In this study we used CO2 core-flood experimental data to develop and calibrate a three-dimensional continuum-scale reactive transport model for simulating CO2-induced rock porosity and permeability evolution. During the core flooding experiments the CO2-equilibrated brine was injected into carbonate rock samples, which were collected from the Arbuckle carbonate reservoir, Wellington, Kansas. We analyzed X-ray computed microtomography (XCMT) characterization data to examine heterogeneous distributions of pore structures, naturally occurring fractures, and mineral phases inside the cores, and then used them to constrain initial model macroscopic properties including porosity, permeability, and mineral compositions. Model chemical reaction kinetics, porosity-permeability, and porosity-surface area correlations were also calibrated with the experimental data. Our model results indicate that the calibrated continuum-scale reactive transport models are able to adequately predict the dissolution behavior of fractured carbonate rocks at a core scale. In addition the model calibration against core flood experiments provides a useful basis for upscaling reactive transport properties used to describe carbonate dissolution from the laboratory (core) scale to field (reservoir) scale.
Understanding the effect of rock-water interactions on rock porosity and permeability under CO2 sequestration and enhanced oil recovery (EOR) conditions is an active area of research [e.g. 1-9]. To this purpose, there is a need to develop reliable reactive transport models, which can accurately represent essential features of CO2-fluid-rock interactions at laboratory and field scales. This is challenging, because adaptation of pore scale models needed to describe fluid-mineral geochemical interactions is limited in scale by their high computational cost. For this reason a macro-scale continuum model based on Darcy flow approximations is often used to simulate reactive transport processes, and evaluate geochemical and transport impacts of rock-fluid interactions at large scales. An important challenge in modeling reactive transport and mineral chemical alteration processes, particularly in a heterogeneous and fractured rock system, is to constrain effective macroscopic parameters (e.g. porosity-permeability relationships, chemical rate kinetics, porosity-reactive-surface-area correlations) required by the continuum-scale models. Model predictions made in the absence of proper calibration or upscaling of these governing transport and reaction parameters are highly uncertain. A number of studies [e.g. 6-12] have suggested that the use of either pore-scale simulations or experimental data to calibrate the continuum-scale models could help reduce such uncertainty.