Simulação composicional de reservatório contendo CO2: estratégias de lumping e estocagem em aquíferos

Abstract

A highly complex work consists of reservoir simulation containing a mixture of hydrocarbons with a high CO2 content, simultaneously, associating precision, speed, and economy in computational processes with the current available resources. In this way, through an algorithm proposed by NIETO-DRAGHI and other researchers (2008), a program was created to group the compounds of the mixture through the method called dynamical cluster, in order to perform the lumping of fluids automatically at from some of its properties, such as critical pressure, temperature and volume, acentric factor and molecular mass. An analysis, called sensitivity test, was carried out to evaluate the influence of each parameter considered in the program, using experimental tests and application of fluids with different CO2 contents to verify the impact, on the results, of the change in its composition. A study of the variation in the number of pseudocomponents for oils with different CO2 compositions was carried out, in which six arrangements (5, 7, 9, 11, 15 and 24 pseudo-components) were compared and, in each one, the equations were applied of Peng Robinson (PR) and Soave Redlich Kwong (SRK). Reservoir tests were carried out to evaluate the behavior of vari-ables, using two production fluids, in 9 different arrangements, with and without experimental data (the latter, called the theoretical case), in a reservoir model found in the literature, the in order to compare and show the efficiency of the results presented by the phase envelopes and the algorithm. In addition, the accuracy of CMG thermodynamic database, the different production scenarios with gas or water injection and the impact of the occurrence of chemical reactions on the storage of CO2 in oil reservoirs and/or aquifers were also analyzed. Thus, all the variables necessary to calculate the properties of the pseudocomponents proved to be relevant in the lumping results, with greater rele-vance for the critical pressure, with errors of up to 2%. Regarding the reduction in the number of pseudocomponents, the analyzes indicated a better performance of the curves with 9 pseudocompo-nents (when compared to those obtained by the mixture with 24 components) and it was not obser-ved a direct relationship between the number of components and the quality of the results. In the reservoir test, from the different fluids and arrangements studied and the use of the thermodynamic database, it was found that, regardless of the number of pseudocomponents, optimistic results for the theoretical cases (analyses without experimental data), with a factor of recovery (FR) of, appro-ximately, 48%. As for the cases with experimental data, the reductions from 24 to 9 and to 15 pseu-docomponents provided differences in the results, with more satisfactory behavior for the two oils with 9 pseudocomponents (FR of 49% and 41% for fluids with 8.24% and 3.55% of CO2, respecti-vely) and use of the Peng-Robinson equation, this divergence is justified by the difference in the CO2 composition. Finally, in the study of the impact of the occurrence of chemical reactions in the storage of CO2 in an aquifer, the importance and contribution of the reaction parameters were noti-ceable, with favorable results of CO2 mineralization for the tests with higher temperature values (130 °C) and salinity (200,000 ppm).