Modeling of fluid-rock interactions during storage of CO2 in geological reservoirs

The injection of CO₂ into geological reservoirs will cause a shift in the geochemical equilibrium between pore fluid and the rock matrix, which leads to a series of reactions between the injected CO₂, the fluid and the minerals exposed to the pore fluid. Generally, fluid-rock-reactions are kinetically controlled and interact with flow and transport processes in a complex way, e.g. by enlarging / reducing effective porosities and fluid-permeabilities due to mineral dissolution / precipitation, which may lead to increased fluid flow and transport rates or a partial sealing or clogging of fluid migration pathways. An accurate numerical simulation and prediction of these process interactions and their effects on storage of CO₂ requires the representation of the kinetic limitations in the model, especially, if reaction time scales significantly exceed transport time scales. For this purpose the multiphase flow and reactive transport code OpenGeoSys-ChemApp (OGS-CA) has been extended to take into account the kinetic nature of fluid/mineral interactions and the resulting feedbacks on flow and transport from porosity changes. The code was verified for the correctness and accuracy of the implemented kinetic reaction module by comparison against other software and analytical solutions. OGS-CA was then applied to study the mineral trapping of CO₂ following an injection into two different types of geological reservoirs in Northern Germany: a deep saline storage formation in an anticline structure intended for CO₂ storage (A) and an almost depleted gas field intended for enhanced gas recovery (B). In both scenarios, geochemical reactions proceed for more than 10.000 years. Batch model simulations of the saline formation (A) predict CO₂ mineral trapping to occur mainly as dawsonite, although dawsonite precipitation proceeds at a very low rate. After more than 10.000 years, the kinetic model is still far from the state predicted by thermodynamic equilibrium modeling. Changes in porosity are less than -1%. In the depleted gas reservoir (B), model simulations predict the mineral trapping of CO₂ as calcite and no precipitation of dawsonite. Reductions in reservoir porosity are also small, but reach up to -10%. On the long term, the kinetic simulations approach a state predicted by equilibrium geochemical modeling. One-dimensional simulations of diffusive CO₂ spreading in the formation brine show only a small influence of the transport of dissolved components at early times. The long term geochemical development in case (B) hence can be approximated reasonably well by kinetic batch modeling.