Uncertainty analysis of impact indicators for the integrity of combined caprock during CO2 geosequestration

CO2 capture and geologic sequestration represent one of the most promising options for reducing atmospheric emissions of CO2. Operator choices and subsurface uncertainties each have significant effects on avoiding the risk of triggering fault failure or intact rock failure that could lead to CO2 leakage. With the application of response surface methodology, a Box–Behnken design of experiments was utilized to evaluate the statistical significance of impact indicators, and uncertainty analysis was performed using a robust model to forecast the optimal combination of impact indicators for maintaining caprock integrity during CO2 injection in saline aquifers of the Ordos Basin, China. The Coulomb failure stress and vertical deformation of 130 statistically designed combinations were experimentally determined using numerical simulation. Then, according to the analysis of variance conducted for the response surface quadratic model, the burying depth, friction angle, Young's modulus, Poisson's ratio, and stress field, as well as quadratic terms such as burying depth ∗ friction angle and Poisson's ratio ∗ Poisson's ratio, were found to be highly significant with respect to the level of Coulomb failure stress. Among the variables, only burying depth was highly significant with respect to the level of vertical deformation. Furthermore, the results were analyzed and an optimal combination of impact indicators for Coulomb failure stress was determined according to the minimum Coulomb failure stress. The optimum combination had a desirability value of 1.000, which demonstrated the fitness of the selected statistical model in analyzing the experimental data.