Vapour-liquid equilibrium data analysis for mixed solvent-electrolyte systems using neural network models

Two-perceptron artificial neural network correlations were proposed for the prediction of vapor-liquid equilibrium for mixed dual-solvents single elec trolyte systems, and validated over an extensive VLE database (2,900 data, 16 binary solvents, 24 salts, 11 cations, 6 anions). Performances of these correlations to predict vapor-phase mole fraction, equilibrium temperature and total pressure, were discussed via comparisons with the experimental da ta, and the UNIFAC electrolyte model (Kikic, Fermeglia & Rasmussen, 1991). Chemical Engineering Science, 46, 2775-2780.) and the Extended UNIQUAC mode l (Iliuta, Thomsen & Rasmussen, 2000). Chemical Engineering Science, submit ted for publication.). The mean absolute deviations in predicted vapor-phas e mole fraction, temperature and pressure for the entire database were 0.02 5, 1.46 K and 1.68 kPa, respectively. (C) 2000 Elsevier Science Ltd. All ri ghts reserved.