Study of dead-end microfiltration features in sequencing batch reactor (SBR) by optimized neural networks

In this research, the structure optimized BP-ANN model was applied to simulate the permeate flux as a function of mixed liquor suspended, temperature, dissolved oxygen, hydraulic retention time, transmembrane pressure and operating time during dead-end microfiltration of activated sludge suspensions and its supernatant from sequencing batch reactor (SBR). This artificial neural network approach was also used to model the chemical oxygen demand (COD) concentration of effluent from SBR. The results showed that the structure optimized single hidden layer neural networks was able to accurately simulate the dynamic behavior of permeate flux and the COD concentration for SBR activated sludge process, and this BP-ANN model possessed higher accuracy than that of C. M. Silva's predictive model and linear multi-regression model.

Research highlights
► Optimal topology configuration of ANN was used to predict the MFM flux from MBR.
► Optimal topology configuration of ANN was used to predict supernatant's COD from MBR.
► The predicted result of ANN is better than that of M. Silva's and regression models.