| Modelling of CO2-H2O mixture flows in a porous media under subcritical conditions remains a challenging issue for carbon sequestration and possible leakage scenarios. Currently, there is no widely used and generally accepted numerical model that can simulate three-phase flows with both gaseous and liquid CO2-rich phases. We propose a new compositional modelling approach for sub- and supercritical three-phase flows of water, liquid CO2 and gaseous CO2. The new approach is based on the calculation of the thermodynamic potential of the mixture as a function of pressure, total enthalpy and mixture composition and storing it values as a spline table, which is then used for the hydrodynamic simulation. A three-parametric generalisation of the Peng-Robinson equation of state is used to fit the experimental data on CO2-H2O mixture properties.
Using the developed approach, we assess several sample problems of CO2 injection in shallow reservoirs for the purpose of testing the model. We provide the simulation results for a simple 1D problem with a homogeneous reservoir and for a more complicated 2D problem with a highly heterogeneous reservoir using data from the 10th SPE comparative project reservoir. We analyse the temperature variations in the reservoir due to the dissolution of CO2 in water and the evaporation of liquid CO2 under subcritical conditions. The interplay of these processes results in a complicated non-monotonic temperature distribution. At different distances from the CO2 injection point, the temperature can either decrease or increase with respect to the reservoir temperature before injection. The main phenomenon responsible for the considerable temperature decline around the CO2 injection point is the liquid CO2 evaporation process.
We also consider parallel simulations of supercritical CO2 plume evolution at Johansen formation. Firstly, we consider a test scenario using a simplified geological model. Both the free CO2 phase saturation and the integral parameters of the CO2 distribution predicted by our approach are in a good agreement with commonly used simulators TOUGH2 and ECLIPSE. Secondly, we consider the plume evolution using a complicated geological model (NPD5 model). For better resolution of the plume, we use local grid refinements. We simulate 100 years of CO2 injection. The total injection volume is 400 Mt of CO2. After the injection is stopped, we continue the simulation to evaluate the preferable directions of CO2 migration in the formation. We analyse the influence of different trapping effects on the storage. Immediately after the well is stopped, the main part of the injected CO2 is in free mobile phase. Afterward, while CO2 is migrating upward, the role of residual and solubility trapping is rising. We evaluate an optimized location of the injection well that makes possible a more complete use of anticlines for structural trapping. The optimized location of the well makes possible a more effective trapping of the injected CO2 in the formation.
This work was supported financially by the Russian Foundation for Basic Research (Project No. 12-01-31117) and a Grant from the President of the Russian Federation for the Support of the Young Scientists and the Leading Scientific Schools (Project No. SP-2222.2012.5, NSh-1303.2012.1). |