Computational Exploration of Thermodynamic Models of Geological CO₂ Injection

Abstract

This thesis investigates the behavior of carbon dioxide flow in porous media through high-fidelity computational modeling, with a specific focus on the impact of the Span-Wagner equation of state (EOS). Accurate modeling of CO₂ transport in subsurface environments is essential for applications such as carbon capture and storage (CCS). We model the entire flow from injection, down throughout a vertical pipe and into a porous reservoir. To this end, we utilize the MOOSE (Multiphysics Object-Oriented Simulation Environment) framework developed by Idaho National Laboratory to perform finite element simulations. A key contribution of this work is the successful coupling of a porous rock domain with a one-dimensional pipe flow simulation in Julia, enabling a broader representation of injection scenarios. The study examines how the thermodynamic accuracy of the Span-Wagner Equation of State influences flow characteristics, in comparison to the Ideal Gas Equation of State. Through a series of coupled pipe-reservoir simulations, we assess variations in pressure and density as CO₂ is injected from the pipe into the porous medium. The model can detect phase change conditions, allowing us to predict the maximum mass flux that can be achieved below the liquefaction threshold, as defined by the binodal curve in the CO₂ phase diagram at a given temperature. The results highlight the importance of EOS selection in predicting multiphase flow behavior, especially under conditions relevant to geological storage. Furthermore, we find that the Ideal Gas EOS underpredicts injection rates under the same conditions. This integrated modeling approach advances the understanding of thermodynamic effects in coupled subsurface flow systems and supports the development of reliable tools for large-scale carbon storage applications.