Understanding how porosity gradients can make a better filter using homogenization theory

Abstract

Filters whose porosity decreases with depth are often more efficient at removing solute from a fluid than filters with a uniform porosity. We investigate this phenomenon via an extension of homogenization theory that accounts for a macroscale variation in microstructure. In the first stage of the paper, we homogenize the problems of flow through a filter with a near-periodic microstructure and of solute transport owing to advection, diffusion and filter adsorption. In the second stage, we use the computationally efficient homogenized equations to investigate and quantify why porosity gradients can improve filter efficiency. We find that a porosity gradient has a much larger effect on the uniformity of adsorption than it does on the total adsorption. This allows us to understand how a decreasing porosity can lead to a greater filter efficiency, by lowering the risk of localized blocking while maintaining the rate of total contaminant removal.

The asymptotic homogenization elasticity tensor properties for composites with material discontinuities

“…High performance parallel computing would be then necessary to solve the cell problems via independent instances, thus reducing the global computational time. In [5,10], it is also shown how to bypass this issue for specific geometrical setting (including, for example, a collection of spherical obstacles) in the context of advection-diffusion and porous media flow problems.…”

The role of the microvascular network structure on diffusion and consumption of anticancer drugs

“…As expected from the analysis of the Stokes' probles reported in [13], the chosen microstructure geometry reflects in negligible extra-diagonal components, when compared to the diagonal ones. We therefore account for the isotropic representations (39) and (40) for the purpose of our analysis. As done in [13] for the hydraulic conductivity, the impact of tortuosity is tested on the effective diffusivities by varying the geometrical configurations of Ω n and Ω t .…”

“…In this case, one cell problem for each macroscale point should be solved, and the challenge resides in the development of efficient algorithms, which should couple macro and micro spatial variations via parallel computing together with adequate image segmentation. Alternatively, for a range of simplified microstructure, it is possible to parametrically vary the parameter(s) (such as the radius of circular objects) characterizing the microscale geometry, as done for example in [39] for advection-diffusion of solute through a macroscopically varying porous structure made of circular obstacles. Note that the solutions obtained in this work, as it usually happens for models that make use of asymptotic homogenization, are usually good over the domain interior, but may be inaccurate in regions close to the domain boundary.…”

Designing asymmetric multilayered membrane filters with improved performance

“…While a full CFD simulation provides excellent insight into how an individual particle is trapped, computational costs associated with keeping track of all the particles within a complicated pore structure makes it impractical on a large scale. 1 Dalwadi et al [13] use homogenization theory to explore the improved filtration observed in a continuous porosity-graded filter. The model presented describes the motion of contaminants within a continuous media as they are transported via advection and diffusion.…”