Modeling of two-stage anaerobic onsite wastewater sanitation system to predict effluent soluble chemical oxygen demand through machine learning

Abstract

The present research aims to predict effluent soluble chemical oxygen demand (SCOD) in anaerobic digestion (AD) process using machine-learning based approach. Anaerobic digestion is a highly sensitive process and depends upon several environmental and operational factors, such as temperature, flow, and load. Therefore, predicting output characteristics using modeling is important not only for process monitoring and control, but also to reduce the operating cost of the treatment plant. It is difficult to predict COD in a real time mode, so it is better to use Complex Mathematical Modeling (CMM) for simulating AD process and forecasting output parameters. Therefore, different Machine Learning algorithms, such as Linear Regression, Decision Tree, Random Forest and Artificial Neural Networks, have been used for predicting effluent SCOD using data acquired from in situ anaerobic wastewater treatment system. The result of the predicted data using different algorithms were compared with experimental data of anaerobic system. It was observed that the Artificial Neural Networks is the most effective simulation technique that correlated with the experimental data with the mean absolute percentage error of 10.63 and R2 score of 0.96. This research proposes an efficient and reliable integrated modeling method for early prediction of the water quality in wastewater treatment.