Large-scale models for CO2 injection, migration, and leakage

The Celia Group continues to develop a suite of models to simulate CO2 injection, migration, and potential leakage, with applications to a number of potential injection sites. These models predict pressure buildup in the formation, movement of both CO2 and brine within the injection formation and across confining units into other formations, capillary trapping of CO2, and dissolution of CO2 into the brine phase (that is, solubility trapping). The models are developed in a multi-scale framework so that scaling issues in both space and time are addressed explicitly and large-scale simulations can be performed easily.

To show how these models behave, the researchers have systematically compared the models, which are based on the assumption of vertical equilibrium, with full three-dimensional simulations using the industry standard Eclipse. For most formations suitable for CO2 injection there are excellent matches in results (Figure 6).

Figure 6. Comparison of radial saturation profiles from Eclipse (top row) and vertically integrated simulations (bottom row) using different combinations of saturation‐relative permeability and saturation‐capillary pressure relationships. The first column use values from a study by Ebigbo et al. (2007), the second one is from Pruess et al. (2002), the third and fourth are from Zhou et al. (2008, 2010) . Figure from Court et al. (2012a).

Another recent study shows how local capillary pressure effects, which lead to a capillary transition zone in the vertical saturation profile, can impact the development and migration of a CO2 plume. An application to the Johannson formation, under the North Sea, shows the relative importance of the capillary transition zone, convective mixing associated with large-scale dissolution, and buoyant migration. Finally, a recent study focused on leakage along old wells shows how the estimated leakage rates for a field in the Alberta Basin can be related to the statistical properties used to define the permeability of the wells in the field. Such a relationship allows both operators and regulators to determine the range of statistical properties required in order to reach a specified target threshold for leakage.


Basin-scale modeling including active reservoir management

Celia and colleagues have applied their modeling tools to several sedimentary basins that have practical importance, including the Illinois Basin, the Michigan Basin, and the Alberta Basin. In the Illinois Basin, they have focused on injection into the Mount Simon formation, considering scenarios in which most of the emissions from power plants and ethanol plants within the basin are captured and injected. This leads to injection of about 140 Mt CO2/yr.

Simulation results show that such a large-scale injection strategy is feasible. Without any spatial optimization of the injection operations, the CO2 plumes all remain relatively small while the pressure pulse expands across most of the basin (see Figure 7, left and middle panels). None of the injection well pressures exceeds estimated fracture pressures, but the spatial extent of pressures that exceed the threshold pressure defining the Area of Review is large. Both the large spatial extent of the pressure footprint and the overlapping footprints among neighboring injection operations motivates an expanded strategy to control the spatial extent of the pressure field.

The researchers have also examined the impacts of brine production for the Illinois Basin, as shown in the right panel of below (Figure 7). Note that extraction of brine volumes similar to the volume of injected CO2 isolates individual injection operations and greatly reduces the overall areas of review. The challenge now is to find beneficial uses for the extracted fluid, which leads to broad considerations of geothermal energy, water management and overall resource management and utilization. Studies similar to those being pursued for the Illinois Basin are ongoing for the Michigan and Alberta Basins.

Figure 7. Results from full‐scale Illinois Basin simulations. Left image shows CO2 plumes, middle image is pressure increase with no active management, right image is pressure increase with active management of pressure.