Assessment of the simultaneous regression of liquid-liquid and vapor-liquid equilibria data of binary systems using NRTL and artificial neural networks

Abstract

A numerical analysis on the phase equilibrium modeling performance of the NRTL equation in binary mixtures using different data regression strategies was carried out. Several formulations for the binary interaction parameter estimation were studied including a novel approach using an artificial neural network model for the simultaneous correlation of liquid-liquid and vapor-liquid equilibria data. The artificial neural network strategy substantially improved the estimation of the liquid-liquid equilibrium data in comparison with the traditional parameter estimation strategies utilized for the thermodynamic models. A detailed analysis on the performance of the model in terms of the calculation of homogeneous and heterogeneous azeotropes was also performed. This study showed that the application of artificial intelligence tools is an interesting alternative to improve the capabilities of local composition models for the liquid-liquid and vapor-liquid equilibria calculations.