活性污泥模型研究综述:从机理拓展到混合智能建模OA

Activated sludge models(ASMs)have been widely applied in wastewater treatment since 1987.According to the database of the Web of Science(1997-2025),this study systematically reviewed the research progress of the ASM by considering model extension,parameter calibration,and hybrid modeling.First,the development from ASM1 to ASM3 and their limitations of complex microbial metabolism were summarized.Results showed that introducing new state variables and pathways remained important to improving model applicability.Parameter calibration strategies were discussed by integrating sensitivity analysis with intelligent optimization algorithms,which could enhance model reliability.Moreover,process optimization,treatment of special wastewaters,and resource recovery of ASMs were analyzed based on real applications.Five paradigms for integrating ASMs with machine learning,including residual correction,parameter learning,input augmentation,structure replacement,and constraint-embedded approaches,were elucidated,together with the comparison of their characteristics and scenarios.Lastly,perspectives on enhanced mechanistic representation,adaptive calibration,digital twin platforms,and cross-scale modeling were presented to support intelligent wastewater treatment.