Performance prediction of trace metals and cod in wastewater treatment using artificial neural network

Abstract

Artificial intelligence is finding its ways into the mainstream of day-to-day operations. Novel AI application techniques such as the artificial neural network (ANN), fuzzy logic, genetic algorithms and expert systems have gained popularity in the fourth industrial revolution era. Due to the chemical composition, inherent complexity, incoherent flow rate and higher safety factor in the effective operation of the biological wastewater treatment process, the AI-based model was extensively tested in managing the wastewater treatment operations. The interrelationship between COD and trace metals was studied using an AI-based prediction model with ANNs incorporated in MATLAB. A supervised learning algorithm was used for training the ANNs and to relate input data to output data. The training was aimed at estimating, validating, predicting the parameters by an error function minimization. The goodness of the prediction was attained with the coefficient of determination (R2) of 0.98–0.99, a sum of square error (SSE) 0.00029–0.1598, room mean-square error (RMSE) of 0.0049–0.8673, mean squared error (MSE) 2.7059e-14 to 2.3175e-15. The ANNs models were found to be a robust tool for predicting WWTP performance. The predictive approaches can be used in the prediction of environmental management and other emerging technologies. This will meet the cost-effectiveness, effective environmental and technical criteria with a wide range of big-data support and implementation of the sustainable development goals, circular bio-economy and industry 4.0.