Efficient Computational Methods for Modeling Fluid-Rock Interactions in Geologic CO2 Storage

Abstract:

Geological storage of CO2 can only be realistically modeled by taking into account reactive fluid-rock interactions. After supercritical CO2 is injected underground, it partially dissolves into the resident brine and increases its acidity. This increase in acidity combined with high underground temperatures stimulates geochemical reactions that can dramatically alter rock properties, such as porosity and permeability, as well as the capability of brine to dissolve more mobile supercritical CO2. Combining these geochemical reactions with reactive transport simulations is then essential for understanding the long-term fate of injected CO2 and estimating carbon trapping via solubility and mineral precipitation mechanisms.

In this talk we present efficient and versatile computational methods for modeling fluid-rock interactions. The talk starts with an overview of common methodologies based on the solution of the law of mass-action equations for calculating equilibria between aqueous, gaseous, and mineral species. This is followed by a discussion of the deficiencies and drawbacks of these methods. We then present alternative approaches based on the minimization of Gibbs energy that resolves the disadvantages of the former methods, and show how the new algorithms can be used for modeling chemical kinetics and multiphase reactive transport under partial equilibrium assumptions (e.g., aqueous and gaseous species in equilibrium, minerals controlled by either kinetics or equilibrium). We demonstrate the efficiency of these novel equilibrium methods in large-scale reactive transport simulations of geological storage of CO2, where thousands to millions of equilibrium calculations are performed during each time step. All these methods for equilibrium, kinetics, and reactive transport calculations are available in Reaktoro (www.reaktoro.org), a unified framework for modeling chemically reactive systems.