北京水务Issue(5)：7-11,5.DOI:10.19671/j.1673-4637.2025.05.002
基于GESD的非居民用水异常分级识别研究
Study on graded identification of abnormal non-residential water consumption based on GESD
刘伟 1刘建国 1徐锁军 1齐艳冰 2张航2
作者信息
- 1. 北京市节水用水管理事务中心 北京 101117
- 2. 北京市水科学技术研究院 北京 100048
- 折叠
摘要
Abstract
In recent years,Beijing aggregates water consumption data from 189,000 non-residential water meters every month.These data are characterized by large volume,wide coverage,and fluctuations corresponding to production and business activities.Generally,there is a lack of effective big data analysis methods to support the effective identification of abnormal water consumption.To improve the quality of water consumption data and enhance the scientific decision-making,precise governance,and efficient services in water conservation management,a technical method for anomaly diagnosis of non-residential water consumption data was constructed based on the Generalized Extreme Studentized Deviate(GESD),and the results of abnormal water volume identification were classified by the degree of anomaly.The research results showed that in 2023,the number of non-residential water meters with abnormal water volume ranged from 961 to 2,191,accounting for 0.51％to 1.16％of the total number of water meters.Among them,water meters with Grade Ⅰ and Grade Ⅱ water volume anomalies were the focus of water consumption management.The number of such abnormal meters ranged from 240 to 807,accounting for 28.3％of the total number of water meters with abnormal water volume.The research results provide technical support for improving the accuracy of water consumption statistics and optimizing the water consumption statistical system.关键词
非居民用水/GESD/用水异常/异常分级Key words
non-residential water consumption/GESD/abnormal water consumption/anomaly grading分类
土木建筑引用本文复制引用
刘伟,刘建国,徐锁军,齐艳冰,张航..基于GESD的非居民用水异常分级识别研究[J].北京水务,2025,(5):7-11,5.
