Fault Diagnosis Of Water Supply Pipe Network Leakage Based On BP Neural Network

With the rapid development of cities,the increasing use of water,the frequent occurrence of leakage has caused serious water waste.Leakage control has become an important research topic in the water supply industry.In order to effectively control the leakage and locate the leakage fault in time,it is necessary to analyze the causes of leakage and the mechanism of leakage,and research and develop effective leakage positioning and control management techniques.In view of the problems existing in the current research process of the fault diagnosis and analysis technology for water supply network leakage,this paper has carried out the following researches:(1)Analysis of the cause of leakage of water supply pipe network;(2)Exploring the relationship between the change of leakage amount and the pressure change of pipe network when leakage of water supply pipe network;(3)Leakage state of water supply pipe network based on BP neural network Discriminate and locate the leakage position;(4)Comparison of BP algorithm based on discriminating state of water supply pipe network;(5)BP neural network leakage location based on monitoring point changes,making BP neural network water supply pipe network leakage fault intelligent diagnosis technology better applied in practice.At present,most of the domestic and international water supply pipe network is based on the SCADA system to locate the pipeline leakage,but the actual pipe network is buried underground for many years,and its internal structure and corrosion are difficult to grasp.Coupled with factors such as insensitivity to monitoring points,the effect of leaking point positioning is not satisfactory.Based on the experimental simulation platform of the water supply pipe network in the laboratory,based on the experimental platform pipe network information and real-time monitoring data,the EPANET is used for hydraulic modeling and checking.By using EPANET to simulate the leakage state of the pipe network,and then studying the relationship between the change of leakage amount and the pressure change of the pipe network,it provides a theoretical basis for the research of leakage network location technology of water supply pipe network.By studying the pressure changes in the leakage state,based on the simulation platform,the BP neural network leakage fault diagnosis model is established by the training sample,the number of hidden layers and the number of nodes,and the location of the leakage point in the pipe network and the water flow state at the time of leakage.The data input to the BP network is trained,fitted and verified.It is concluded that the location of the leakage point can be located by discriminating the leakage state through the BP network.In order to further improve the accuracy of BP neural network model for fault location,the three BP algorithms for pattern recognition are compared based on the first set of experiments,and the best BP algorithm for fault diagnosis of water supply network leakage is obtained.On the basis of the online monitoring data of the pipe network,in order to diagnose the location of the fault at the first time,through the analysis of the monitoring value change of the leakage moment,a more practical BP network diagnosis method based on the monitoring point change BP is proposed.Under the premise of repeatedly testing the internal structure of the established BP network model to obtain the best network topology,verify the rationality of the built BP network model,in order to achieve timely and accurate diagnosis of pipe network leakage,effective control of leakage water provides powerful technical support.