Permeability and selectivity prediction of poly (4-methyl 1-pentane) membrane modified by nanoparticles in gas separation through artificial intelligent systems

In this work, the effects of operative parameters on CH4, CO2, O2, and N2 membrane gas separation for poly (4-methyl-1-pentane) (PMP) membrane modified by adding nanoparticles of TiO2, ZnO, and Al2O3 are assessed and investigated. The operative parameters were type and percentage of nanoparticles, and cross membrane pressure. The membrane permeability and selectivity were selected as the responses and indexes of separation process performance. To design the experimental layout, design of experiment methodology (DoE) techniques were used. Further, the separation process was modeled and simulated using artificial intelligence (AI) methods. So, a robust black-box model based on radial basis function (RBF) network was developed and trained with the ability for predicting the performance of membrane process. The developed model could simulate the process and predict the permeability with R2-validation of 0.9. Finally, it was found that addition of nanoparticles and increasing the operative pressure had positive effects on membrane performance. Maximum permeability values for O2, N2, CO2 and CH4 were 181.58, 52.09, 550.85, and 54.26, respectively. The maximum values of validation-R2 of optimum structure for CO2/N2 and CO2/CH4 selectivity were 0.8697 and 0.7028, respectively.