Research On Simulation And Prediction Method Of Water Supply Network

As an important municipal infrastructure,the water supply system is the key to ensuring the stability of the people’s lives and the main constraint on the sustainable development of the city.At present,the management of water supply system in most cities is simple and crude.Due to the lack of technical research,most of the decision-making depends on human experience,which leads to water supply enterprises consuming a lot of energy,and the economic benefits are far from expected.In addition,because of the fact that the management of water supply network is short of three-dimensional display,simulation treatment of pipe burst accident and prediction analysis of water supply scheduling,it is unable to complete the demand of intelligent water supply.Aiming at these problems,this paper monitors the water pressure,flow and other information of the water supply network in real time,so as to achieve the purpose of dynamically tracking the operation status of the water supply network system.At the same time,the three-dimensional simulation and predictive scheduling are carried out to meet the needs of decision-making analysis and operation management of water supply network system.Finally,taking Fugu County of Shaanxi Province as an example to implement the simulation and prediction system of water supply network,which verifies the effectiveness and practicability of the method in this paper.This paper mainly completes the following work:(1)The three-dimensional simulation technology is applied to the system of water supply network.Based on the principle of triangle patch reorganization,this paper realizes the automatic construction of three-dimensional pipe network by using the basic data of pipe network.In addition,network management functions such as flow direction display,transparency adjustment and valve positioning are added.Furthermore,taking graph theory as the theoretical basis,this paper uses breadth-first search algorithm to study the the topological relationship of water supply network and complements the valve closing analysis in case of water supply network accident.Finally,the three-dimensional display and dynamic accident simulation processing of water supply network are realized.(2)Data statistical analysis technology is researched,including data synchronization technology of heterogeneous database,data visualization chart technology based on ECharts,and web crawler technology.The real-time synchronization of monitoring point pressure and flow is achieved through data synchronization technology of heterogeneous database,and the chart visualization display of data statistical analysis is realized based on Echarts.Finally,the weather data is obtained through web crawler technology,which provides the data analysis basis for the subsequent experimental research on residential water consumption forecast.(3)The Gated Circulation Unit(GRU)is applied to the prediction of the residential water consumption’s trend and the four-year residential water data of Fugu County,Shaanxi Province is selected as the experimental object.Firstly,pre-process the characteristic factors that affect the forecast of residential water use including weather,temperature,holidays,etc.Then use Recurrent Neural Network(RNN),Long Short-Term Memory Network(LSTM)and GRU to build a prediction model of daily water consumption of residents.By comparing the experimental results,it is proved that the GRU model has higher prediction accuracy than the other two models,and the average training time of the model is reduced by 29.73%compared with the LSTM model.(4)Based on the research of the above key technologies,this paper takes the construction of the Shaanxi Fugu water supply network information management system as an example,uses the JavaEE framework Spring,Spring MVC,Mybatis to implement the MVC design pattern.Then uses Bootstrap,AngularJS and other popular web development technologies to implement responsive layout.Besides,it adopts the deep learning framework Keras to conduct algorithm experiments and analysis of residential water consumption prediction.Thus,the paper completes the three modules of three-dimensional simulation on pipe network,data statistical analysis,and residential water consumption prediction.As a result,the research of the simulation and prediction system of water supply network has been realized.