Dispersivity calculation in digital twins of multiscale porous materials using the micro-continuum approach

TL;DR

This paper introduces a micro-continuum approach to calculate dispersivity in multiscale porous materials within digital twins, enabling seamless modeling of flow and transport across different pore resolutions.

Contribution

The work develops a novel derivation of the closure problem and a multi-stage process to compute effective dispersivity tensors using high-resolution imaging and DNS, integrating them into multiscale models.

Findings

Validated method against fully-resolved simulations in 2D micromodels.

Applied approach to hierarchical ceramic foams and carbonate rocks.

Established dispersivity as a function of Péclet number for multiscale modeling.

Abstract

The micro-continuum method is a novel approach to simulate flow and transport in multiscale porous materials. For such materials, the domain can be divided into three sub-domains depending on the local porosity {\epsilon}: fully resolved solid phase, for which {\epsilon}=0, fully resolved pores, for which {\epsilon}=1.0, and unresolved pores, for which 0<{\epsilon}<1.0. For such domains, the flow can be solved using the Darcy-Brinkman-Stokes (DBS) equation, which offers a seamless transition between unresolved pores, where flow is described by Darcy's law, and resolved pores, where flow is described by the Navier-Stokes equations. Species transport can then be modelled using a volume-averaged equation. In this work, we present a derivation of the closure problem for the micro-continuum approach. Effective dispersivity tensors can then be calculated through a multi-stage process. First, high resolution images are chosen for characterizing the structure of the unresolved pores. Porosity, permeability and effective dispersivity for the unresolved part are calculated by solving a closure problem based on Direct Numerical Simulation (DNS) in the high-resolution images. The effective dispersivity is then expressed as a function of the P\'eclet number, which describes the ratio of advective to diffusive transport. This relationship, along with porosity and permeability, is then integrated into the multiscale domain and the effective dispersivity tensor for the full image is calculated. Our novel method is validated by comparison with the numerical solution obtained for a fully-resolved simulation in a multiscale 2D micromodel. It is then applied to obtain an effective dispersivity model in digital twins for two multiscale materials: hierarchical ceramic foams and microporous carbonate rocks.