Improved Fluid Characterization for Miscible Gas Floods

by
Azubuike Michael Egwuenu, M.S.E.

University of Texas at Austin, 2004
Supervisor: Russell T. Johns

Injection of gases into a reservoir for enhanced oil recovery results in complex fluid phase behavior that cannot be modeled by black oil simulators. This interaction of flow and phase behavior is best captured by fully compositional simulators. A drawback of fully compositional simulators is that they require accurate reservoir fluid characterizations by equations of state (EOS) to capture the phase interactions in miscible gas floods. Another disadvantage is that EOS are computationally intensive. An EOS is typically tuned to standard PVT data, which may include multicontact experiments and swelling tests. The standard method of tuning, however, does not incorporate important displacement parameters such as the minimum miscibility pressure or enrichment (MMP or MME) or the likely compositions that result in a reservoir from condensing-vaporizing displacements.

The currently available correlations for the estimation of MMP for CO₂ floods in literature are not robust enough for the wide range of reservoir oil types and conditions. The MMP correlation available in literature for impure CO₂ floods is so limited in its application that it is only suitable for a certain West Texas reservoir oil type.

This thesis presents new MMP correlations for the displacement of multicomponent oil by CO₂ and impure CO₂. The approach is to use recently developed analytical theory for MMP calculations from EOS to generate MMP correlations for displacements by pure and impure CO₂. The advantage of this approach is that MMPs for a wide range of temperatures and reservoir fluids can be calculated quickly and accurately without introducing uncertainties associated with slim-tube MMPs and other numerical methods. The improved MMP correlation is based solely on the reservoir temperature, molecular weight of C₇₊, and percentage of intermediates (C₂-C₆) in the oil. The MMPs from the improved correlation are compared to currently used correlations and 41 experimentally measured MMPs. Correlations are also developed for impure CO₂ floods, where the injection stream may contain up to 40% methane. The new correlations are significantly more accurate and applicable than currently used correlations.

An improved tuning procedure for miscible gas floods that can more accurately represent the interaction of flow and phase behavior is also presented. Two displacements are used to demonstrate the approach; an eleven-component CO₂ flood and a twelve-component enriched gas flood. The MOC analytical theory is used to determine the MME (or MMP) of both lumped and unlumped pseudocomponent EOS models.

The results show that by tuning to the MME or MMP, fewer pseudocomponents were required to obtain the same degree of accuracy as the standard approach with more pseudocomponents. The lumped model gave good oil recovery prediction and composition velocities even when lumped to four pseudocomponents. The MME or MMP is shown to be the key tuning parameter. Furthermore, for the cases studied, the lumped models tuned with the proposed characterization method gave more reliable results compared to traditional tuning methods over a wide range of dispersion found in reservoirs. The reduced number of components will reduce the computational time required for compositional simulators.

Back to theses index