Dynamic Modelling of Aerobic Granular Sludge Artificial Neural Networks

Abstract

Aerobic Granular Sludge (AGS) technology is a promising development in the field of aerobic wastewater treatment system. Aerobic granulation usually happened in sequencing batch reactors (SBRs) system. Most available models for the system are structurally complex with the nonlinearity and uncertainty of the system makes it hard to predict. A reliable model of AGS is essential in order to provide a tool for predicting its performance. This paper proposes a dynamic neural network approach to predict the dynamic behavior of aerobic granular sludge SBRs. The developed model will be applied to predict the performance of AGS in terms of the removal of Chemical Oxygen Demand (COD). The simulation uses the experimental data obtained from the sequencing batch reactor under three different conditions of temperature (30˚C, 40˚C and 50˚C). The overall results indicated that the dynamic of aerobic granular sludge SBR can be successfully estimated using dynamic neural network model, particularly at high temperature.

Support Vector Regression Modelling of an Aerobic Granular Sludge in Sequential Batch Reactor

“…The studies of modelling of aerobic granulation are still in their infancy. Three previous studies of aerobic granulation were performed by [ 21 , 22 , 55 ]. Table 4 tabulates all the outcomes of these models, including this work.…”

RSM and ANN Comparative Modelling with a Granulation Treatment in Mixed Waters

“…Artificial neural networks (ANNs) can be employed to develop conceptual adaptive models for wastewater treatment processes that expand the boundaries of research and development [3]. ANN-based models have proved to be efficient and robust tools that can easily overcome the shortcomings of conventional mechanistic models [3,8]. However, research conducted so far on neural network-based models that may predict the behavior of an AGS system in wastewater treatment has been minimal [3,5,8,9].…”

“…ANN-based models have proved to be efficient and robust tools that can easily overcome the shortcomings of conventional mechanistic models [3,8]. However, research conducted so far on neural network-based models that may predict the behavior of an AGS system in wastewater treatment has been minimal [3,5,8,9]. The main challenge of modeling an AGS system resides in the complex physical, biological, and chemical processes involved, which include internal interactions among process variables and sludge characteristics [8].…”

“…However, research conducted so far on neural network-based models that may predict the behavior of an AGS system in wastewater treatment has been minimal [3,5,8,9]. The main challenge of modeling an AGS system resides in the complex physical, biological, and chemical processes involved, which include internal interactions among process variables and sludge characteristics [8]. On the other hand, the response surface methodology (RSM) is a versatile mathematical tool used for experimental design, modeling, and optimization [10].…”