Dynamic ultrafiltration modelling for a submerged membrane bioreactor based on neural networks

Abstract:

A neural network (NN) model to predict the membrane performance in a scale pilot submerged membrane bioreactor (SMBR) has been developed. This SMBR has been used for the aerobic treatment of municipal wastewater. The aim of this work was to predict the dynamic behavior of the transmembrane pressure (TMP) as a function of three operating parameters, i.e., permeate flux, the mixed liquor suspended solids and the hydraulic cleaning cycles. The applied NN was a multilayer perceptron network (MLPN) and the Levenberg-Marquardt (LM) optimization algorithm was employed for its training. © 2010 Taylor & Francis Group, London.

Año de publicación:

2010

Keywords:

• Neural networks
• Membrane bioreactor
• Ultrafiltration

Fuente:

scopus