This paper proposes a new method to interpret transient pressure data from gas condensate reservoirs. Our method models two-phase gas-condensate flow in the reservoir using traditonal solutions to flow equations for single-phase, slightly compressible liquid flow by employing a pseudopressure function and fluid properties derived from conventional laboratory experiments.
The method is especially useful for well test analysis when reservoir pressure is near or below the dewpoint, where existing methods fail. Their assumption of steady-state flow yields a relation between reservoir condensate saturation and pressure that is not valid at these low pressures.
In this paper, we show that an accurate relation between reservoir condensate saturation and pressure can be computed from the assumption of a three-zone radial model, in which steady-state fluid flow is assumed in the first zone, the condensate is immobile in the second zone, and only the gas phase exists in the third zone. (Whether the third zone exists or not is determined by reservoir average pressure.) The second zone becomes increasingly important at lower pressures.
The essential problem in multiphase well test analysis is to establish the correct relationship between saturation and pressure needed to calculate pseudopressure. We verified our calculated relationship between reservoir condensate oil saturation and pressure by comparing our analytical results to results generated by compositional simulation. The comparison also showed that our proposed relationship between reservoir condensate saturation and pressure leads to accurate test interpretations and is much more satisfactory than previous methods.
Once the correct relation between reservoir condensate oil saturation and pressure is obtained, our method can compute and correlate pseudopressure with either producing or shut-in time. Sensitivity studies for various reservoir and fluid conditions support our conclusions under a wide variety of conditions.
Our test interpretation method provides good estimates of reservoir permeability and initial reservoir pressure. Mechanical skin is usually overestimated, but only modestly (typically about one unit of skin).
Our method requires lab constant volume depletion (CVD) data, gas and condensate oil relative permeability, and producing GOR in addition to transient pressure data. The effects of non-Darcy flow, wellbore storage, and capillary number-dependent relative permeability are not considered. The production wells are fully penetrated.