Statistical based hydromechanical models to estimate poroelastic effects of CO2 injection into a closed reservoir

In this study, a statistical‐based methodology is used to evaluate the poroelastic effects of injection during CO2 geologic sequestration in a closed reservoir. We constructed a series of hydromechanical models to evaluate the poroelastic effect of injection by quantifying total stress changes inside the reservoir, reservoir displacement, and surface uplift. The models are representative of closed carbonate reef reservoirs of the Michigan basin. A combination of experimental design for seven independent parameters (depth, caprock and reservoir Young's modulus, caprock and reservoir Poisson's ratio, pressure, and Biot's coefficient) and response surface modeling was used to develop statistical‐based reduced‐order models. We performed 147 numerical simulations to develop simplified models. The poroelastic model responses were captured using a standard quadratic model with full interaction terms, as well as a reduced model with only statistically significant coefficients. The interpretation of reduced‐order models, using R2 loss method, shows that while surface uplift depends mainly on the depth of the reservoir, stress increase depends mainly on Biot's coefficient. The developed statistical‐based models provide a quick tool to evaluate the poroelastic effect of injection into the closed carbonate reservoirs of the Michigan basin. Reduced‐order models were then combined with a Monte Carlo simulation to perform poroelastic uncertainty analyses and achieve a better understanding of the poroelastic performance of CO2 storage in the closed carbonate system of the Michigan basin. © 2020 Society of Chemical Industry and John Wiley & Sons, Ltd.