عنوان مقاله [English]
In this research, the performance, efficiency, and properties of anti-fouling and flux of poly vinylidene fluoride (PVDF) nano-composite membranes with concentration of 15wt.% and 18 wt.%, mixed with different functional carbon nanotubes (-OH, -COOH, -NH2), were made and studied using phase inversion and normal methylpyrrolidone (NMP) solvent; moreover, the nano-composite membranes tested for flux, fouling, contact angle, porosity and protein rejection rate. Also, by using empirical test results, (1) flux and (2) fouling parameters were modeled based on the input variables including nanoparticle percentage, polymer percentage, porosity, contact angle and protein rejection rate. In this model, four intelligent systems including multiple layer percepton, radial basis function, least squares support vector machine and adaptive neuro-fuzzy inference system and three optimization algorithms including generic algorithm, simulated annealing and particle swarm optimization have been used. The results showed that for both flux and fouling parameters, the best models are GA-RBF and Conjugate-ANFIS with high correlation coefficient. In the next section, modeling was used to obtain optimal values of the best models made for both outputs (minimum fouling and maximum flux) and then the combined algorithm of the genetic algorithm and particle swarm optimization values were obtained. Afterward, by using optimization results for each type of polymer (15wt% and 18wt%), the membranes were made in the laboratory, and then flux, fouling, contact angle and porosity tests were performed, and the results were compared with the results of the model. Finally, the results showed that 0.07 wt.% single-walled carbon nanotube-PVDF nanocomposite membrane functionalized with hydroxyl group and 0.17 wt.% single-walled carbon nanotube-PVDF nanocomposite membrane functionalized with hydroxyl group had the best performance with the polymers of 15 wt.% and 18 wt.% of PVDF respectively.