Study On The Identification Of The Water Distribution Network Burst In The Hunnan New District Based On Data Analysis

China's total water resources is abundant,but the per capita share is less than the world average.And the distribution of water is very uneven in time and space.There are serious water shortages in the north and northwest of China.In this context,China's water supply management level is not high due to historical reasons.Loss of water supply is serious,the pipeline is often burst in many cities nationwide.In a short period of time the loss of water is very large due to burst pipe.This is a huge damage to the water supply for the surrounding area and the economic benefits of the water division.Timely detection and handling of pipe accident has a positive significance to the regional water supply management and protection.In this paper,the information of the abnormal condition is obtained by simulating the explosion condition in the micro hydraulic model of the water supply pipe network.The pressure monitoring data and the simulation data are processed and analyzed using the algorithm.The method of identifying the abnormal condition data and the method of analyzing the position of the leak point are established.Through the study of neural network algorithm,self-organizing mapping algorithm and sparse self-coding and learning Java programming language,we can edit the computer platform to achieve the algorithm and test run.The compiled program provides a core computing tool for the later analysis model.Because the actual pipe network is difficult to carry out burst test,so use the micro-hydraulic model to simulate the explosion situation.The establishment of micro-hydraulic models is a time-consuming and laborious task.Through the investigation and analysis of the detailed information of the water supply network of H area,we get the information of pipe network topology,valve distribution,terrain elevation,pipe diameter distribution,water source water supply management,pressure monitoring history,user water consumption and so on.Organize these information and enter the modeling software as required,and the software will perform hydraulic calculations.The model can be checked by the measured pressure monitoring value,so that the model can reflect the operation of the actual pipe network.The neural network algorithm is a mathematical model established by the basic principle of human brain system.It can use the internal large number of weight value parameters to study the nonlinear relationship between analog input value and output value.We can achieve the prediction of water supply pressure through a lot of historical data training.The pressure at the time of bursting and the results of the timing prediction are compared because the water supply pressure is changing every hour.According to the numerical distribution of the difference between the two operating conditions,it is found that the difference under normal working condition is subject to normal distribution.This is an important basis for the identification of working conditions.The numerical relationship between the pressure of each monitoring point under normal working condition is simulated by neural network,and the spatial prediction relation model is formed.The results of the model under normal and abnormal conditions are compared and analyzed,and it is found that the variance is much larger than normal.This is an important basis for job identification.A classifier model of data sequence is established by combining sparse self-coding and SOM self-organizing map network.The model is used to cluster the normal working pressure information and the simulated abnormal pressure information,so as to establish the classification information base of the pipe network as the basis for the analysis of pressure monitoring information.Combining the difference between the measured sequence of the pressure and the sequence of the time series is studied,and the possible area of the burst is analyzed according to the distribution of the difference.Through the new simulated burst pipe to verify the situation,the successful identification of the conditions for the abnormal conditions.The possible location of the leakage point is analyzed by the method of regional location analysis,and the position of the actual leakage point falls within the analysis area.