Compositional reservoir simulators require volumetric and compositional predictions of reservoir fluids from an Equation of State (EOS). This paper presents a method to optimize the parameters of a General Cubic EOS. The generalized cubic has four parameters that are related to the compositions of mixtures through pure component critical properties and binary interaction parameters.
The method to improve density predictions is a two-step process. First, the EOS parameters ξc, b, and γ are optimized from a nonlinear least squares regression of pure component volumetric data. This step insures that the critical constraints and the fugacity criterion are met. The second step extends the predictions to binary mixtures by introducing quadratic mixing rules to the parameters a, b, α, and β, including binary interaction parameters kij. Two iterative techniques were developed to establish values for the kij.
The average density predictions of three single-phase multicomponent hydrocarbon systems and 26 pure components were within 1% and 0.5% of their respective experimental values. The commonly used Peng-Robinson EOS (PR) predicted densities of these multicomponent hydrocarbon mixtures and pure components within 4% and 3%, respectively. Although emphasis was placed on improving density predictions, phase composition predictions of seven multicomponent hydrocarbon systems were also compared to the predictions from the PR EOS.