Rethinking Activated Sludge Modeling: A Critical Review of Modeling Strategies and the Role of Hybrid Integration

作者：Jaydev Zaveri, Guangyu Li, Zijian Wang, Yan Yuan, Péter Budai, Imre Takács, April Z. Gu · 发表于：Water Environment Research · 年份：2025 · DOI：10.1002/wer.70181 · 被引用次数：8 · 研究领域：Wastewater Treatment and Nitrogen Removal

Efficient control of wastewater treatment is vital for achieving the United Nations' Sustainable Development Goals (SDGs). Increasing industrialization and population growth threatens aquatic ecosystems, necessitating effective clean water solutions and enhanced efficiency through advanced technologies. These technologies need to focus on innovative methods to reduce energy use and reliance on external chemicals. Conventional mechanistic (henceforth referred to as white-box) modeling techniques, such as Activated Sludge Models, have significantly advanced our understanding of activated sludge processes by providing a structured and mechanistic framework. However, challenges remain in capturing dynamic microbial interactions and responding to influent variability, particularly in real-world applications that are outside the realm of the mechanistic framework. This paper explores the use of artificial intelligence (AI) (henceforth referred to as black-box) techniques in modeling practices, particularly machine learning, a subset of AI, to address the limitations faced by white-box models. AI models provide a powerful data-driven approach to wastewater treatment modeling, enabling them to identify complex patterns from large datasets. However, their effectiveness depends on high-quality training data, and their reliance on statistical correlations limits interpretability. This review advocates for the integration of hybrid modeling in the current modeling practice, which combine...