Abstract
Summary
Steady-state three-phase gas/oil/brine relative permeabilities were measured in a carbonate core under CO2 flooding permeabilities were measured in a carbonate core under CO2 flooding conditions. Results show that the relative permeability of each phase depends only on the saturation of that phase instead of on phase depends only on the saturation of that phase instead of on two saturations, as many previous studies have concluded. All previously reported gas/oil/brine relative permeability studies previously reported gas/oil/brinerelative permeability studies have been conducted with low-pressure N2 gas or air. In this work, CO2 gas, oil, and brine were injected into a carbonate core at 71 degrees C and 9.65 MPa so that the phase behavior and flow would be similar to reservoir conditions. Results show that significant differences exist between the three-phase gas/oil/brine relative permeabilities measured when the gas is CO2 and those measured permeabilities measured when the gas is CO2 and those measured when the gas was N2.
Introduction
Three-phase relative permeability relations are needed for the design of CO2field projects; for accurate prediction, through numerical reservoir simulation, of CO2 flood performance; and for modeling of production and injection problems. The literature contains empirical, mechanistic, and pore-level models used to predict the relative permeability relationships when two phases are predict the relative permeability relationships when two phases are flowing simultaneously in a porous, permeable medium. Because of the limited amount of consistent experimental data available to determine the model parameters accurately, only the simpler models are usually considered. The typical description given when one extends one of these models to three-phase gas/oil/brine flow assumes that one liquid phase strongly wets the rock matrix, the gas phase is "totally nonwetting," and the second liquid phase is of "intermediate wettability." In these cases, one also assumes that the relative permeability of the wetting and totally nonwetting phases depends only on their respective saturations. One then applies the respective two-phaserelative permeability relation as the three-phase relation for these two phases and need only derive an expression for the relative permeability of the intermediate-wetting phase (kro for a strongly water-wet medium in three-phasegas/oil/brine phase (kro for a strongly water-wet medium in three-phasegas/oil/brine flow). The models for kro most frequently applied include the Corey et al., Naar and Wygal, Land, and Stone models. More recent models or significant model modifications include work by Parmeswar and Maerefat, Fayers, Parker et al., Aleman and Slattery, Baker, and Delshad and Pope. Manjnath and Honarpour and Parmeswar and Maerefat reviewed the models, and Baker and Delshadand Pope recently provided detailed comparisons. Experimental studies of three-phase relative permeability (water/oil/gas) were reported as early as1941 and have continued to trickle into the literature. Oak et al. and Maloney et al. reviewed these experimental studies in detail. Oak et al, presented a very well-documented experimental study resulting from presented a very well-documented experimental study resulting from painstaking attention to procedural detail; their results, along painstaking attention to procedural detail; their results, along with those of Schneider and Owens, are analyzed in more detail below. Maloney et al. (with preliminary work reported by Parmeswar et al.) presented what they described as viscosity Parmeswar et al.) presented what they described as viscosity effects on three-phase relative permeability but, while acknowledging the obvious importance of saturation history, made no mention of it with regard to their own experiments. Maini et al. 18 reported three-phase relative permeabilities measured at 100 degrees C and 3.45 MPa forN2/mineral oil/distilled water in Ottawa sandpacks. Although a modest number of studies have been done in which three-phase relative permeability data have been reported, there is no universally accepted conclusion as to the shape of the isoperms when the data are plotted on a saturation ternary; in fact, oil and gas isoperms are reported as concave, linear, and convex toward the respective apex on the saturation ternary, and brine isoperms are given as concave and linear. In addition, a significant number of the experiments were performed under unsteady-state flow conditions. The interested researcher should evaluate these results with regard to saturation hysteresis, viscous instability, and experimental methods used. What separates our three-phaserelative permeability study from previous studies is our use of CO2 gas instead of air or N2, resulting in phase properties consistent with those observed in afield CO2 flood. The experiments were performed under steady-state conditions, minimizing instability phenomena and allowing for control of saturation history. Saturations were determined by tracer injection.
Experimental
A single 5 × 45-cm dolomite core was used for these experiments. The core was cut from a block of outcrop dolomite quarried from the Guelph formation in Sandusky County, OH. Commonly called Baker dolomite, the Guelph is a primary, sedimentary grainstone dolomite with intergranular porosity described in some detail by Meister. Its granular matrix gives it a somewhat homogeneous appearance, although there are zones of varying permeability throughout the core, which are evident by detailed visual inspection as well as by X-ray computed-tomography scanner analysis. Permeability measurements of 2.54-cm core plugs showed that permeability varied by as much as a factor of two along the core length. The core PV was found to be 168.4 cm3 at 10.3-MPa overburden,3.45-MPa core pressure (24 degrees C), yielding a porosity of 0.202. The overall brine permeability measured over the total core length (Sw = 1) was 24md. Three fluids were pumped through the core: 0.020-kg/kg CaCl2 brine, n-decane, and CO2. Phase properties were estimated with phase-behavior data according to Dria's methods. From these, the phase-behavior data according toDria's methods. From these, the viscosities, flow rates, and fractional flows were calculated at the mean core pressure. The equipment used in this experimental investigations was arranged to achieve steady-state, nonrecirculating three-phase water/oil/gas flow through the core at a temperature of 71 degrees C and average core pressures of about 9.65 MPa. The once-through flow design was necessary for continuous collection and analysis of effluent needed for determination of tracer response. Flows through the core were controlled through constant injection rate coupled with downstream pressure control. Saturations in this study were attained through steady-state injection of brine, decane, and CO2 into the core until constant pressure drops over each core section (measured through four pressure drops over each core section (measured through four internal pressure ports), constant pressure drop measured between the inlet and outlet core faces, and constant effluent flow rates were observed. Steady-state overall pressure drops for the various experiments ranged from 35 to 480 kPa, with most between 205 and 345 kPa. The three-phase flow rates were designed to maintain total fluxes, u=q/A, of about0.12 m/d.
SPERE
P. 143
Publisher
Society of Petroleum Engineers (SPE)
Subject
Process Chemistry and Technology