Modeling of Cosolvents in a Fully-Implicit Surfactant Flood Simulator Using the Three-Level Framework

Abstract

Abstract Cosolvents are commonly injected along with surfactants for successful enhanced oil recovery as they help control aqueous stability, salinity gradient, and microemulsion phase viscosity. Therefore, modeling capability for numerical simulation of cosolvent injection is essential in helping design optimal surfactant floods. Also, the numerical implementation in the simulator should be fully implicit, fully coupled, and highly-scalable to enable full-field models and the higher resolutions often required by chemical flood simulations. We propose a novel numerical approach to model cosolvents in a fully implicit, fully coupled, parallel, four-phase surfactant flood simulator using the three-level (phase/pseudocomponent/pure component) framework. Three pseudoalcohol components are introduced to the framework for efficient modeling of surfactant phase behavior with alcohols that are partitioned to pseudooil, pseudowater, and pseudosurfactant, respectively. They consist of pure alcohol components which are partitioned to the same pseudocomponent and are distributed to phases as required by the phase behavior equations. New nonlinear solution variables of concentrations are proposed to model transport of pure alcohols, their partitioning into pseudcomponents, and distribution of the pseudoalcohols to phases, along with corresponding equations. The physical properties critical for surfactant flood simulation such as interfacial tension, phase relative permeability, viscosity, and mass density are extended to consider the effect of alcohols. It is shown that the new numerical approach significantly simplifies implementation of the cosolvent simulation functionality. This is because time consuming and error prone conversion between variables and derivatives, and local iterative solve for the concentrations, are not needed. This simplification enables us to significantly reduce implementation efforts, even within the fully implicit, fully coupled framework. The implementation is validated with various test cases against a widely referenced chemical flood simulator. A large-scale surfactant/polymer flood case with cosolvent injection is successfully simulated with all the important physical processes modeled, with the simulator exhibiting good performance. Large field scale, four-phase chemical flood simulations with surfactant phase behavior with cosolvents are now practically achievable with the novel numerical approach using the three-level framework without compromising comprehensive physics.