Practical Limitations in Obtaining PVT Data for Gas Condensate Systems

Abstract

Abstract This paper examines the accuracy of laboratory generated PVT data for gas condensate reservoir fluids. Although there are recognised difficulties in obtaining representative samples from such reservoirs, there are also significant uncertainties in the results of laboratory studies performed using the samples. The largest uncertainties are associated with the measurements of small liquid volumes, dew points and chemical compositions. Reservoir engineers should be aware of these uncertainties and should weight PVT data accordingly, when used as input parameters to reservoir models. Introduction The choice of the most profitable development scheme for any gas condensate reservoir is dependent upon various factors, including:field size and locationlocal market for separated gas and condensatetax regimephase behaviour of reservoir fluid Development schemes vary in the degree of pressure maintenance and in the pressure maintenance medium, which can be recycled gas, nitrogen or even water. Compositional simulation models are being used increasingly to study gas condensate reservoir depletion processes particularly when gas injection and recycling are used for pressure maintenance. Prediction of the phase behaviour of petroleum systems using the traditional cubic equation of state (EOS) is inherently unreliable. Inaccuracies in the phase behaviour prediction may be attributed to a poor theoretical basis of the model and to a lack of detail in the reservoir fluid analysis and heavy end (heptanes-plus) characterisation. If experimental PVT data are not used to tune the EOS, then the quantitative prediction is of little value. Hence, the development of reservoir simulation models is critically dependent on the qualify of experimental PVT data. PVT data-are generated in the laboratory using samples taken in the field during well testing. Frequently, the users of EOS packages do not appreciate the practical limitations in both obtaining samples from a gas condensate reservoir and, subsequently, measuring the physical and chemical properties of the samples. This paper discusses briefly the methods for obtaining samples from gas condensate reservoirs and then examines uncertainties in tile results of standard PVT studies, with specific reference to a study carried out at the BP Research Centre at Sunbury (RCS). Quantification of these uncertainties is extremely important because it allows correct weighting to be put on the experimental results used when tuning the EOS. It is not possible to make general statements concerning the accuracy of PVT data, because different laboratories perform studies in different ways and no two reservoir fluid systems or studies are identical.