Seasonal Setpoints Optimization of WWTP DO Control Based on Artificial Neural Networks Performance Indices Prediction

• Published in: Computer Aided Chemical Engineering Vol. 53; pp. 1615 - 1620
• Main Authors: Mihály, Norbert B., Cristea, Vasile M.
• Format: Book Chapter
• Language: English
• Published: 2024
• Subjects: aeration control; genetic algorithm; greenhouse gas emissions; performance indices; seasonal artificial neural network; wastewater treatment plant; genetic algorithm; greenhouse gas emissions; seasonal artificial neural network; wastewater treatment plant; performance indices; aeration control
• ISBN: 9780443288241; 0443288240
• ISSN: 1570-7946
• DOI: 10.1016/B978-0-443-28824-1.50270-2

Adaptive and optimal setpoints for the aeration control system are necessary to maintain the high performance of wastewater treatment plant operation under generally variable seasonal weather conditions. By effectively achieving the optimization aims, models with shorter computation times and more dependable forecasts are highly valued components for real-time optimization activities. In order to forecast the wastewater treatment plant's Greenhouse Gas Emissions and Effluent Quality performance indicators, artificial neural network models were designed, trained, and evaluated. We took into consideration the nonlinear autoregressive network with exogenous inputs network type. The models of artificial neural networks were developed using data particular to each season. Using evolutionary algorithm optimizations and two distinct selection techniques based on the Pareto fronts of the two considered performance indicators, the optimal artificial neural network architecture and hyperparameters were identified for each of the four seasons. When tested, the trained network models showed high forecast accuracy for all seasons, with mean absolute percentage error values for the greenhouse gas emissions reaching 2.88% and the effluent quality index up to 4.25%. The optimization of aeration led to improvements in Effluent Equality, Greenhouse Gas Emissions, and Operational Cost performance throughout all seasons. The improvements ranged from 0.40% for Greenhouse Gas emissions to 13.31% for Effluent Quality Index.