Influence of Lumping and Equation of State Tuning Methods on the Sub-Salt Reservoirs Simulation

The Brazilian pre-salt oil reservoirs contain a considerable amount of methane, CO₂ and volatile fractions under high pressure. Several authors suggest that the compositional reservoir simulation is fundamental for the description of their properties and prediction of oil production. Although this technique improves the quality of reservoir management, due to computational limitation, simplifications of fluid data are often recommended, affecting the quality of the compositional simulation. Petroleum reservoir behavior cannot be simulated in the most accurate way using present day computational resources. Simplifications such as lumping and tuning of equation of state (EOS) to describe mixtures have influence on the phase behavior. Fluid characterization becomes an important source of uncertainties in reservoir simulation because PVT data tuning is an inverse problem with multiple possible answers. The objective of this study was to show the impact of the description of the oil behavior under typical Brazilian sub-salt conditions considering oil production and computational effort. This was done through the study of EOS and comparison of the complete fluid with the lumped fluid descriptions. Since simplifications are adopted and PVT data-handling errors are introduced, it was necessary to detect and minimize them. The analysis of this study used comparisons of phase envelopes, MMP and reservoir simulations in a synthetic model. The studied model was based on the second model of SPE 10^th comparative study, heterogeneous, using oil similar to that found in the Brazilian pre-salt, with 42° API, 3.55% molar CO₂, and GOR 132.9 ft³ std/ft³. This work emphasizes the importance not only of experimental data quality but also shows the impact of the choice of regression and lumping methods. It also shows the importance of fluid modeling to obtain reliability in the reservoir simulation process.