Capillary Pressure Sensitivity in Modeling CO2 Injection

Abstract

Carbon dioxide (CO2) sequestration consists in injecting and storage the gas into a geologic formation as a means of mitigating the greenhouse effect. Among the storage sites, saline aquifers are very promising because of their large capacity and wide availability. In this work, we perform a sensitivity analysis on the response of a flow simulator to study the impact of capillary pressure in CO2 injection and storage. Capillary pressure is represented as a exponential function of CO2 saturation. The simultaneous flow of CO2 and brine in porous media is based on the well known Black-Oil formulation, applied to two-phase fluid flow. It considers that CO2 may dissolve in the brine but the brine is not allowed to vaporize into the CO2 phase. This formulation uses the PVT data as a simplified thermodynamic model. The numerical solution is obtained applying the public-domain software BOAST, which solves the differential equations by finite differences with the IMPES algorithm. Besides, we build a suitable geological model based on fractal porosity and clay content and taking into account the variation of properties with pore pressure and CO2 saturation. This model also includes embedded mudstone layers of very low permeability that accumulate CO2 but also allow its upward migration. Relative permeability and capillary pressure as functions of CO2 saturation are taken from published data, using the Leverett J-function to scale capillary pressure values. The numerical results show that the fluid simulator is able to represent the CO2 injection and storage. As injection proceeds, part of the injected fluid migrates upwards across the mudstone layers. The simulations are performed using different capillary pressure curves. When capillary pressure is higher, CO2 upward migration is slower and, consequently, different zones of accumulations below the layers and high CO2 saturations levels between layers are obtained. Thus, a precise estimation of the capillary pressure function is needed to perfom realistic long term predictions of the spatial-time CO2 distribution in the underground.