Ternary Gas Permeation Through Synthesized PDMS Membranes: Experimental and CFD Simulation Based on Sorption-Dependent System Using Neural Network Model

In this study, a predictive model for the separation of gases via a polydimethylsiloxane (PDMS) membrane has been developed. This model takes into account the effects of gas composition and pressure at the membrane surfaces on the gas sorption and diffusion coefficients in the membrane. Computational fluid dynamics (CFD) modeling has been employed in order to predict the behavior of a gas mixture containing C3H8, CH4, and H-2 at various operating conditions and three zones (upstream, downstream, and membrane body). Artificial neural network (ANN) modeling has been applied to predict sorption and diffusion coefficients of each component of the gas mixture in the membrane. A procedure of calculation has been applied to combine the CFD modeling and the ANN modeling in order to predict sorption, diffusion, and composition of each component at various sites of the membrane. The results determined using the developed prediction model have been found to be in agreement with those determined using experimental investigations with an average error of 10.21%. POLYM. ENG. SCI., 54:215-226, 2014. (c) 2013 Society of Plastics Engineers