Challenges In Modeling Fluid Systems With Unusual Contents Of Dissolved CO2 In Hydrocarbon Systems: Colombian Sedimentary Basins Case Of Study

Several fields in the Caguán-Putumayo Basin of southern Colombia have produced oil with CO2 dissolved in the crude oil at reservoir conditions above and below 50% mol over the last 5 decades. These fields are currently undergoing CO2 EoR and CCUS studies, which require that historically used PVT models be re-evaluated using the state-of-the-art thermodynamics available for this purpose. The most appropriate model is presented for this Colombian basin where the hydrocarbon systems present are very different from each other, have unusual and variable contents of originally dissolved CO2, and some of them correspond to fluid systems that are close to the critical point under reservoir conditions, increasing the difficulty of modeling, due to their supercritical fluid condition and their PVT properties at reservoir pressures and temperatures. The main objective of the study was to build a Peng-Robinson Cubic Equation of State (PREOS) so that a predictive model with minimal or no regression could be generated. Furthermore, this model allows us to determine the volumetric and phase properties of the different fluid systems and to evaluate the diffusion phenomena that dominate the gradients of compositional variation with depth. The proposed-fluid-model is currently being used in numerical reservoir simulation, improved recovery processes, and CCUS studies. An integrated analysis of the basin was performed during its productive life, allowing a compositional analysis of the available fluids, considering crude with API gravity of 25° to 50° and dissolved CO2 between 10% and 90% mol in the live fluid in the reservoir. Subsequently, with the application of the three-parameter gamma function, the modeling of the different compositional distributions typical of the different API gravities found in the basin was conducted, characterizing, and adjusting the three-parameter PREOS for the different fluid systems. PVT properties were calculated, obtaining a deviation of less than 10% concerning the experimental data, good enough for the basin scale. After adjusting the PREOS parameters, finally, the compositional variation with depth models was generated. The main finding of the predictive model developed was the generation of different scenarios where it was identified that the vertical distribution of the components is linked to pressure diffusion, especially for those that have a greater quantity of heavy components inside the fluid. In addition, it was found that CO2 is significantly sensitive to thermal diffusion within the hydrocarbon reservoirs, so a direct relationship between the thermal effect and the acceleration of the compositional gradient was identified.