The increasing levels of CO2 in the atmosphere are the main cause of climate change, and this calls for actions to reduce CO2 emissions. In this context, underground geological storage (UGS) of CO2 is deemed as one of the most promising technologies to control CO2 emissions. It is one of the possible final destinations of the carbon capture and storage (CCS) chain. However, its deployment at an industrial scale relies on the development of tools capable of forecasting the long-term behavior of injected CO2, which can be used in assessing the project feasibility and assisting project implementation and monitoring. In this context, robust and reliable mathematical modeling of CO2 UGS by means of numerical simulations is instrumental to the wide acceptance and implementation of the technology.
Injection of CO2 in mineral formations can trigger many physical and chemical phenomena, which makes the task of simulating CO2 UGS extremely challenging. CO2 can be distributed into fluid and mineral phases and many chemical reactions can occur. Simulators capable of giving a sound physical description of CO2 UGS should account for fluid flow in porous media, adsorption of certain components, phase splitting of components, and chemical reactions, among other physical and chemical phenomena. Because of the latter two mentioned, algorithms for performing chemical and phase equilibrium (CPE) calculations are at the center of CO2 UGS simulation.
In this thesis, we provide novel numerical simulation tools applied to CO2 UGS in saline aquifers. The contributions of the thesis are both of theoretical nature, by expanding the general theory of CPE algorithm, and of practical nature, by applying the new findings to CO2 UGS. From the theoretical point of view, new types of CPE calculations were developed. From the practical point of view, the path of certain mineral reactions was studied in scenarios of interest to CO2 UGS, e.g., by using the newly-developed algorithms. In addition, insights on salt clogging during CO2 UGS (risk of injectivity impairment in saline aquifers) were acquired. Finally, a fully implicit compositional simulator capable of simulating both reactive and non-reactive flow in porous media was developed with the RAND algorithm at its core. Overall, the study has provided a basis for using the CPE algorithms in various simulation analyses of CO2 UGS involving multiphase geochemical reactions, and the theoretical methods developed can potentially be applied to other reactive transport processes.