Modeling of the activated sludge process by using artificial neural networks with automated architecture screening

In this study, a MATLAB script was developed to aid in the development of artificial neural network (ANN) models by screening out the better ANN architectures for the cases studied. Then, the script was applied for modeling of activated sludge process (ASP) for two different cases. In the first one, a hypothetical wastewater treatment plant (WWTP) was considered. The input and output data for the training of the ANN models were generated using a simulation model, which was an implementation of the Activated Sludge Model No. 1 (ASM1). The results indicated high correlation coefficient (R) between the observed and predicted output variables, reaching up to 0.980. In the second case, ANN modeling of ASP in the Iskenderun Wastewater Treatment Plant (IskWWTP) was studied. Resulting maximum R value was 0.795 for the predicted effluent chemical oxygen demand (CODeff) values. Moreover, CODeff was forecasted using another effluent parameter.