Incremental Machine Learning and Genetic Algorithm for Energy Optimization of Biological Nutrient Removal in Wastewater Treatment Plants

Abstract
Wastewater treatment plants (WWTPs) play a crucial role in municipal infrastructure, but their energy consumption remains a significant concern. Among the various components of WWTPs, the aeration system in biological reactors stands out as an essential contributor to high energy usage. This system accounts for over 50% of the plant's total power consumption, removing organics and nitrogen. Supervisory Control and Data Acquisition (SCADA) systems are commonly employed to monitor dissolved oxygen (DO) concentration and regulate aeration blower to maintain a specific DO setpoint. However, despite the prevalence of SCADA systems, many WWTPs still grapple with challenges such as over-aeration and under-aeration caused by diurnal wastewater loading cycles, resulting in increased energy usage. This research introduces a predictive aeration optimization tool tailored to a full-scale biological nutrient removal WWTP to address this issue. A concept-drift independent incremental-learning model based on K-Nearest Neighbor (KNN) was developed to predict air blower flow rates, achieving an R2 value greater than 85%. This incremental-learning model further serves as an objective function for a Genetic Algorithm (GA) optimization to minimize air blower flow rates while ensuring that final effluent properties meet treatment quality limits in compliance with regulatory requirements. The model was trained and validated using online sensors and laboratory data collected between 2012 and 2022, with measurements recorded from 10 minutes to daily intervals. The optimization approach reduced aeration requirements by 14% without compromising effluent quality, resulting in a theoretical average of 2,035.7 kWh in daily energy savings. Furthermore, the model was implemented over a period of 30 days using PI-Datalink and PyXLL by utilizing Microsoft Excel user environment, yielding a prediction error less than 9% and 18% reduction in aeration requirement. This tool presents a novel predictive optimization approach in a familiar and user-friendly environment to equip operators with optimized air blower flow rates based on wastewater characteristics and effluent requirements, thereby enhancing energy efficiency.
Description
Keywords
Wastewater treatment plants, Energy Optimization, Machine learning, Incremental learning, K-Nearest Neighbor (KNN), Genetic Algorithm
Citation
Monday, C. (2024). Incremental machine learning and genetic algorithm for energy optimization of biological nutrient removal in wastewater treatment plants (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.