Research On Leakage Control Of Water Supply Pipe Network Based On Hydraulic Model

This paper mainly studied the partition optimization of water supply network based on hydraulic model and real-time location of leakage loss,which provided reference for leakage control management.The research work and main conclusions were as follows:1.?The MicroGE model of the target area pipe network was established using WaterGEMs software.Based on the pipe network and flow data collected by the GIS system,the water consumption curve was determined,the C-value of the pipe section Hai Zeng William coefficient was estimated,the topology was simplified and the node flow was distributed,and finally the delayed hydraulic model consisting of 727 node,1048 strip segments was formed.Through checking model based on the WaterGEMs Darwin corrector,the results showed that the six pressure points and the three flow points were highly identical with the result,which satisfied the model checking criteria and can provide the next partition optimization model and leakage location model with the data support.2?The pipe network partition optimization model based on graph theory was established.In order to solve the problem of strong randomness of experience partition,this paper proposed a method of partition calculation based on the shortest path principle of graph theory,and evaluated the leakage level of pipeline network by using nighttime minimum flow as evaluation index.Applying the model to the study area,the results showed that the scheme leakage after partitioning was lower than that before partitioning,and with the different weight functions selected,the leak control effects of each partition were quite different..The use of graph theory for pipe network partitioning not only improved the efficiency and accuracy of the partition,but also effectively controlled and reduced leakage.3?The leakage positioning model based on a genetic algorithm(GA)to optimize BP neural network was established.Because BP neural network had the disadvantage of difficult convergence,this paper optimized GA based on BP neural network.Firstly,by analyzing the variation characteristics of the pressure monitoring points,a leakage database of 76 leaky databases of the target area and 380 sets of samples were established,and then the GA-BP and BP neural network were used to train the leakage database.The results showed that the former had a smaller mean square error,a fewer iteration and a higher data fitting degree than the latter.The leakage positioning model of GA-BP was constructed to predict the pipeline of the study area.The results showed that the prediction error of the leak point is within 200m,and the model can be applied to the actual pipe network.