Towards Accurate Density and Interfacial Tension Modeling for Carbon Dioxide/Water Mixtures

Abstract

Phase behavior of carbon dioxide/water (CO2/H2O) binary mixtures plays an important role in various CO2-based industry processes, including CO2 injection for enhanced oil recovery and CO2 storage in saline aquifers. Engineering design of such processes requires an appropriate thermodynamic model that can well capture the vapor-liquid equilibria (VLE), liquid-liquid equilibria (LLE), phase densities, and interfacial tension (IFT) of CO2/H2O mixtures. This work aims to screen such a model out of a number of promising candidate models. A special attention is given towards the phase density predictions as well as IFT predictions. A comprehensive analysis reveals that Peng-Robinson Equation of State (PR EOS) (Peng and Robinson, 1976), Twu alpha function (Twu et al., 1991), Huron-Vidal mixing rule (Huron and Vidal, 1979), and Abudour et al. (2013) volume translation model is the optimum model which yields average absolute percentage errors (abbreviated as %AAD) of 6.52% and 2.88% in reproducing the experimental phase composition data (i.e., 195 data points) and density data (i.e., 855 data points) collected in the literature over 278.00-478.35 K and 2.20-1291.90 bar. After reliable modeling of phase compositions and densities for CO2/H2O mixtures has been achieved with the optimal thermodynamic model, a new empirical IFT correlation for CO2/H2O mixtures is proposed through a nonlinear regression of the measured IFT data collected from the literature over 278.15-477.59 K and 1.00-1200.96 bar (i.e., a total of 778 data points for CO2/H2O mixtures with 589 training data and 189 test data). The inputs of the IFT model are the phase compositions and densities calculated by the aforementioned PR EOS model. Although the newly proposed IFT correlation only slightly improves the prediction accuracy yielded by the refitted Chen and Yang’s correlation (Chen and Yang, 2019), the proposed empirical correlation avoids the inconsistent prediction trend present in Chen and Yang’s model (Chen and Yang, 2019) and yields smooth IFT predictions.