Modeling interfacial tension of normal alkane-supercritical CO2 systems: Application to gas injection processes

[Display omitted] To study the gas injection scenario for successful implementation of enhanced oil recovery (EOR) processes, the prediction of interfacial tension (IFT) between injected gas and the crude oil is of paramount significance. In the present study, some intelligent methods were developed for determining IFT values between supercritical CO2 and normal alkanes. IFT was considered as a function of temperature, pressure, and molecular weight of normal alkanes. The developed methods were Multilayer perceptron (MLP), Genetic Algorithm Radial Basis Function (GA-RBF), and Conjugate Hybrid-PSO ANFIS (CHPSO-ANFIS). The average absolute percent relative errors (AAREs) for the stated techniques were found to be 2.59%, 1.39%, and 1.81%, respectively, showing that GA-RBF is the most efficient technique. This model was then compared to the other previously developed models in literature. It was also found that the current GA-RBF model with AARE of 1.39% surpasses the previously developed models. Finally, the results of the Leverage approach showed that GA-RBF model could be trusted to predict the IFT of the normal alkane- supercritical CO2 systems in the used range of pressure, temperature, and n-alkanes. This study represents the most reliable technique in predicting the IFT value between supercritical CO2 and normal alkanes to be applied in studies on gas injection processes.