Forecasting wastewater flows and pollutant loads: A comparison of data-driven models within the urban water system framework

作     者：Giberti, Matteo Dereli, Recep Kaan Bahramian, Majid Flynn, Damian Casey, Eoin 

作者机构：Univ Coll Dublin Sch Chem & Bioproc Engn Dublin Ireland Technol Univ Dublin Environm Sustainabil & Hlth Inst Dublin Ireland Univ Coll Dublin Sch Elect & Elect Engn Dublin Ireland 

出 版 物：《JOURNAL OF ENVIRONMENTAL CHEMICAL ENGINEERING》 (J. Environ. Chem. Eng.)

年 卷 期：2024年第12卷第5期

核心收录：

学科分类：0830[工学-环境科学与工程（可授工学、理学、农学学位）] 0817[工学-化学工程与技术] 08[工学] 

基　　金：Science Foundation Ireland (SFI) under the SFI Strategic Partnership Programme [SFI/15/SPP/E3125] 

主　　题：Autoregressive models Benchmark simulation model Forecasting Wastewater Neural network 

摘      要：The ability to forecast key features of wastewater treatment plant (WWTP) influent, is emerging as an important tool to enable advanced WWTP operational strategies. Various data driven models have been reported in the scientific literature for WWTP influent forecast however, there has been no quantitative comparison of them against each other. The present study is the first to undertake this comparison utilising a high-frequency reference dataset generated from the Urban Water System model. Specifically, the performance of autoregressive models was compared against time-delay networks, nonlinear autoregressive networks and long shortterm memory networks. Time-delay networks were generally found to outperform the other tested methods, although the reliability of the generated forecasts decreases as the prediction horizon exceeds one hour. While longer prediction horizons would be desirable, there is a trade-off between model accuracy and overall optimisation of plant operation. This study also highlights the challenge of dealing with both concentration and flow variations suggesting the need for future research to analyse the separate impacts of flow and concentration variations on model performance.