SPE Members*Now with BHP Petroleum Pty. Ltd. Abstract From an analysis of the constant volume depletion experiments performed on 80gas-condensates, it is concluded that serious problems exist in the detection of the actual volume of retrograde liquid in the PVT cell. The measured liquid volume is usually only a fraction of that required to give realistic liquid densities. The frequently observed concave shape of the retrograde liquid saturation curve close to the dew point is not real and is the result of experimental error. It is recommended that the data quality be improved by directly measuring the retrograde liquid density because it provides an excellent control over the liquid saturation, especially when this saturation is low. The properties of the fluid initially in the PVT cell are normally determined mathematically. However, this often only approximates the true properties and deviations can have consequences for the derived liquid properties. The direct measurement of the composition and propertie scan easily solve this problem and is therefore strongly recommended. The analysis of the experiments leads to the following statements for constant volume depletion experiments . liquid densities are monotonically decreasing with pressure . B: and R: are both monotonically increasing with pressure. Here B: is defined as the inverted partial density of the C5+ component fraction in the liquid, while R: is the mass ratio of the C4- fraction over the C5+ mass fraction. Introduction The importance of a good knowledge of the phase behaviour and compositional properties of reservoir fluids for the development of gas-condensate reservoirs and the subsequent processing of the well stream is widely recognised. Substantial financial investment is made on the acquisition of reliable samples of a reservoir fluid and on measuring the phase behaviour or PVT propertiest. Great reliance is placed directly on the resulting data and often even more on a derived equation of state'mathematical representation. In this paper we address the subject of the quality of experimental PVT data obtained from the standard experiment performed on gas-condensates, the constant volume depletion(CVD). The experiment is related directly to the way the information is applied by reservoir engineers (reservoir material balance calculations and numerical reservoir simulations) and also serves to provide a set of data on which to base the development of a mathematical model of the fluid. This so called equation of state model is then used for all kinds of purposes, eg separator design, gas processing plant design, compositional simulations, vertical performance calculation setc. Recently a variation of the black oil PVT representation in terms of masses of pseudocomponents was presented which derives its parameters directly from experimental data without the need for equation of state simulations of separator systems etc., and which allows black oil reservoir simulations to be conducted with improved accuracy, especially for gas-condensates. P. 525^

Many methods of characterizing reservoir fluids thermodynamically yield equation-of-state (EOS) inodels with a large number of components.. For simulation studies, this detailed compositional model is subsequently xeplaced by a simplified hmdel in which compositions are described with a limited number of groups or pseudocomponents. This paper discusses methods of translating group (pseudocomponent) prepositions into detailed component compositions, a procedure that wiJ be referred to as "inverse lumping. " Inverse lumping can be used to calculate phase properties in tie group (@udocomponent) model. 'fhis ehninates the med to construct ad-hoc prope~correlations. Another application would be to inverse lump the welf streams as predictti by~e simulator so that the results can be used in a fully compositional separator-system simulator.httroduetion Thestudy of recove~processes in which the tmmsfer of components between phases is impor@nt often starts with the characterization" of the system of reservoir fluids in terms of many real components (such as methane and C02) and hypothetical component% the hypothetical components generelly represent the C7+ fraction of the mixture. 1 These components are the elements of an EOS mcdel for the phase behavior of the reservoir fluid system. When one proceeds to the numerical simulation of flow behavior, it is often impractical to use this detailed fluid model, and it is replaced by a simplified model in which thp compositional elements are a few groups of components rather @m many individual components.~.3 (These groups ace often cdkdpseudocomponents, but in this paper we shall uw "gro"p,, to emphasize the composite mture.) The simplified model can bean EOS model too, or it can be of the tabular or diagrammatical variety. In the latter case, the many-component EOS model is used to calculate in adwmce a selection of the phase equilibria expected to cccur during the simulation the resulting comprent compositions ere added to obtain group compositions (lumping), and these group compositions are stored as the basis for interpolation d"~g X8 simulation rumThe switch fmm components to groups as compositional elemerm is calkd "lumping" and obviously results in a loss of information. The purpose of this paper. is to present methods of retrieving the lost information to a good appnxirnation with little computatioml effort. This retrieval will be called ''irwerse Iumping. " It tmnslates phase equilibria of the group model into phase equilibria in terms of the components of the original model. An important application lies in the calculation of pha% properties (densities and viwo~ties) in the group model. Traditiormlly, this is done by treating ch group as a pure component (pseudocomponent) and assigning physical parameters to it a compositional correlation developed for actual pure components is then used m obtain phase properties, or simple mixing rules are used. This method necessitates a manypammeter mnlineqr tit of the physical parameters of the pseudccompmtmm to reproduce the correct...

No abstract

No abstract

This paper was prepared for presentation at the 7th Offshore South East Asia Conference held in Singapore, 2-5 February 1988. This paper was selected for presentation by a programme committee following review of information in an abstract. Contents of the paper have not been reviewed by OSEA and are subject to correction by the author(s). ABSTRACTA procedure for evaluating the quality and consistency of PVT data is presented. This procedure uses the material balance and the associated error analysis for investigating experimental accuracy, and to identify the sources of erratic data. The use of this method to 'correct' erratic data is illustrated for a near-critical or volatile oil and a gascondensate.