Modeling of Activated Sludge Process Using Multi-Layer Perceptron Neural Networks

Mathematical Modeling of the activated sludge process (ASP) enhances the understanding of the process and improves the quality of the effluent released. However, as the process is complex and nonlinear, mathematical modeling of the process has been a challenge. In this study, multi-layer perceptron neural networks (MLP-ANN) are investigated to predict water quality parameters for better control of wastewater treatment plants employing an activated sludge process. The study area selected was in a central district of the southern state of India. The parameters to be investigated are biochemical oxygen demand (BOD), suspended solids (SS), and pH. The model is evaluated based on statistical parameters of correlation coefficient R and mean square error (MSE). The neural network toolbox of MATLAB 2015b is used for modeling and simulation study. It has been found that effluent biochemical oxygen demand was predicted with a maximum correlation coefficient of 0.927 and minimum mean square error of 0.0022, effluent suspended solids were predicted with a maximum correlation coefficient value of 0.947 and minimum mean square value of 0.0058, effluent pH was predicted with a maximum correlation coefficient value of 0.8299 and minimum mean square value of 0.0132.