Density Changes at Supercritical and Near-Critical Conditions by Increasing CO2 Content in Synthetic Hydrocarbon Mixtures – A Comparison Between Experiments and Simulation Predictions

Abstract

Abstract Carbon dioxide (CO2) injection is a well-known EOR-method to reduce residual oil in the pore network of oil reservoirs. It is also increasingly used as a means of mitigating the greenhouse gas emissions problem by storing it in geological formations. A key parameter to such attempts is the density of the rich CO2 mixture, which is formed downhole in the injection well, since it affects the swelling potential, oil formation volume factor, viscosity, hydrostatic gradient, fluid distribution and formation pore pressure. The density of the crude oil-CO2 mixture depends on the pressure-temperature conditions, the CO2 concentration and the dominant hydrogen compounds in the crude oil, i.e. whether they are aliphatics, aromatics, or naphtenics (cyclic structures). The PVT properties of the different CO2-hydrocarbon mixtures vary greatly and the available experimental data for tuning PVT simulators are scarce, especially for ternary mixtures at high pressures and CO2 concentrations. This study investigates the effect of CO2 concentration on the density of ternary mixtures containing CO2, methane, and a pure liquid hydrocarbon, which is either an alkane, aromatic or cycloalkane compound. The liquid hydrocarbons used in the study were normal heptane (n-C7), toluene (Tol) and cyclohexane (c-C6). The measurements were conducted at variable compositions, at temperatures of 50, 70, and 90 °C, and at pressures ranging between 100 and 517 bar. The ternary mixtures were: Methane, toluene and CO2 at 1:1 molar ratio and CO2 concentrations of 14%, 27% and 72%, Methane, cyclohexane and CO2 at 1:1 molar ratio and CO2 concentrations of 19%, 47% and 68%. Methane, n-heptane and CO2 at constant molar hydrocarbon ratio (C1/n-C7) of 2:1 and varying CO2 concentrations of 23% and 75%, Some of the rich CO2 mixtures exhibited retrograde condensation behaviour at high temperatures. The results were compared against predictions from an EoS model (Peng Robinson Equation of State), coupled with volume shift parameters. The comparison between the simulation calculations and the experimental data indicated good agreement in the densities, but significant deviations in the boiling point pressures (Pb). As a result, the EoS model can be safely used to predict the CO2 mass storage potential of reservoirs of known pore volume such as the depleted ones.