Influence of Geological Parameters on Co2 Storage Prediction in Deep Saline Aquifer at Industrial Scale

Here we examine the consequences of uncertainty with respect to geological parameters, using large-scale 2D models. We also investigate ways to reduce prediction uncertainty, either by showing how a parameter’s influence is negligible for a given project design, or by showing for which parameters additional data will significantly increase the quality of prediction. TOUGH2/ECO2N is used to simulate the injection of millions of tonnes of CO2 for the specific case of the Dogger Aquifer (a carbonate aquifer in the Paris Basin), which has substantial lateral and vertical heterogeneities and few associated data. The parameters of interest are spatial variability and correlation length of permeability, absolute permeability, pore compressibility, cap rock permeability, and relative permeability curves. Several numerical models of permeability are constructed: two uniform cases (two values of permeability) and 200 geostatistical initial realizations, which are modified according to the studied parameters. Results are compared in terms of propagation of pressure perturbations, injectivity (pressure in the vicinity of the well), and gas migration and dissolution. The results indicate that: (1) The pore compressibility, and the absolute value and spatial variability of permeability have the strongest influence on pressure propagation and injectivity. Relative permeability curves and correlation lengths have a weaker influence at the peak of pressure, but tend to increase the variations in maximum/minimum cases. (2) Relative permeability curves and heterogeneities have a significant impact on prediction of gas dissolution and migration. Finally, we also investigate the possibility of reducing the number of simulations. INTRODUCTION In terms of available volume, deep saline aquifers seem suitable for CO2 storage projects at industrial scale, yet the perturbations resulting from the injection of millions of tonnes of CO2 will constrain the storage capacity of these aquifers. The extent and intensity of these perturbations, as well as CO2 plume migration, will depend on the geological characteristics of the reservoir. However, the scarcity of data related to these aquifers leads to uncertain predictions of realistic injectivity and storage capacity. At the regional scale, several storage projects may be involved in the same groundwater system. Therefore the quality of predictions in large-scale studies is critical to insure the feasibility of these projects. Several studies at basin scale, such as Nicot (2008), Birkholzer at al. (2009), Zhou et al. (2010), Person et al. (2010), have shown that pressure buildup due to massive injection may extend far from the injection point, causing issues in terms of capacity (e.g., pressure interferences between neighboring sites) and posing risks for natural groundwater resources. Some of these studies have also described the influence of pore compressibility, reservoir permeability, and cap rock permeability on pressure-perturbation results at large scale. Depending on the specific study and model used—especially for homogeneous or layered models—boundary conditions (open/closed systems), permeability, and pore compressibility will have a decisive impact on pressure perturbations, especially compared to other parameters such as porosity, temperature, residual gas saturation, anisotropy, or cap rock permeability—as indicated by Buscheck et al. (2012), Schäfer et al. (2011), Chadwick et al. (2010), Flett et al. (2005). ha l-0 07 77 89 4, v er si on 1 18 J an 2 01 3 Author manuscript, published in "TOUGH Symposium 2012, United States (2012)"