多目标大规模供水管网监测点的优化选址  

Multiobjective genetic algorithms for optimal monitoring station placement in large water distribution systems

作　　者：刘书明[1] 王欢欢[1] 徐鹏[2] 徐速[2] 

机构地区：[1]清华大学环境学院,北京100084 [2]北京清华城市规划设计研究院,北京100085

出　　处：《清华大学学报（自然科学版）》2013年第1期78-83,共6页Journal of Tsinghua University(Science and Technology)

基　　金："十一五"水体污染控制与治理重大专项经费资助项目(2009ZX07425-006)

摘　　要：针对供水管网水力模型中监测点布置的优化计算不易得到确定结果、运行数据不能完全代表管网状况的问题,提出反映管网节点水量和节点压力变化的节点覆盖水量和节点相关水压的多目标遗传算法,综合灵敏度分析、爆管分析,结合测压点布设原则、管网拓扑结构、水力计算结果最终确定测压点优化选址的方法。该方法用于城镇大规模实际管网中,采用管网节点中的0.3%个节点为水压监测点即可覆盖约45%的管网节点与43%的管网水量,通过与现有布设原则比较,发现其结果与布设原则匹配度较高。该方法提高了管网模型测压点的精度和准度,也代表了管网中水压变化和水量影响,可有效消除人为因素影响和提高监控管网漏水、爆管、能耗的效果。Supervisory control and data acquision （SCADA） systems are used to estimate the energy consumption and water quality in water distribution systems （WDS）. The number and locations of the monitoring stations are generally selected prior to the SCADA layout; thus, the station placement is critical. The aim of this study is to develop a method to optimally locate monitoring stations as a multiobjective optimization problem. The problem is solved using a multiobjective genetic algorithm （MOGA） which maximizes the number of covering nodes and the monitoring station demand. This method is then combined with the topology and a hydraulic sensitivity analysis of burst pipes in a large scale urban WDS. Monitoring stations in only 0. 3% of the nodes can monitor 45% of the nodes and 43% of the demand in a large urban network. The result matches the performance with the layout. The method improves monitoring station placement accuracy and balances the influence of node pressure and demand. In addition, this reduces jamming and provides information for improved operations in terms of energy consumption, leakage and bursting.

关 键 词：给水管网 优化选址 压力监测点 多目标遗传算法 

分 类 号：TU991.33[建筑科学—市政工程]