On the analytical extension of gas condensate plus fraction

Abstract

Heavy fractions play a vital role in gas condensate fluid phase behavior. Therefore, accurate determination of molar distribution in characterization of a gas condensate fluid is a crucial task. In the absence of enough experimental compositional data, mathematical functions are used to model the molar distribution of fluid heavy fractions. Different models have been developed over time and some of them such as exponential model and 3-parameters gamma distribution function gained popularity in industry and have been used in commercial simulators.Although advanced gas chromatographs are now able to analyze the fluid composition to C45+ or even higher, some researchers still believe simple distribution functions can be used to characterize the heavy fractions so that time and money be saved. A new molar distribution function called Four Coefficients Model (4CM) is being introduced recently. As author's claim, there is always discontinuity in SCN8 and SCN13 in gas condensate sample which none of the previous models are able to show them as those models are generated based on continuity in molar distribution. The authors believe that the model could be added to simulation packages.In this study, the 4CM model is evaluated and compared with accepted popular molar distribution models to see the generality of this new model. 42 sets of gas condensate sample along with samples which used in developing 4CM are used to implement this study.Results show that coefficients of 4CM are not universal and should be calculated for each fluid type, specifically. To do so, composition analysis must be reported at least to C14+. According to the results, different fluids exhibit different molar distribution; some are continuous, while some show discontinuity only in C8 and some show discontinuity in both C8 and C13. Therefore, 4CM with coefficients introduced by Raffie et al. may unable to model various molar distribution behaviors, as it is developed based on samples which all have discontinuities in both C8 and C13. However, when coefficients of 4CM are obtained for a specific reservoir, they give the best prediction of molar distribution when apply to other samples of the same reservoir too. Results also show that, by optimizing the alpha parameter in 3-parameters gamma distribution function, this model is being capable of predicting discontinuity in C8, if any. So, it is still the powerful method in predicting molar distribution of the fluid.