Two-phase flow behavior in CO₂ geological storage considering spatial parameter heterogeneity

Saline aquifer rocks exhibit significant spatial randomness due to geological sedimentation processes. To address the issue of the heterogeneity of rock formations in numerical simulations, it is common practice to homogenize rock layers with similar lithologies. However, the acceptability of the errors generated during homogenized computations is a major concern and should be investigated. Therefore, to study the influence of heterogeneity at the storage site on the CO2 migration behavior, the Monte Carlo simulation-random finite element method (MCS-RFEM) was combined with a CO2 two-phase flow model to compare the effects of the coefficient of variation (C-v) and correlation length (lambda(x)) of random reservoir permeability fields on the migration distance and extent of CO2 storage under the same mean conditions. The results showed that higher C-v and lambda(x) values significantly reduced the CO2 migration distance while increasing the spread extent. Compared to the homogeneous model, at a lambda(x) value of 100 m, the CO2 migration distance decreased by 5.05%, while the profile sweep area increased by 6.20%. Concurrently, with increasing C-v, the area with a CO2 volume fraction higher than 0.75 decreased by 20.22%, while an increase in lambda(x) resulted in a 42.35% increase in the area with a CO2 volume fraction higher than 0.75. Therefore, reservoirs with high C-v and low lambda(x) values are more suitable for safely storing CO2. (c) 2023 Society of Chemical Industry and John Wiley & Sons, Ltd.