Research On Leakage Control Based On Water Supply Network Modeling

Urban water supply network is an important infrastructure to ensure people's life and promote urban development.However,with the advancement of urbanization in China,the leakage rate is high.Due to the hidden nature of the underground water network,some pipelines are often found to be destroyed after a long period of time.This not only results in the loss of treated water resources and affects the economic efficiency of the water supply companies,impurities may enter the pipeline through the cracks and may also pollute the water quality,bringing security risks.Leakage will cause the network operating conditions to change,while the pressure data monitored by pressure monitoring points implies the information of network operating status.Therefore,this paper studies and analyzes the effect of different levels of pipeline damage on the monitoring point pressure in the pipe network,and uses the BP neural network technology to establish the nonlinear relationship between the leakage position and the pressure change rate of the pressure measurement points.The realization of the input of only a few pressure measurement point pressure change data can reversely deduced the position of the leakage point,achieve the purpose of real-time monitoring of leakage,and provide theoretical basis and technical support for the extensive application in the actual pipe network in the future.First,use the EPANET software to establish a hydraulic model of the actual water supply network.Through the analysis of water sales records and on-site investigations,the water users in the study area were divided into urban water consumption,rural household water,business service water,office water,industrial water and collective unit water according to the nature of water use.For the six categories of users,the actual amount of water consumed for one week was measured,and then the curve of 24-hour change of the amount of water consumption for each type of user was obtained through curve fitting.Through field surveys,nine cast iron pipes with different pipe diameters,different buried years,and different locations were selected.Three-point pressure measurement was used to determine the resistance coefficient of the pipes.Fitting the resistance coefficient model of water supply pipelines with pipe laying age and pipe diameter as coefficient::C=65.345e^{0.01537??-1950?}+0.03827D+15.39,which provides important parameters for the construction of the hydraulic model.Furthermore,the topology of the water network is simplified,the distribution of node flow for large user water consumption,small user water consumption,unmeasured water quantity and water leakage of pipe network in the study area is carried out by the combination of various distribution modes.Finally,a simulation hydraulic model of water supply network with 122 nodes and 132 sections is established.The pipelines were selected in the upper,middle and lower reaches of the network of the established hydraulic model to simulate the leakage with different damage levels?the leakage area ratios Ka were 0.01,0.03,0.05,0.1 and 0.2,respectively?.Finally,70 buried pipes in the target network are selected to simulate the leakage of different damage degrees,and the pressure variation data of the pressure monitoring point are obtained.These data are used as training samples to train the neural network to establish the nonlinear relationship between the leakage point position and the nodal pressures,a real-time leaky point location model based on BP neural network is constructed.Untrained pipelines are selected for leak point location experiments.The results show that the model's leakage point location diagnosis was feasible.