Algerian Journal of Engineering Research
Volume 5, Numéro 2, Pages 12-27

Regression Of Nrtl Parameters From Liquid–liquid Equilibria For Water + Ethanol + Solvent (dichloromethane, Diethyl Ether And Chloroform) Using Particle Swarm Optimization And Discussions At T =293.15 K

Authors : Hebboul Sabrina .


Experimental data taken from previous scientific studies of the liquid-liquid equilibrium (LLE) for ternary and quaternary systems {water +ethanol + dichloromethane (DCM), chloroform (CHCl3) or diethyl ether (DEE)} were correlated using the nonrandom two-liquid (NRTL) model. In the chemical engineering, there are metaheuristic methods are used to obtain binary interaction parameters of this model, which are the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO). In addition, the quality of the NRTL model predictions was assessed by calculating the variance of Root-Mean-Square Error (RMSE); it is used measure of accuracy to compare between the experimental equilibrium data and predicted data. Moreover, the lowest value of RMSE is always better than the highest. As a result, the PSO’s RMSE values were better than the GA’s RMSE values. Accordingly, the PSO was the best optimization method for these systems.


Genetic Algorithm, Liquid-Liquid Extraction, Nonrandom Two-liquid model and Particle Swarm Optimization.