Activated sludge models at the crossroad of artificial intelligence—A perspective on advancing process modeling

Abstract

The introduction of Activated Sludge Models No. 1 (ASM1) in the early 1980s has led to a decade-long experience in applying these models and demonstrating their maturity for the wastewater treatment plants’ design and operation. However, these models have reached their limits concerning complexity and application accuracy. A case in point is that despite many extensions of the ASMs proposed to describe N2O production dynamics in the activated sludge plants, these models remain too complicated and yet to be validated. This perspective paper presents a new vision to advance process modeling by explicitly integrating the information about the microbial community as measured by molecular data in activated sludge models. In this new research area, we propose to harness the synergy between the rich molecular data from advanced gene sequencing technology with its integration through artificial intelligence with process engineering models. This is an interdisciplinary research area enabling the two separate disciplines, namely environmental biotechnology, to join forces and work together with the modeling and engineering community to perform new understanding and model-based engineering for sustainable WWTPs of the future.

Modeling, simulation and control of biological and chemical P-removal processes for membrane bioreactors (MBRs) from lab to full-scale applications: State of the art

“…Artificial intelligence-based model controls such as artificial neural networks (ANN), deep learning neural networks (DNN), fuzzy techniques, and hybrid controls (neuro-fuzzy) are adapted in wastewater (Garikiparthy et al, 2016). However, their success is not satisfactory for complex phenomena such as P-removal (Sin and Al, 2021). Logarithm-based phosphorus control systems are deployed and are automated using SCADA to reduce consumption of the precipitants (to reduce cost) and maintain effluent compliance.…”

Retechnologization of biological sewage water treatment facilities using mathematical modelling

“…In the current situation, with the «traditional» scheme, the low efficiency of biological treatment processes, namely in the operation of aeration tanks, is revealed, and achieving the requirements for the discharge of pollutants is impossible without the use of special technologies that require the reconstruction of treatment facilities and the re-technology of biological treatment facilities. The simulation of the operation of biological treatment facilities was carried out using a computer program based on the mathematical models ASM (ASM2d, ASM3), EAWAG, which were first published in 2000-2003 [14,20,21]. The developer of the models is a special group for mathematical modeling of the International Water Association [22].…”

A Review of Computational Modeling in Wastewater Treatment Processes

“…In an attempt to adapt the models to changes in WWTPs, i.e., process upgrades and introduction of new treatments, or even more strict effluent discharge limits, new models and/or extensions to existing models have been developed . However, these changes result in an increase in the model complexity, making them too parametrized and difficult to calibrate . Therefore, their popularity has decreased over the last years, as can be observed in Figure , where the number of publications related to the Activate Sludge Model (one of the most used mechanistic models) is in decline.…”