Document Type : Research Paper
Faculty of Advanced Technologies, Nano Chemical Eng. Dep., Shiraz University, Shiraz, I. R. Iran
Faculty of Chemical Engineering, University of Tehran, Tehran, I. R. Iran
School of Electrical and Computer Engineering, Shiraz University, Shiraz, I. R. Iran
In this work, several machine learning techniques are presented for nanofiltration modeling. According to the results, specific errors are defined. The rejection due to Nanofiltration increases with pressure but decreases with increasing the concentration of chloride ion. Methods of machine learning represent the rejection of nanofiltration as a function of concentration, pH, pressure and also the experimental rejection. The results are in promising agreement with the experimental data taken from the literature. Six methods for modeling and prediction of rejection by nanofiltration membranes are presented in this study. The models have been trained and tested with a selected data set. Three defined matrices have been used to analyze the performance of the models.