An Equation of State to Model Microemulsion Phase Behavior in Presence of Co-Solvents Using Average Solubilization Theory

Abstract

Co-solvents are used with surfactants in modern chemical enhanced oil recovery (CEOR) formulations to avoid formation of viscous microemulsion phases (and reduce costs) in porous media. Modeling the effect of co-solvents on phase behavior is critical to CEOR reservoir simulations. The state-of-the-art is to use HLD (Hydrophilic Lipophilic Difference) with a modified form of NAC (Net Average Curvature) as an Equation of State (EoS) to model microemulsion phase behavior. In this paper, we use an alternative EoS flash algorithm and couple it with an alcohol partitioning model to predict physical phase behavior. In this paper, we show that the net curvature equation in NAC is not valid for overall compositions away from typical experimental conditions, specifically in Type I and II systems. Alternatively, we use experimental evidence to correlate the harmonic average of oil and brine solubilization ratios to HLD. We use the average solubilization ratio equation with boundary conditions that allow for microemulsion phase type regions to be well defined, thus making the flash calculations robust. To model the co-solvent effect, we couple the newly developed average solubilization theory (AST) based EoS with the Prouvost-Pope-Rouse model to capture co-solvent partitioning across oil, brine and microemulsion phases. The resulting AST theory allows for a HLD based EoS to predict physical two-phase regions with no discontinuity in phase behavior thereby making it a more robust alternative to HLD-NAC. We used 80 phase behavior experiments over a wide range of hydrocarbons and temperatures to validate our approach. The coefficient of determination between the actual experimental data and the predicted model output was found to be above 0.9.