A parameter sensitivity analysis using an EOS for optimal characterization of Cupiagua reservoir fluids

Abstract

: : Issued also on microfiche from Lange Micrographics. Cupiagua field is located in the foothill belt of the Llanos basin in eastern Colombia. This field has large reserves of retrograde gas in the Mirador (Tertiary), Barrio, and Guadalupe (Cretaceous) formations that are typically beyond 12,000 ft with in-situ pressures and temperatures higher than 5,500 psi and [], respectively. The Cupiagua hydrocarbon system contains favorable thermodynamic factors (composition, pressure, and temperature) for a gas injection process to improve the hydrocarbon recovery. To carry out this process successfully, we need a good characterization of the reservoir hydrocarbons to predict thermodynamic fluid behavior. Further, this characterization is to be used in a compositional simulation to predict the distribution of the fluid saturation and pressure through the reservoir. In this work, the characterization of the Cupiagua fluid system was accomplished by use of the modified Pang-Robinson equation of state (PR-EOS). Experimental data from four welts at difference depth were used. The characterization procedure developed for this purpose includes defining EOS properties, pseudoization, mixing rules, and tuning the molecular weight of the heavy fraction, weighting factors, and regression of the Da and Db parameters of the EOS. Schemes of 10 and 7 pseudocomponents were used to calculate the EOS parameters. A unique description for surface and reservoir conditions was used in the simulation process. The results were compared with the experimental data to find the best predictor scheme. We found that a good match of experimental data was obtained for both the 10-and J-pseudocomponent schemes. The 10-psudocomponent scheme was selected to illustrate the matching process between lab data and EOS data. A compositional con-elation of the composition with depth was found and used to extrapolate the composition throughout the reservoir. Finally, a three-dimensional composition simulation was carried out using the both schemes. The 10-pseucomponent scheme showed good results and high stability during the entire simulation. The 7-pseudocomponent scheme presented high instability at higher rates of production, and the computing time was not improved substantially with respect to the computing time of the 10-pseudocomponent scheme.