Summary
In miscible flooding, injection of solvent is often combined with water to reduce the mobility contrast between injected and displaced fluids and control the degree of fingering. Using traditional fractional-flow theory, Stalkup estimated the optimum water-solvent ratio (or WAG ratio) when viscous fingering effects are ignored, by imposing that the solvent and water fronts travel at the same speed. Here we study how the displacement efficiency and the mobility ratio across the solvent front vary with the WAG ratio when fingering is included in the analysis. We do so by computing analytical solutions to a 1D model of two-phase, three-component, first-contact miscible flow that includes the macroscopic effects of viscous fingering. The macroscopic model, originally proposed by Blunt and Christie (1993, 1994), employs an extension of the Koval fingering model to multiphase flows. The premise is that the only parameter of the model—the effective mobility ratio—must be calibrated dynamically until self-consistency is achieved between the input value and the mobility contrast across the solvent front. This model has been extensively validated by means of high-resolution simulations that capture the details of viscous fingering and carefully-designed laboratory experiments.
The results of this paper suggest that, while the prediction of the optimum WAG ratio does not change dramatically by incorporating the effects of viscous fingering, it is beneficial to inject more solvent than estimated by Stalkup's method. We show that, in this case, both the pore volumes injected (PVI) for complete oil recovery and the degree of fingering are minimized.
Introduction
Solvent flooding is a commonly used technology for enhanced oil recovery in hydrocarbon reservoirs, which aims at developing miscibility, thereby mobilizing the residual oil and enhancing the mobility of the hydrocarbon phase (Stalkup 1983; Lake 1989). Despite its high local displacement efficiency, the overall effectiveness of solvent injection may be compromised by viscous fingering, channeling, and gravity override, all of which contribute negatively to sweep efficiency (Christie and Bond 1987; Christie 1989; Christie et al. 1993; Chang et al. 1994; Tchelepi and Orr 1994). In this paper, we focus on the effect of viscous fingering; that is, the instability that occurs when a low-viscosity fluid (solvent) is injected into a formation filled with more viscous fluids (water and oil).
Mobility control of the injected solvent can be achieved by simultaneous coinjection of water—typically in alternating water and solvent slugs (WAG) (Caudle and Dyes 1958). In this way, the mobility contrast between the injected and displaced fluids is reduced, thereby limiting the degree of fingering.
There is an optimum ratio of water to solvent that maximizes recovery—in the sense of minimizing the number of pore volumes injected—while providing effective mobility control. For linear floods in homogeneous media, and without consideration of viscous fingering effects, a graphical construction of the optimum WAG ratio was given by Stalkup (1983) for both secondary floods (water/solvent injection into a medium filled with mobile oil and immobile water) and tertiary floods (water-solvent injection into a medium filled with mobile water and immobile oil). The design condition imposed in Stalkup's method is that the velocity of the water and solvent fronts be the same. Walsh and Lake (1989) performed an interesting analysis of the WAG ratio (the ratio of injected water to solvent) on the displacement efficiency for secondary and tertiary floods, using fractional-flow theory. They did not include the effects of viscous fingering, but they estimated the mobility contrast across the solvent front as a measure of the severity of fingering.