Using Neural Networks to Predict Treatment Process Performance

Abstract

This article discusses the application of a simple artificial neural network (ANN) model to predict microfiltration/ultrafiltration (MF/UF) performance with reasonable accuracy using data typically gathered online in an MF/UF facility. By developing a site‐specific ANN model, operators of MF/UF facilities could predict the performance of critical parameters such as transmembrane pressure (TMP) and plan accordingly. For instance, when the model predicts a higher rate of fouling based on feedwater quality, operating conditions could be modified to reduce the rate of increase of TMP, thereby reducing the required cleaning frequency of MF/UF systems.