Prediction of microfiltration membrane fouling using artificial neural network models

In this study, artificial neural network (ANN) models were applied to predict the performance of microfiltration (MF) system for water treatment. A series of bench scale experiments were conducted at critical flux and supra-critical flux conditions with various permeate fluxes and feed water qualities. The effects of operating parameters on membrane performance were evaluated based on the comparison of transmembrane pressure (TMP) as a function of operating time. The ANN models used five input variables including permeate flux (Jw), feed water turbidity (Turf), UV254, time (h), and backwash frequency for predicting corresponding TMP. The modeling results indicated that there was an excellent agreement between the experimental data and predicted values. Nevertheless, selection of database for training is important for the accuracy of ANN prediction. Relative weights of each input variable were calculated to find out key operational factors affecting the performance of MF system.