Modeling of interfacial tension in binary mixtures of CH4, CO2, and N2 - alkanes using gene expression programming and equation of state

Interfacial tension (IFT) plays a key role in enhanced oil recovery (EOR) processes that involve the injection of light hydrocarbon and non-hydrocarbon gases into oil reservoirs. Crude oil is a complex mixture of different hydrocarbons, its phase behavior is unknown in presence of miscible gas under reservoir conditions; therefore, most of the IFT measurement experiments are done using a certain fluid, such as n-alkanes. CO2 is a common gas used for EOR purposes. Since N2 and CH4 impurities in the injected CO2 stream substantially influence the IFTs and the MMP of the system, several experimental studies conducted on (N2 + n-alkane) and (CH4 + n-alkane) systems over a wide range of temperature and pressure to provide a comprehensive IFT database. Accurate prediction of this property under the thermodynamic conditions encountered in petroleum reservoirs is of paramount importance for achieving the highest amount of oil recovery. In this study, two different models are developed for prediction of IFT in binary mixtures containing (CO2 + n-alkane), (N2 + n-alkane), and (CH4 + n-alkane). These models include the thermodynamic-based model of Weinaug and Katz 1943 [1] combined with PR-EoS, and the mathematical-based model of Gene Expression Programming (GEP). A preprocessing based on Z-score method is applied on the assembled dataset to remove the outliers and duplicates from the data. The GEP and EoS models were able to predict accurate IFT results in all three systems with the EOS providing slightly inferior results in systems containing N2 gas. For some experimental observations the EoS model failed to provide any IFT results. In addition, to have a better understanding of the performance of the developed models, they were compared for IFT values less and greater than or equal to 1 mN/m for all the systems. The average squared coefficient of determination (R2) for the GEP model in CH4, CO2, and N2 – alkanes systems were 0.92, 0.94, and 0.91, and for EoS model were 0.94, 0.87, and 0.66, respectively. The findings of this study can help for a better understanding of interfacial tension under the thermodynamic conditions encountered in petroleum reservoirs is of paramount importance for achieving the highest amount of oil recovery. •2206 selected values for the interfacial tension between n-alkanes and different gases were gathered and modeled.•Gene Expression Programming and Equation of State were used to model the gathered data.•Outlier diagnosis was performed