Jones, Stanley C., SPE, Marathon Oil Co.
Abstract
Displacements were conducted in Berea cores to gain insight into the mechanism of tertiary oil displacement and propagation by a micellar slug. Contrary to expectation, propagation by a micellar slug. Contrary to expectation, the first oil mobilized by micellar fluid was among the first oil (instead of the last oil) to be produced, giving the appearance of either viscous fingering or of unusually large dispersion. To eliminate the possibility of unfavorable mobility ratios caused by oil/water/surfactant interaction, we conducted several runs in which an injected hydrocarbon displaced another hydrocarbon, initially at residual saturation. In other experiments, water (the wetting phase) at irreducible saturation was displaced by a distinguishable injected aqueous phase. Injected hydrocarbon appeared in the produced fluids immediately after oil breakthrough, yielding behavior similar to the micellar-slug experiments. Even with a favorable viscosity ratio of less than 0.01, the apparent dispersion was huge. However, mixing zones in the wetting-phase displacements were quite normal and similar to those observed for single-phase flow. Nonwetting-phase fronts (injected hydrocarbon displacing resident hydrocarbon) are smeared much more than wetting-phase fronts because the entrance of hydrocarbon into smaller water-filled pore throats is delayed until the capillary entrance pressure is overcome by differences in the flowing oil and water pressure gradients. Oil might not be displaced from the smaller pores until long after oil breakthrough. Nonwetting-phase dispersion, which occurs in many EOR processes, can be expected to be one or two orders of magnitude greater than dispersion measured in single-phase-flow experiments. Entrance of the wetting phase, however, is not delayed; hence, wetting-phase Mixing zones are short.
Introduction
Experiments for this study were inspired by the question: How is residual oil, which has been mobilized by a micellar slug, transported? More specifically, does the first oil mobilized by a slug (near the injection end of a core) contact and mobilize oil downstream from it, which displaces more oil even farther downstream? If this were the case, the first oil to be produced would be the most-downstream oil (i.e., oil nearest the outlet). The last oil produced would be the first oil mobilized from the produced would be the first oil mobilized from the injection end of the core.
This scheme is somewhat analogous to pushing a broom across a floor covered with a heavy layer of dust. The first dust encountered by the broom stays next to the broom. As the accumulated layer of dust in front of the broom becomes adequately compacted, it pushes dust ahead of it to from an ever-widening band or "dust bank" ahead of the broom. The dust farthest ahead of the broom is the first to be pushed into the dustpan, and the dust first encountered by the broom is the last to be pushed in. Or is this concept all wrong? Another model postulates that the oil first contacted by a micellar slug is mobilized and quickly travels away from the slug so that the downstream oil is contacted and mobilized by the slug, not by the first-mobilized oil. If this process were to proceed to its logical conclusion, the first-produced oil would proceed to its logical conclusion, the first-produced oil would be from the inlet end of the core, and the last-produced from the outlet end. Either of these two extremes would be modified by dispersion, which smears sharp fronts by mixing displaced and displacing fluids. Dispersion in porous media has been investigated extensively. Perkins and Johnston have reviewed several studies, mostly involving single-phase flow. The simultaneous injection of the water with light hydrocarbon solvents is a technique used to reduce solvent mobility and viscous fingering. Raimondi et al. performed steady-state experiments in which flowing performed steady-state experiments in which flowing water and oil were miscibly displaced by the simultaneous injection of water and solvent. They found that the longitudinal mixing coefficient for the hydrocarbon phase increased sharply with increasing water above the irreducible saturation. The displacement of the wetting phase was not greatly affected by the presence of the nonwetting phase. However, a large amount of oil that initially phase. However, a large amount of oil that initially seemed to be trapped by water was eventually recovered by continued solvent injection. Raimondi and Torcaso later found that some oil, particularly at high water-to-solvent injection ratios, was particularly at high water-to-solvent injection ratios, was trapped permanently, provided that injection rates, ratios, and pressure drops were unchanged in switching from water/oil to water/solvent injection. Fitzgerald and Nielsen also found that only part of the in-place crude was recovered by solvent injection. Moreover, solvent appeared in the effluent shortly after oil breakthrough. Oil recovery was further decreased when solvent and water were injected simultaneously. Thomas et al. reported slightly increased wetting-phase longitudinal mixing during simultaneous water/oil injection as the wetting-phase saturation decreased. Non-wetting-phase mixing increased substantially as the nonwetting-phase saturation decreased from 100%.
SPEJ
p. 101
Improved pore network models to simulate single-phase flow in porous media by coupling with lattice Boltzmann method