Effect of model complexity and relative permeability inputs on geologic carbon sequestration modeling

When modeling geologic carbon sequestration, predicting the saturation of carbon dioxide (CO (sub 2) ) over space and time, the distribution of CO (sub 2) mass over time, and pressure changes over time are primary concerns. Modeling may be done to address short-term concerns such as determining the saturation of CO (sub 2) at the end of the injection period or long-term concerns such as estimating the mass of dissolved CO (sub 2) decades or centuries after injection ceases. Model complexity describes the physics included in the model and encompasses how simply key input data can be described--homogeneous or heterogeneous permeability, homogeneous or heterogeneous capillary pressure and simple or hysteretic relative permeability. Studies have shown that the relative permeability function can significantly control long-term CO (sub 2) distribution. The petrophysical and geological parameters of any specific reservoir are typically uncertain, which motivates studies of parameter sensitivity. Permeability contrast, irreducible gas and water saturation, trapping saturation, and capillary entry pressure will be included in this sensitivity analysis. We assess the impact of changes in model complexity and these basic parameters have on CO (sub 2) mass distribution and injection pressure. The vertical transport of CO (sub 2) , as measured by the mass of CO (sub 2) moving into a caprock or baffle, was significantly reduced as the model became more complex by introducing permeability anisotropy (36.8% to 11.3% of total CO (sub 2) mass) and by introducing capillary pressure to the caprock (36.8% to 4.8% of total CO (sub 2) mass). While both are significant, the modeling results indicate that adding capillary pressure to the caprock was the dominant factor. At the end of the simulation period (150 years), CO (sub 2) was present as a supercritical fluid or dissolved in the brine. The amount of CO (sub 2) dissolved in the brine was significantly lower after adding capillary pressure (23.1% to 14.2%) or increasing the residual gas saturation (23.1% to 16.3%). The amount of CO (sub 2) dissolved in the brine was slightly higher after converting from homogeneous to heterogeneous geology (23.1% to 28.2%). Changes in injection pressure were observed when the permeability was reduced by adding permeability anisotropy but was increased by introducing heterogeneous geology or by changes in relative permeability.