High-Performance Modeling of Rock-Fluid Interactions Driven by Chemical Alterations in Mineral CO2 Storage

Abstract A wide range of geological and geotechnical processes, spanning micro- to macro-scales, are fundamentally governed by dynamic rock-fluid interactions. These interactions are intricately influenced by the coupled effects of rheological transformations and alterations to the rock matrix, driven by metasomatic reactions. Understanding these dynamics is crucial, especially in advancing carbon capture and storage (CCS) technologies. While mafic and ultramafic rocks offer a promising solution through mineral carbonation, securely locking CO2 within stable mineral formations, carbonate reservoirs also present a viable option for long-term CO2 storage. In these reservoirs, CO2 can be trapped through a combination of structural, solubility, and mineral trapping mechanisms, where dissolution and precipitation reactions play a key role in enhancing storage efficiency and sealing potential leakage pathways. However, injecting large volumes of CO2 into carbonate reservoirs can lead to dissolution and rock weakening, potentially inducing the formation of new pathways and increasing the risk of leakage. CO2 injection into mafic and ultramafic rocks can lead to pore clogging due to the formation of solid carbonate minerals, reducing permeability and potentially limiting further CO2 injection. This highlights the importance of accurate modeling to minimize the adverse effects of chemical alterations. Due to the challenges of directly observing these processes in situ or replicating them under laboratory conditions, numerical simulations have emerged as an indispensable tool for understanding and quantifying these phenomena. High-performance numerical simulation has become the primary tool for gaining insights into these complex interactions, enabling us to explore scenarios and mechanisms that would otherwise be inaccessible. Within this work, we employ advanced interdisciplinary approaches to examine the fundamental dynamics of non-isochemical, multicomponent fluid flow in deformable, reactive rock systems. A robust, two-way coupled hydro-mechanical-chemical model is implemented using accelerated numerical algorithms effectively merged with GPU-optimized solvers. The chemical response of the rock matrix involves not only the dissolution and precipitation of salts, carbonates, and clays but also significant alterations in primary mineral paragenesis, such as (de)hydration and (de)carbonation, along with transformations within anhydrous paragenesis. Additionally, the model accounts for the effects of impurities in the CO2 stream, capturing their influence on mineral reactions. These processes are accompanied by dynamic modifications in rock density, porosity, and permeability. Such alterations influence fluid transport dynamics. These complex chemical interactions are fully integrated into the coupled hydro-mechanical-chemical model. The proposed solver executes simultaneous calculations for hundreds of minerals and species, providing real-time insights into dynamic mineral phase distributions without predefined constraints. Numerical results showcase its capability to reproduce natural carbonation processes and assess the effects of CO2 solubility and porosity changes driven by solid density variations resulting from (de)carbonation during fluid infiltration. We demonstrate the model's potential to advance our understanding and predictive capabilities of reactive flow modeling in subsurface systems, offering vital advancements for optimizing CO2 sequestration and addressing critical challenges in climate change mitigation.