Gas sorption in H2-selective mixed matrix membranes: Experimental and neural network modeling

Robust artificial neural network (ANN) was developed to forecast sorption of gases in membranes comprised of porous nanoparticles dispersed homogenously within polymer matrix. The main purpose of this study was to predict sorption of light gases (H-2, CH4, CO2) within mixed matrix membranes (MMMs) as function of critical temperature, nanoparticles loading and upstream pressure. Collected data were distributed into three portions of training (70%), validation (19%), and testing (11%). The optimum network structure was determined by trial-error method (4:6:2:1) and was applied for modeling the gas sorption. The prediction results were remarkably agreed with the experimental data with MSE of 0.00005 and correlation coefficient of 0.9994. Copyright (C) 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.