Francesca Bellamoli, Marco Vian, Mattia Di Iorio, Farid Melgani
{"title":"通过主动学习策略对污水处理厂的异常分类进行领域适应。","authors":"Francesca Bellamoli, Marco Vian, Mattia Di Iorio, Farid Melgani","doi":"10.2166/wst.2024.387","DOIUrl":null,"url":null,"abstract":"<p><p>The increasing use of intermittent aeration controllers in wastewater treatment plants (WWTPs) aims to reduce aeration costs via continuous ammonia and oxygen measurements but faces challenges in detecting sensor and process anomalies. Applying machine learning to this unbalanced, multivariate, multiclass classification challenge requires much data, difficult to obtain from a new plant. This study develops a machine learning algorithm to identify anomalies in intermittent aeration WWTPs, adaptable to new plants with limited data. Utilizing active learning, the method iteratively selects samples from the target domain to fine-tune a gradient-boosting model initially trained on data from 17 plants. Three sampling strategies were tested, with low probability and high entropy sampling proving effective in early adaptation, achieving an F2-score close to the optimal with minimal sample use. The objective is to deploy these models as decision support systems for WWTP management, providing a strategy for efficient model adaptation to new plants, and optimizing labeling efforts<b>.</b></p>","PeriodicalId":23653,"journal":{"name":"Water Science and Technology","volume":"90 11","pages":"3123-3138"},"PeriodicalIF":2.5000,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Domain adaptation through active learning strategies for anomaly classification in wastewater treatment plants.\",\"authors\":\"Francesca Bellamoli, Marco Vian, Mattia Di Iorio, Farid Melgani\",\"doi\":\"10.2166/wst.2024.387\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>The increasing use of intermittent aeration controllers in wastewater treatment plants (WWTPs) aims to reduce aeration costs via continuous ammonia and oxygen measurements but faces challenges in detecting sensor and process anomalies. Applying machine learning to this unbalanced, multivariate, multiclass classification challenge requires much data, difficult to obtain from a new plant. This study develops a machine learning algorithm to identify anomalies in intermittent aeration WWTPs, adaptable to new plants with limited data. Utilizing active learning, the method iteratively selects samples from the target domain to fine-tune a gradient-boosting model initially trained on data from 17 plants. Three sampling strategies were tested, with low probability and high entropy sampling proving effective in early adaptation, achieving an F2-score close to the optimal with minimal sample use. The objective is to deploy these models as decision support systems for WWTP management, providing a strategy for efficient model adaptation to new plants, and optimizing labeling efforts<b>.</b></p>\",\"PeriodicalId\":23653,\"journal\":{\"name\":\"Water Science and Technology\",\"volume\":\"90 11\",\"pages\":\"3123-3138\"},\"PeriodicalIF\":2.5000,\"publicationDate\":\"2024-12-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Water Science and Technology\",\"FirstCategoryId\":\"93\",\"ListUrlMain\":\"https://doi.org/10.2166/wst.2024.387\",\"RegionNum\":4,\"RegionCategory\":\"环境科学与生态学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2024/11/27 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q3\",\"JCRName\":\"ENGINEERING, ENVIRONMENTAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Water Science and Technology","FirstCategoryId":"93","ListUrlMain":"https://doi.org/10.2166/wst.2024.387","RegionNum":4,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/11/27 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"ENGINEERING, ENVIRONMENTAL","Score":null,"Total":0}
Domain adaptation through active learning strategies for anomaly classification in wastewater treatment plants.
The increasing use of intermittent aeration controllers in wastewater treatment plants (WWTPs) aims to reduce aeration costs via continuous ammonia and oxygen measurements but faces challenges in detecting sensor and process anomalies. Applying machine learning to this unbalanced, multivariate, multiclass classification challenge requires much data, difficult to obtain from a new plant. This study develops a machine learning algorithm to identify anomalies in intermittent aeration WWTPs, adaptable to new plants with limited data. Utilizing active learning, the method iteratively selects samples from the target domain to fine-tune a gradient-boosting model initially trained on data from 17 plants. Three sampling strategies were tested, with low probability and high entropy sampling proving effective in early adaptation, achieving an F2-score close to the optimal with minimal sample use. The objective is to deploy these models as decision support systems for WWTP management, providing a strategy for efficient model adaptation to new plants, and optimizing labeling efforts.
期刊介绍:
Water Science and Technology publishes peer-reviewed papers on all aspects of the science and technology of water and wastewater. Papers are selected by a rigorous peer review procedure with the aim of rapid and wide dissemination of research results, development and application of new techniques, and related managerial and policy issues. Scientists, engineers, consultants, managers and policy-makers will find this journal essential as a permanent record of progress of research activities and their practical applications.