Speaker
Description
Dual-permeability materials are porous materials containing distinct regions with two widely differing characteristic pore scales. These materials, as well as the more general case of multiscale materials, are prevalent in both nature and manufactured materials. They can exhibit unique behaviors, particularly when the differing scales lead to different transport and reaction processes However, the same reasons that create these interesting physical behaviors make modeling these systems a challenge. Various methods have been developed for modeling behavior in dual-permeability systems, including the Brinkman equation or similar approaches that augment the Darcy equation with additional terms, such that microporosity can be modeled using a Darcy approach while macroporosity can be modeled using a traditional fluid mechanics approach or an appropriate pore-scale method.
Alternatively, advances in computing power, multiscale imaging technology, and image-based modeling techniques are creating opportunities for direct modeling of dual permeability systems. This implies that all length scales are captured via direct numerical modeling of pore-scale transport. We are using unstructured meshing of the pore space with computational fluid dynamics to test this direct approach. The advantage is that it avoids the limiting assumptions and tricky boundary conditions that must be dealt with when integrating continuum and pore-scale methods. The disadvantage is that, as the difference in characteristic length scales becomes larger, the computational size and and/or stability of the algorithms will eventually become a problem.
In this work, we probe the limits of direct modeling of dual-permeability systems using i) unstructured tetrahedral meshing of the domains, ii) the finite element method for fluid flow, and iii) stochastic particle tracking for solute transport. Mesh generation is performed using modern open-source and in-house algorithms, to maximize flexibility in conforming to a variety of structures. The FEM flow modeling is performed using a locally-mass conservative algorithm from the open-source software package FEniCS. Finally, stochastic particle tracking is used to model mass transfer processes, which ties into the longer-term objectives of the research.
Results emphasize the impressive capabilities of modern mesh generation algorithms, in terms of control and adaptation of the meshes to complicated pore geometries, including both geometric (or CAD-based) structures and 3D digital images. By analyzing the flow, we have quantified the points at which the quality of the solution begins to degrade, which can result from any of the following factors: mesh resolution, structure of refinement, or disparity in overall mesh size. These results help define the limits of direct modeling of dual-scale porous materials.
| Country | USA |
|---|---|
| Acceptance of the Terms & Conditions | Click here to agree |
