Optimization of NRTL interaction parameters for vapor liquid equilibrium using genetic algorithm

The non-ideal behavior of fluid mixtures can be represented at the thermodynamic level by the constituents activity coefficients which are key parameters for phase equilibria predictions. Several empirical and no linear models are available in the literature for the determination of these activity coefficients. However the great majority of these models do require binary interaction parameters which are not always available. Consequently the present work deals with the optimization of NRTL model binary interaction parameters of liquid - vapor systems by means of the genetic algorithm (GA), using experimental data. The obtained results are in a perfect agreement compared with the experimental values reported in DECHEMA data basis, confirming the quality of the optimized interaction parameters as well as the GA reliability for predicting the behavior of such systems."