The Study And Implementation Of Robust Leakage Detection For Underground Water Supply Network Based On Hydraulic Pressure

In recent years,with the rapid development of economy,the demand for water in China is increasing.This puts increasing pressure on the water supply system.Leakage in water supply network is also becoming more and more serious,especially in some water supply network systems which have been put into use earlier.Long-term water leakage will not only waste water resources and cause economic damage,but also lead to insufficient fire water pressure,geological collapse,water pollution and other safety issues.Therefore,leakage hot-spot detection becomes the core task of water supply network,and its high efficiency and intelligence are particularly important.At present,artificial leakage detection technology is usually used to detect leakage points.However,due to its limitations,it is more and more difficult to continuously reduce the leakage rate of water supply pipeline by manual leakage detection.With the development of artificial intelligence,it has become a popular means of monitoring the water supply network in real time actively and mastering the leakage status in time in leakage control.Therefore,this paper aims to build a robust model of water supply network based on pressure to effectively monitor the leakage points in the underground network that are difficult to be found,so as to solve the situation that the underground network is unsupervised and difficult to be supervised for a long time.This paper first describes the existing work of leak detection in water supply network.Based on the analysis of the existing techniques,a robust model based on pressure was proposed to solve the core problems in leak detection and applied in the real work.Specifically,this paper mainly has the following contributions:1.Robust hydraulic model correction method based on grouping measurementBased on water consumption grouping and water balance theory,a robust hydraulic model based on hydraulic pressure is established in this paper.The objective function of this model is the robust Huber loss function.For this model,we use the fm GA algorithm to obtain an estimator.2.In this paper,the new proposed robust optimization model is used to analyze the simulated case,a serious leakage accident in the water supply network system of Beijing Jiaotong University and DMA data in northern England.Experimental results show that the new model can accurately detect leakage points,especially those which have proved difficult to detect.