The prediction of liquid-liquid equilibria of water/alcohol/ionic liquid ternary systems using support vector machine model

Yasin Jamalinejad¹ Fariborz Shaahmadi²

1) Department of Chemical Engineering, Islamic Azad University Dezful branch, Iran
2) Gas Engineering Department, Petroleum University of technology, Ahvaz, Iran

Publication : 2nd International Biennial Oil, Gas, and Petrochemical Conference(opgc2018.ir)
Abstract :
Liquid-liquid extraction is one of the most commonly used in chemical and petrochemical industries for the extraction in azeotropic mixtures. Ionic liquids, considered as ‘’green solvent’’, have been recently utilized as solvents in liquid-liquid extraction because of their negligible vapor pressure. Precise prediction of liquid-liquid equilibria (LLE) for ternary systems such as system water/alcohol/IL are essential. Thermodynamic modeling of these ternary systems (water/alcohol/IL) is very complicated, and for each ternary systems usually reqires several adjusted parameters. Iteligent models can be used as appropriate alternatives for the conceptual models. In this study, support vector machine (SVM) is applied to predict the thermodynamic phase behavior of LLE for water/alcohol/ionic liquid ternary systems. The accuracy of the obtained results of intelligent modeling is investigated by statistical and graphical error analyses.
Keywords : Azeotrope Thermodynamic Liquid-liquid extraction Modeling SVM Ionic liquids