A tertiary miscible hydrocarbon flood is being implemented in Virginia Hills Beaverhill Lake reservoir. The hydrocarbon solvent injected in Virginia Hills, as in many other tertiary recovery schemes, is less viscous and less dense than the fluids in the reservoir. Therefore, it is likely that a significant portion of the reservoir is bypassed due to gravity segregation and viscous fingering. One way to minimize such problems isthe injection of a foam.
This paper investigates the viscous instability effects when solvent and foams of different viscosity are injected. The approach chosen here is to conduct scaled model experiments to study the amount of bypassing due to viscous fingering when the solvent is injected in the tertiary recovery mode. The scaling is not trivial as there are a number of interfaces and the stability or instability of one or more of these interfaces will affect the overall sweep efficiency differently. The application of scaling criteria to a displacement with multiple interfaces such as the tertiary flood in Virginia Hills is discussed.
Experimental results are presented where the areal sweep efficiency is investigated with solvents alone and with thickened solvents which were assumed to represent a stable and homogeneous foam. It has been found that the optimum mobility reduction factor (MRF) of the foam was dependent upon the level of instability. The less unstable displacements reach the optimum recovery at lower MRF values. This means that the surfactant concentration in the foam can be decreased in such areas of the reservoir resulting in cost savings.
The displacement of waterflood residual oil by a hydrocarbon solvent has been conceived as a supplementary process for recovering additional crude oil. In general, the solvent is less dense and less viscous than the fluids present in the formation. Consequently, a small fraction of the reservoir volume may be swept at the time of the solvent breakthrough due to viscous fingering, gravity segregation and reservoir heterogeneity. The discussion will be limited to fluid mechanics aspects in this paper. One method to increase the sweep efficiency is to reduce the solvent mobility by injecting water with it 1. Water injection decreases solvent mobilily by decreasing the relative permeabilily of the reservoir to solvent Because it is not practical to inject simultaneously the solvent and water in the field, they are injected in small alternating slugs (WAG: Water-Alternating-Gas) 2.
Another alternative is to inject the solvent in the form of a foam (or emulsion) which has lower mobility than the solvent itself3. The improvement in recovery due to WAG or foam injection can be estimated either by numerical simulation or by laboratory experiments. When modelling tertiary miscible displacements, it is often not feasible to perform detailed simulations with fine grids to represent solvent channeling through the reservoir. Therefore, simpler models such as the mixing parameter model must be used in coarse grid simulators 4.
In general, the choice of the mixing parameter is an empirical one.
