Effect of an Aqueous Phase on CO2/Tetradecane and CO2/Maljamar-Crude-Oil Systems
- N.R. Pollack (U. of Pittsburgh) | R.M. Enick (U. of Pittsburgh) | D.J. Mangone (U. of Pittsburgh) | B.I. Morsi (U. of Pittsburgh)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- May 1988
- Document Type
- Journal Paper
- 533 - 541
- 1988. Society of Petroleum Engineers
- 5.4 Enhanced Recovery, 5.2 Reservoir Fluid Dynamics, 5.4.1 Waterflooding, 4.3.4 Scale, 4.1.2 Separation and Treating, 5.2.2 Fluid Modeling, Equations of State, 5.4.2 Gas Injection Methods, 5.2.1 Phase Behavior and PVT Measurements, 4.1.5 Processing Equipment, 5.3.2 Multiphase Flow
- 0 in the last 30 days
- 281 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
The effect of the presence of an aqueous phase on the phase behavior of the CO2/tetradecane and the CO2/Maljamar-crude-oil systems has been experimentally determined. Both the salinity and the amount of the aqueous phase were varied to test several methods of modeling. In the first technique, the amount of CO2 "lost" to the aqueous phase was determined using Henry's law, decreasing the overall ratio of CO2 to hydrocarbons, while the Peng-Robinson equation of state (PR EOS) was used to determine the phase distribution of the hydrocarbon phases. In the second technique, the equation of state (EOS) was modified to predict the densities and compositions of not only the hydrocarbon phases, but also the aqueous phase. Simply accounting for CO2 solubility by incorporating Henry's law, used in conjunction with the Poynting correction and an empirical factor for salinity, not only gave results nearly identical with the EOS approach, but also required less computational effort. Both methods gave good agreement with the experimental data.
Many investigations have been performed on CO2/alkane and CO2/crude-oil systems to establish the mechanisms involved in gas miscible-displacement processes. The presence of an aqueous phase, however, should also be accounted for. Brine not only exists in the reservoir before any oil production, but also is injected into the formation during waterflooding to maintain high levels of reservoir pressure and to displace oil. Brine is also injected throughout the CO2 flood to reduce the mobility of the CO2.
The effect of an aqueous phase traditionally has been accounted for by predicting the solubility of CO2 in the aqueous phase and subtracting this amount of CO2 from subsequent EOS flash calculations for the hydrocarbon phases. The purpose of this investigation is to determine experimentally the effect of introducing water or brine into a CO2/hydrocarbon system. An alternative method of predicting the observed phase behavior will also be presented. In this technique, an EOS is first modified to permit an accurate prediction of the water or brine density and vapor pressure. Then the mutual solubilities of CO2 and water or brine are fitted by implementing a mixing rule for asymmetric systems. The EOS is then incorporated into a multiple-phase flash-calculation algorithm. The two calculation schemes will be compared to determine which provides a more realistic description of the experimentally observed effects and the computational requirements associated with each.
The experimental apparatus used in this study, Fig. 1, is typical of those used in the generation of p-x diagrams for CO2/oil systems. CO2 is charged directly into the evacuated 420-cm3 Ruska windowed volatile-oil cell. The CO2 may be compressed to several different pressures by use of mercury and the number of moles calculated with the relationship
Pressure is monitored by a Heise gauge and a transducer, volume is determined by measuring the height of the CO2/mercury interface, temperature is kept constant by the insulated oven, and the compressibility factor may be obtained from standard charts or tables.1 A predetermined number of moles of oil is then displaced into the cell by mercury displacement or by use of one of the two Ruska positive-displacement proportioning pumps. After the cell (which now contains a mixture of known overall molar concentration of CO2 and oil) is rocked to accelerate the equilibrium of phases, it is placed in an upright position and the phase volumes are measured with a cathetometer. This procedure is repeated until the hydrocarbon-rich liquid/vapor (L1/V) bubblepoint or the hydrocarbon-rich/CO2-rich liquid (L1/L2) immiscibility high-pressure limit of 20 MPa [2,900 psi] is attained. The mercury is then slowly withdrawn, and a specified amount of water or brine is introduced into the cell. Mercury is then injected, the cell is rocked and put into an upright position, and phase volumes are measured once again until the L1/V bubblepoint in the presence of an aqueous liquid phase, L3, or 20 MPa [2,900 psi] is reached. To facilitate the comparison of results, the relative phase volumes of the L1/L3/V, L1/L2/L3/V, and L1/L2/L3 systems are normalized to exclude the L3, or aqueous, phase. Any change in the phase behavior caused by the aqueous phase is evidenced by a shift in the p-x diagram, where x corresponds to an overall mole fraction of CO2 or oil, including CO2 in the aqueous phase. If, however, the results are plotted on a p-x diagram in which x corresponds to overall mole fractions based only on CO2 and oil in the hydrocarbon phases and if the only significant effect of the brine was to solubilize CO2, the results would coincide.
The CO2/tetradecane system (the first binary in the CO2/n-alkane series that is characterized by a short critical locus proceeding from the CO2 critical point and a second critical branch that extends from the alkane critical point to high pressures2) was studied at 305 and 343 K [550 and 620°R].3 Equal volumes of oil and water were used in this series of experiments to facilitate the study of four-phase, three-component volumetric behavior4 at the lower temperature. A similar set of experiments with the (CO2+CH4)/C14H30/H2O system was also performed.3 Although these results are not presented, they were consistent with those presented in this study.
|File Size||578 KB||Number of Pages||9|