The Influence of Phase Behavior on Surfactant Flooding

This paper analyzes by mathematical modeling the role of phase behavior in surfactant flooding. In the absence of dispersion, miscible, immiscible, and semimiscible displacements are distinguished by the position of the injected composition relative to the position of the injected composition relative to the binodal envelope and extended tie lines. Even with dispersion, these concepts prove useful in analyzing slug miscibility breakdown in surfactant floods. Introduction Two design philosophies of tertiary oil recovery by surfactant flooding exist. In one, the chemical slug is designed to be miscible in some proportions with reservoir oil and brine, the goal being miscible displacement of resident oil. The second philosophy is to attain, rather than miscibility, philosophy is to attain, rather than miscibility, ultralow interfacial tension (IFT) between the slug fluid and resident oil. Correlations obtained by immiscible displacements of oil from natural and artificial porous media show that the saturation of residual oil (i.e. trapped, unrecoverable oil) decreases as IFT decreases. In reality, the distinction between philosophies is a matter of degree. Miscible displacements have regions of immiscibility. (e.g., the oil/brine bank). Furthermore, advocates of miscible displacements concede that breakdown into immiscible displacement occurs in the later stages of their processes; others argue that the breakdown occurs processes; others argue that the breakdown occurs early and that miscible displacements are, by and large, immiscible. On the other hand, since most slug formulations advocated by both schools are single phases capable of absorbing some amount of oil and phases capable of absorbing some amount of oil and brine without splitting into multiple phases, even chemical flood displacements designed to be immiscible are miscible for some time, however short. A related area of contention concerns the alleged advantages or disadvantages of formulating oil-rich, as opposed to brine-rich, slugs. Another area of contention concerns whether small, high-concentration chemical slugs are preferred to larger, lower-concentration slugs. The purpose of this paper is to shed light on these questions by paper is to shed light on these questions by incorporating equilibrium phase concepts as represented on a ternary diagram into the simulation of surfactant flood displacements. This study indicates that immiscible and miscible displacements are, in fact, closely related. Specifically, miscible recovery of oil is enhanced if the multiphase region of the ternary diagram contains a substantial subregion of ultralow tension. Furthermore, the success of miscible displacements is affected strongly not only by the position of the slug composition relative to the multiphase envelope on a ternary diagram but also by the position of slug composition relative to the tie lines, with better oil recovery attained when the injected composition point lies away from the region through which point lies away from the region through which extended tie lines pass. Thus, this study stresses the importance of the partition coefficient, a parameter shown to be important in an earlier study. For the purpose of this study, two simulation techniques for three-component, one- and two-phase flow in porous media were developed, each with its own restrictions. The first, a method-of-characteristics scheme (extended from a method developed earlier) allows phase volumes to change by solubilization of components phase volumes to change by solubilization of components but considers only continuous injection of micellar fluid, not the more realistic slug injection. The second method is a finite-difference approach that handles slug injection and solubilization and builds in dispersion, which cannot be considered when the method of characteristics is used. Because of the large number of parameters that arise in this study, base-case values (Table 1) of all parameters have been selected. For all results given in parameters have been selected. For all results given in this paper, the value of each parameter is the base-case value, unless otherwise specified. SPEJ P. 411