Impact of the vertical resolution of corner-point grids on CO2 plume migration predictions for layered aquifers

• Published in: International journal of greenhouse gas control 2021-03, Vol.106, Article 103249
• Main Authors: Youssef, AbdAllah A., Tran, L.K., Matthäi, S.K.
• Format: Article
• Language: English
• Subjects: CO2 geo-sequestration; Discretization error; Grid resolution analysis; Plume extent; Plume spreading; Vertical grid resolution
• ISSN: 1750-5836; 1878-0148
• DOI: 10.1016/j.ijggc.2021.103249

•The dependence of plume spreading velocity on vertical grid resolution requires sub-meter vertical resolution for exact plume extent prediction.•Due to gravity override, slow injection leads to a significant wider spread of the plume, which is suppressed if vertical grid resolution is low.•The piecewise linear as opposed to piecewise constant discretization of saturation has a much-reduced dependence on vertical grid resolution.•Highly resolved simulations reveal a counter-intuitive influence of capillary forces, retarding plume spreading by saturation homogenization. Design of subsurface CO2 storage sites largely relies on numeric simulation-based predictions of plume extent and progressive immobilization. In most cases, sensitivity analyses are performed with corner-point grid representations of the geo-model and first-order IFD methods using two-point flux approximation (TPFA). Here, we have conducted a comprehensive analysis of the impact of the vertical resolution of such grids on the predicted plume extent and capacity in a simplified layered aquifer system. Four different CO2 mobility and buoyancy scenarios were analyzed. To minimize grid-orientation effects, the analysis was performed for predominantly grid-axes parallel flow through regularly gridded cross-sectional models with uniform cell size and variable cell width-over-height ratios. The analysis of the role of vertical grid resolution indicates a first-order correlation between plume extent and this parameter. The results also reveal that capillary forces reduce plume extent and enhance aquifer storage for low permeability contrasts between layers. Furthermore, model sensitivity to grid resolution scales with the magnitude of the permeability contrast between layers. Inspection of these results reveals that an underestimation of CO2 mobility at the top of the plume is the root cause of the observed plume retardation. A comparison with two alternative simulators that discretize mobility as piecewise linear within cells as opposed to piecewise constant and are less resolution sensitive confirms this interpretation.