially occupied by the porous medium. Different porous matrices are considered and the influence of the geometry and filtering parameters on the macroscopic quantities is studied. Especially, the focus is placed on the requirements for the kernel type and size to perform filtering accurately, and their impact on the distribution of permeability at the interface. The performance of the typically used models for the permeability is compared to the explicitly filtered results. Lastly, a new model for permeability and the drag force is introduced, taking into account the information about the filtering size and non-uniformity of the velocity field. The model greatly improves the prediction of velocity at the porous-fluid interface and serves as a proof of concept that a successful porous drag model should strive to include information about both parameters. [Display omitted] •Permeability at the porous-fluid interface is highly dependent on the filtering size.•A permeability model including the information about velocity non-uniformity greatly improves velocity prediction at the interface and in the free fluid region.•Filtering Particle-Resolved Simulations in ordered porous media with Gaussian kernel provides much more robust results than using compactly supported kernels.