Prediction of CO2/Crude Oil Phase Behavior Using Supercritical Fluid Chromatography

Abstract

Abstract In this paper, we describe a method for characterization of crude oils for predictions of phase behavior of CO2/crude oil mixtures with an equation of state (EOS). The method includes the use of supercritical fluid chromatography (SFC) with CO2 as the carrier fluid. Pressure-composition diagrams calculated using the Peng-Robinson EOS and this characterization scheme agree well with PVT observations. The advantage of this technique is that it produces predictions of the phase behavior of CO2/crude oil mixtures with much less experimental effort than is required to perform PVT experiments. Introduction Very high local displacement efficiency is possible in multicontact miscible floods because chromatographic separations occur as phases with different compositions move at different flow velocities. Hence, accurate prediction of phase compositions is an essential part of fundamental understanding of the development of miscibility in CO2/crude oil displacements. Development of cubic EOS 's has proceeded to the point that phase equilibrium calculations can be made routinely provided that the crude oil can be characterized adequately. An adequate characterization includes (1) division of the crude oil into an "appropriate" number of pseudocomponents and (2) specification of properties such as critical pressure (Pc), critical temperature (Tc), and acentric factor (w) for each pseudocomponent. In addition, essentially all of the modern EOS's make use of binary interaction parameters to improve the accuracy of calculated phase compositions. The use of pseudocomponents is required because any crude oil contains hundreds of hydrocarbon constituents. Even if all could be identified and the amount of each determined, the computational demands of phase equilibrium calculations would be too high for use in compositional simulations, for example. Some uncertainty inevitably arises from the lumping of real components into pseudocomponents, however, because the parameters required for EOS calculations cannot be measured directly. The approach most commonly used is to estimate Pc, Tc, and w from correlations, and then to adjust those values and to improve agreement between calculated phase behavior and observations from PVT experiments. The disadvantage of that approach is that time-consuming PVT experiments are required if accurate phase equilibrium calculations are to be made. In this paper, we propose an alternate approach to crude oil characterization for CO2 flood applications, and we investigate whether the proposed method can be used to make reasonably accurate a priori predictions of CO2/crude oil phase behavior. A key part of the characterization procedure is the use of SFC. SFC is similar to gas chromatography, except that the carrier fluid is a relatively dense, supercritical fluid. Here we use high-pressure CO2 as the carrier. Mixtures of hydrocarbons are separated because they partition differently between the mobile CO2 phase and the stationary phase in the packed column. Thus, the physical mechanism that produces the separation is similar to the mechanism that produces high displacement efficiency in a multicontact miscible CO2 flood. We make use of SFC with CO2 as a carrier in two ways. We use SFC to measure the composition of the oil by performing simulated distillation. In addition, we show how elution-time data can be used to estimate not available elsewhere for CO2 and moderately heavy hydrocarbon components. The proposed characterization procedure was tested for three CO2/crude oil systems. Reported below are comparisons of predicted phase behavior with PVT observations for mixtures of CO2 and oil from the Means, Maljamar, and Wasson fields. P. 71^