Machine Learning-Based Quantitative Model of Coupled Water-Energy-Chemical-Carbon Efficiency of Wastewater Treatment Plants

As typical urban facilities processing several types of resources, wastewater treatment plants (WWTPs) consume chemicals and electricity to remove pollutants, generating substantial greenhouse gas (GHG) emissions. Identifying synergistic mechanisms for reducing pollution and carbon emissions from WWTPs is essential to optimizing energy and chemical use. In this study, we propose an integrated water-energy-chemical-carbon (WECC) efficiency model that reveals the synergistic pathways for pollutant removal and carbon reduction. The model identifies optimized operational parameters by assessing the WECC efficiency of resource inputs, pollutant removal, and carbon emission from WWTPs, and predicts the optimal influent and effluent quality based on a multilayer perceptron (MLP) model. The results show that carbon source chemicals in the assessed WWTP are the main limiting factors for synergies, accounting for 40% of total chemical-related emissions. However, introducing carbon source chemicals to increase the influent organic matter concentration can reduce energy and carbon intensities by up to 34.64% and 40.96%, respectively. For other factors, synergistic effects on pollution removal and carbon reduction are expected to be achieved by increasing influent COD/TN ratio, setting a reasonable safe discharge threshold, and optimizing chemical use. This machine learning-based model of the coupled WECC efficiency is expected to improve the performance of WWTPs.