An Investigation of Phase Behavior-Macroscopic Bypassing Interaction in CO2 Flooding

Abstract

Abstract CO2-crude coreflood experiments and high-resolution 2-D CO2-crude displacement simulations in which viscous fingering is represented explicitly suggest that there is a synergistic interaction between multiple-contact CO2-crude phase behavior and macroscopic bypassing that causes the "ultimate" oil recovery (when the system has been swept) to be lower in the unstable case than in the stable case. Assuming this effect is present in field applications of CO2 flooding, then corefloods in which fingering is absent, for whatever reason, should not be used as direct indicators of field-scale displacement efficiency since they will yield optimistic predictions, all other factors being equal between the laboratory and the field. Introduction This paper presents results from one phase of a systematic Shell investigation aimed towards understanding mechanisms of viscous-dominated displacements of waterflood-residual, light, undersaturated crude oils by CO2 flooding (Fig. 1). A previous paper by Gardner, Orr, and Patel dealt with stable, paper by Gardner, Orr, and Patel dealt with stable, secondary (no mobile water) displacements of recombined Wasson crude. The present paper is concerned with unstable, secondary displacements of that same oil. In the previous study, potentially significant factors investigated were CO2-crude oil phase behavior, longitudinal dispersion, and relative permeability (as it turned out, relative permeability was of negligible importance). Bypassing was excluded. With the present work, we bring that factor into the picture. In particular, the added impact of macroscopic bypassing in the form of viscous fingering is examined. It should be noted, as indicated in Fig. 1, that neither the previous nor the present study addresses the effect of mobile water, examined on a separate branch of our overall program. The fact that gravity is neglected has already been implied by use of the term "viscous-dominated". Also excluded are the effects of severe heterogeneity Finally, the present investigation deals specifically only with what happens at a single pressure above the critical point on a pressure-CO2 concentration diagram. This in turn means above the so-called "minimum miscibility pressure", which cannot be greater than the pressure at the critical point. Nonetheless, we feel the results documented here provide valuable insight of a general nature into provide valuable insight of a general nature into the role of viscous instabilities in vaporizing gas drive type processes. They also bring up some important considerations involving finite lateral boundaries that should be borne in mind when designing and/or interpreting CO2 coreflood experiments. Both experimental and theoretical results are documented in this paper. The experimental results come from CO2-crude and first-contact miscible CO2-Soltrol TM corefloods. (Soltrol TM, or more specifically Soltrol TM 130, is a bottoms product of an alkylation unit and is manufactured by Phillips Petroleum Company. It has a normal boiling point range equivalent to that of C11-C14, and is composed of 99.9% t-butyl groups.) The major theoretical results reported in the paper are from high resolution, two-dimensional simulations of unstable CO2-crude and first-contact miscible displacements in which viscous fingering is represented explicitly. These simulations play a key role in interpreting the experiments. EXPERIMENTAL BASIS OF INVESTIGATION Secondary CO2-Wasson Crude and CO2-Soltrol TM Displacements in Berea Sandstone, Plus 1-D CO2-Wasson Crude Simulation; Phase Behavior-Bypassing Synergism Three sets of experimental data from the basis of the investigation documented here. First, data from CO2-Wasson crude and CO2-Soltrol TM secondary, viscous-dominated corefloods carried out at Shell, in conjunction with 1-D CO2-Wasson crude simulation results, suggest that the combination of multiple-contact CO2-crude phase behavior and macroscopic bypassing reduces not only the rate of oil recovery with throughput, but also the "ultimate" recovery. Evidence of this type is shown in Fig. 2. P. 163