How modeling assumptions shape predictions of convective mixing of carbon dioxide

TL;DR

This study examines how different modeling assumptions impact predictions of convective mixing in porous media relevant to CO2 storage, highlighting the importance of fluid properties, boundary conditions, and dimensionality.

Contribution

It provides a comprehensive analysis of how modeling assumptions influence convective mixing predictions, emphasizing the effects of density relationships and interface conditions.

Findings

Mixing is governed by mean scalar dissipation across cases.

Density-concentration relationships affect convection and maximum density position.

Simplified models can deviate in mixing rate predictions by up to 10-100%."

Abstract

We investigate how models of fluid properties and boundary conditions influence predictions of convective mixing in confined porous media, with relevance to subsurface carbon dioxide storage. Using high-resolution simulations at high Rayleigh-Darcy numbers (O(10$^4$)), we analyze miscible fluids with linear, nonlinear, and non-monotonic density-concentration relationships under fixed- and free-interface in 2D and 3D. We show that, across all cases, mixing is governed by the mean scalar dissipation, providing a unifying framework for convective-diffusive interactions. The density-concentration relationship affects mixing via the effective density contrast driving convection and the position of the maximum density. Free interfaces enhance early-time mixing through deformation, while long-term behavior depends on fluid properties and dimensionality. We demonstrate that simplified modeling assumptions (e.g., monotonic density laws or 2D flow) can lead to deviations in predicted mixing rates of up to O(10-100)\%. These results offer guidance for model selection and improving predictions of convective mixing in geophysical systems.