Abstract
Membrane separation of CO2 from natural gas components is known to be more efficient at high feed gas pressure. However, one of the main challenges of high pressure operations is penetrant - induced plasticization phenomenon. An approximate model for predicting the plasticization pressure was developed using the fundamental theory of dual - mode sorption model and the total immobilization assumption of the penetrant concentration in the Langmuir sorption site The assumption reduces the permeability - pressure relationship to system of quadratic equations whose coefficients were used to derive an expression for estimating the plasticization pressure. The resulting expression showed that, the plasticization pressure is a multivariable function of polymer properties. The support vector (SVR) algorithm was then used to correlate between the plasticization pressure, and the fractional free volume and the glass transition temperature of the polymers. Results obtained revealed that, with a careful combination of these two properties, the SVR can be successfully employed to predict the plasticization pressure of membrane used in high pressure CO2 removal from natural gas. The correlation coefficient for the training and the testing are 0.8837 and 0.9433, respectively. (C) 2015 Elsevier B.V. All rights reserved