Comparison of Stephan Maxwell Model with neural network model for Pevaporation process of water-ethanol

Abstract

The recent development of solvent and temperature-resistant, deliquescent zeolite Hydroxysodalite (HS) membranes has created it possible to beat the on top of limitations of deliquescent polymeric membranes. Mineral membranes have uniform and molecular-sized pores, and that they separate molecules supported variations within the molecules’ sorption and diffusion properties. Sturdy electricity interaction between ionic sites and water molecules (due to its extremely polar nature) makes the mineral HS membrane terribly deliquescent. During this study, experiments were conducted with varied Ethanol–water mixtures at 25. Total flux for Ethanol–water mixtures was found to vary from 0.319 to 0.226 kg/m2.h with increasing ethyl alcohol concentration from 1 to 20 wt.%. The precise nanoporous structure of the mineral cage helps in an exceedingly partial molecular sieving of the massive solvent molecules resulting in high separation factors. The comparison of the experimental results with the results of the Stephan Maxwell (S.M.) model additionally because the neural network model was disbursed during this study. The results showed that the neural network model might predict experimental results higher than writer Maxwell's model.