Ali Haghighat Mesbahi - Department of Chemical Engineering, polymer engineering group. Isfahan University of Technology, Isfahan
Morteza Sadeghi - Department of Chemical Engineering, polymer engineering group. Isfahan University of Technology, Isfahan
Dariush Semnani - Department of textile engineering, Isfahan University of technology, Isfahan, Iran

Essential importance of gas separation in industry implies that modeling these processes is enormously considerable. Complicity and nonlinearity of these process lead to utilizing intelligence systems. In this study we present a model on effect of five input variable such as:1-Weight percent of polyurethane hard segment, 2-Type of polyurethane soft segment(Polypropylene glycol(PPG), polytetramethylene-glycol (PTMG) and polycaprolactone (CAPA225)),3- Molar ratio of H dexamethylene diisocyanate (HDI) 4-Molar ratio of Isophorone diisocyanate (IPDI) and 5- Toluene diisocyanate (TDI) in polyurethane (PU) membrane on N2,O2 ,CH4 and CO2 permeability via adaptive neuro-fuzzy inference system (ANFIS) as powerful intelligence method. We applied Genfis3 for ANFIS modeling. Genfis3 use fuzzy c-means (FCM) clustering. Owing to the fact that we had had small number of samples, leave one out cross-validation (LOOCV) selected for evaluation the ANFIS performance. Four networks corresponding to four gases have been investigated. The best performance has belonged to net on O2permeability with the average relative deviation (ARD %) equal to 6.94 for testing sets. We gain good result that show ANFIS is powerful tool in modeling permeability of polyurethane (PU) membrane