融合凝聚层次聚类和互相关函数的城市排水管网监测点优化布置研究

Optimizing the layout of urban drainage pipeline monitoring points using agglomerative hierarchical clustering and cross-correlation functions

摘要

【目的】随着城市排水管网在洪水管理和水污染控制中的重要性日益凸显，对其监测网络的优化布置提出了迫切需求。【方法】提出一种高效、准确的城市排水管网监测点优化布置方法，以监测和识别管网污水非法排放。采用暴雨管理模型(SWMM)模拟城市排水管网中污染物的运移过程，获取污染物时间序列数据；利用凝聚层次聚类算法对数据进行分类处理，确定排水管网中最优监测点的数量；通过互相关函数评估节点间的相关性，选择互相关系数最大的节点作为监测点的位置，以确保全面反映簇内的污染情况。【结果】使用SWMM手册中的排水管网案例验证了所提方法的有效性，选择部署3个监测点为最优选择。监测点布置效果评价表明，监测点数为3时，监测可靠度达到91.86%,监测平均响应时间为3.26 min。凝聚层次聚类在监测点布置效果上优于K-means算法，并且在全面覆盖排水管网上游、中游和下游方面表现更为出色。【结论】提出的城市排水管网监测点布局优化方法为城市排水管网监测与管理提供了新的技术手段，为相关领域的研究提供了新思路。

Abstract

[Objective] With the increasing importance of urban drainage pipelines in flood management and water pollution control, there is an urgent need for optimized monitoring networks. [Methods] An efficient and accurate method for optimizing the layout of monitoring points in urban drainage pipelines to detect and identify illegal sewage discharge was proposec. A Storm Water Management Model(SWMM) was utilized to simulate pollutant transport within urban drainage pipelines and generate time-series data. An agglomerative hierarchical clustering algorithm was applied to classify the data and determine the optimal number of monitoring points in the drainage network. Node correlation was assessed through cross-correlation functions, selecting nodes with the highest coefficients to ensure comprehensive pollution conditions within the clusters. [Results] The proposed method was validated using a drainage network case from the SWMM manual, and it was determined that deploying three monitoring points is the optimal choice. The evaluation revealed a reliability of 91.86% and an average response time of 3.26 minutes at three monitoring points. Agglomerative hierarchical clustering outperformed the K-means algorithm in terms of monitoring point layout effectiveness, especially in covering the upstream, midstream, and downstream sections of the drainage network. [Conclusion] This method offers a new technical means for optimizing the layout of monitoring points in urban drainage pipelines, enhancing urban drainage monitoring and management and offering new perspectivesfor related research areas.

关键词

城市排水管网 / 监测点优化布置 / 凝聚层次聚类算法 / 互相关函数 / SWMM模型

Key words

urban drainage pipelines / monitoring point optimization / agglomerative hierarchical clustering algorithm / cross-correlation function / SWMM model

基金资助

深圳市可持续发展科技专项项目(KCXFZ20201221173407021)

国家自然科学基金项目(52109025)

山东省自然科学基金项目(ZR2021ME030)