High Pressure Phase Behavior of Asymmetric Mixtures for Oil Production

Development of a petroleum reservoir requires an accurate description of the fluid phase behaviour in order to determine the reservoir fluid type, estimate the gas and oil in-place, and analyze or simulate the production process with changes in the state, composition and properties of fluid phases. Although the subject has been studied over decades, there are still many long-standing challenges.

Among them, the description of the density and phase equilibrium of highly asymmetric mixtures related to reservoir fluids is an important issue especially for high pressure and high temperature (HPHT) reservoirs. Most reservoir fluids are by nature comprised of molecules with large contrast in molecular size and property, thus forming asymmetric mixtures.

The high asymmetry can lead to a lot of practical problems in the phase behaviour description, such as the non-ideal mixing in the density and viscosity modeling and relatively large phase envelopes that vary significantly with composition. In addition, other problems, like formation of additional liquid or solid hydrocarbon phases and difficulty in the saturation pressure calculations in the near critical region, are also related to the asymmetry.

This PhD thesis is dedicated to the study of the high-pressure phase behaviour of asymmetric mixtures related to reservoir fluids, whose data are generally scarce in the literature. Instead of continuing measurement of well-defined mixtures mimicking reservoir fluids, we prepared live fluid systems by combining a light gas component, including carbon dioxide, nitrogen, and methane, and stock tank oil (STO).

The obtained asymmetric multicomponent mixtures can be considered as pseudo binaries. They are better than well-defined mixtures in mimicking real live fluid samples, which are difficult to obtain from HPHT reservoirs. The prepared methane + STO, nitrogen + STO, and carbon dioxide + STO systems cover a wide composition range.

Their density and phase equilibrium data were systematically measured at temperatures from 298.15 to 463.15 K and pressures up to 1400 bar. The densities of the three gas + STO systems as well as the STO itself were measured through a vibrating tube densimeter Antar Paar DMA HPA while the phase equilibrium of the three systems was studied through a full visibility PVT 240/1500 cell manufactured by Sanchez Technologies.

The phase equilibrium study also ensured that the density measurement was conducted at sufficiently high pressures corresponding to single phase. From the experimental densities, we also determined the isothermal compressibility values by differentiating the Tammann-Tait equations fitted to the densities, and the pseudo excess volumes of gas + STO mixtures by assuming that the synthetic mixture was composed of just a gas component and an oil component.

In the phase equilibrium part, we determined the phase envelopes, relative volumes, and liquid volume fractions below the saturation point. Our measurement has produced valuable HPHT density and phase equilibrium data for evaluating and further improving thermodynamic models for HPHT reservoir fluids, thus supporting the relevant industrial applications on exploring and developing the high-pressure reservoirs.

The obtained data can also be used for relevant gas injection modeling. The measured density and phase equilibrium data were modeled by three equations of state (EoS) including Soave-Redlich-Kwong (SRK), Peng-Robinson (PR) and Perturbed Chain Statistical Associating Fluid Theory (PC-SAFT). For SRK and PR, their volume translated versions SRK-VT and PR-VT were also used to improve the performance in the density descriptions.

In the present work, C7+ characterization of the STO is based on Pedersen’s characterization method with different sets of correlations selected for these EoS models. For PC-SAFT, the correlations developed by Yan et al. were used. It is impossible to find one model that performs consistently better than the other models.

The performance of these models in saturation pressure is case dependent. For volumetric properties, volume translation is essential for SRK and PR, and the performance of PC-SAFT is similar to that of PR-VT or SRK-VT. It is found that the deviation in the calculated STO density has certain correlation with the deviation in the live oil density, which shows the importance of the STO density modeling in the live oil density description.

It is also found that the deviations in calculated excess volumes are small relative to the total volume. This can be utilized to estimate the live oil density from the measured STO densities at different pressures and calculated excess volumes, which can potentially reduce the amount of experimental work on the more difficult live oil density measurement.

The PhD thesis also includes an experimental and modeling study carried out at the Equinor research center on systems containing methane/natural gas and mono-ethylene glycol (MEG)/water. The systems are of direct relevance to tackle gas hydrate issues in the subsea pipelines. We measured the density and interfacial tension (IFT) of methane + MEG, natural gas + MEG, and natural gas + MEG + water at temperatures of 278.15K, 293.15K and 323.15K to support the actual needs of natural gas processing industry.

The vapor and liquid densities were measured through a vibrating tube densimeter Anton Paar DMA HPM while the IFT was measured by the pendant drop method. Meanwhile, we collected relevant density, solubility, and IFT data of the systems involving methane, MEG and water in order to compare with our measurement results and to test models.

In the related modeling part, we tested the performance of the cubic plus association (CPA) EoS using the measured IFT and density data as well as the collected IFT, density and methane solubility data. CPA was coupled with the parachor method and the linear gradient theory to model the IFT data. The measured and collected data provide a more complete picture of the thermodynamic properties for MEG containing systems and they can contribute to evaluating and further improving the current models.