Application of Genetic Algorithm to Parameter Estimation in Liquid-Liquid Phase Equilibrium Modeling

Parameter estimation problems for liquid–liquid equilibrium data modeling are challenging due to the nonlinearity of thermodynamic models. In this work, the stochastic global optimization technique, namely Genetic Algorithm is applied to calculate the interaction and non randomness parameters of NRTL model. Forty ternary extraction systems containing benzene, hexane and different ionic liquids as solvent at various temperatures reported in literature are selected. The parameters of NRTL model are calculated and the global rmsd value of 0.0076 for 357 tie-lines shows that these parameters can be considered as the global parameters of NRTL model in the studied ternary systems. The global parameters are applied in all systems and rmsd values for them are reported. The results showed that Genetic Algorithm as a powerful and effective tool can be used to optimize highly nonlinear problems in phase equilibrium modeling.