净水技术2025，Vol.44Issue(6)：47-53,7.DOI:10.15890/j.cnki.jsjs.2025.06.006
基于谱聚类分区算法的供水管网漏损区域识别技术
Identification for Water Supply Network Leakage Area Based on Spectral Clustering Partitioning Algorithm
摘要
Abstract
[Objective]The current water supply network leakage problem is widespread and increasingly prominent.How to effectively identify the location where leakage occurs and reduce the leakage rate has been a hot topic of discussion in water supply industry in recent years.The study introduces a virtual pressure partitioning approach to enable regional leakage identification in water supply networks,aiming to improve the precision of leakage detection and thereby reduce the overall leakage rate.[Methods]In this study,the spectral clustering algorithm was utilized to divide the pipe network into virtual zones according to the topology of the water supply pipe network.The same leakage ratio coefficient was taken for the water nodes within the same zone,and this coefficient was used as the decision variable.Then,the regional distribution of leakage in the hydraulic model was inferred based on the strengthen elite genetic algorithm(SEGA),with the minimization of the difference between the calculated value of the pressure at pressure points and the actual value in hydraulic model taken as the objective function.The feasibility of the leakage region identification technology was verified through actual engineering cases.[Results]When applied to the water supply network in Nanqiao Town,Fengxian District,Shanghai,results indicated that compared to directly allocating leakage volume proportionally in the hydraulic model,the simulated values at pressure monitoring points were more stable and closer to actual values.Furthermore,it was inferred that the distribution of leakage water volume in this model was also more in line with the actual leakage situation.[Conclusion]The leakage area identification technology can provide a theoretical reference in terms of the priority order of leak detection work,point out the areas that need to be focused on for leakage control in the pipeline network system,and indicate the priority areas and directions for the hardware leak detection work that ultimately achieves precise location of leakage points.关键词
供水管网/谱聚类/漏损识别/水力模型/压力监测点Key words
water supply pipelines network/spectral clustering/leakage identification/hydraulic model/pressure monitoring points分类
土木建筑引用本文复制引用
王静莹,黄强,周斌,安淑萍,汪家泉,信昆仑..基于谱聚类分区算法的供水管网漏损区域识别技术[J].净水技术,2025,44(6):47-53,7.基金项目
国家自然科学基金(52270093,51978494) （52270093,51978494）
