A review of mechanistic and data-driven models of aerobic granular sludge. Issue 3 (June 2022)

• Record Type: Journal Article
• Authors: Zaghloul, Mohamed Sherif; Achari, Gopal
• Is Part Of: Journal of environmental chemical engineering. Volume 10:Issue 3(2022)
• Publication Date: 2022-06
• Subjects: Aerobic granular sludge -- Modeling -- Kinetics -- Biofilm models -- Computational fluid dynamics -- Machine learning; Chemical engineering -- Environmental aspects -- Periodicals; Environmental engineering -- Periodicals; Chemical engineering -- Environmental aspects; Environmental engineering; Periodicals; 660.0286
• DOI: 10.1016/j.jece.2022.107500
• Languages: English
• ISSNs: 2213-2929

Abstract

Abstract: The advantages of aerobic granular sludge sequencing batch reactors over conventional wastewater treatment methods are propelling the technology forward to full-scale application. The development of simulation models plays an influential role in understanding the process dynamics and predicting the process behavior, which is critical during the transition from laboratory and pilot studies to the design of full-scale plants. Simulation models allow virtual testing with approximate results to guide the expensive real-life implementation. This work reviews the current state of the literature on modeling aerobic granular sludge sequencing batch reactors, focusing on the objectives of the models, modeling methods, and the current trends. The most common modeling approaches were found to adopt mathematical models with many assumptions and process simplifications that were usually adopted from preceding studies. Mathematical modeling provided a fundamental understanding of the micro and macro scale bio-chem-physical processes that simultaneously occur inside aerobic granular sludge reactors. The common conclusion derived from these studies was that mathematical modeling could be overly complicated and computationally demanding when the models were more comprehensive. This review explores the current trend in the literature to develop models that can provide good performance while keeping the modeling objectives in mind. Further, the application of different machine learning and data-driven models is investigated. Finally, this review provides suggestions for future research needed to achieve better comprehensive models for the full aerobic granular sludge process with the fewest assumptions. Graphical Abstract: ga1 Highlights: Kinetic models are the most commonly available models in the literature. Machine learning is effective and gaining popularity. Image analysis and CFD are promising modeling approaches. The intended application of the model dictates the method to use. More comprehensive yet computationally efficient models are needed.