Modeling of filtration process using PSO-neural network

Abstract

Modeling of membrane filtration process is a challenging task because it involves many interactions from biological and physical operation behavior. Membrane fouling in filtration process is too complex to understand and to derive a robust model become very difficult. The aim of this paper is to study the potential of neural network based dynamic model for submerged membrane filtration process. The purpose of the model is to represent the dynamic behavior of the filtration process therefore the model can be utilized in the prediction and control. The neural network model was trained using particle swarm optimization (PSO) technique. Three methods of PSO are compared to obtained an optimal model which are random PSO (RPSO), constriction factor PSO (CPSO) and inertia weight PSO (IW-PSO). In the data collection, a random step was applied to the suction pump in order to obtained the permeate flux and transmembrane pressure (TMP) dynamic. The model was evaluated in term of %R2, root mean square error (RMSE,) and mean absolute deviation (MAD). The result of proposed modeling technique showed that the neural network with PSO is capable to model the dynamic behavior of the filtration process.