Research On Water Supply Network Leakage Detection Syste

The leakage of water supply network is one of the main hidden dangers in domestic water supply.The leakage monitoring of water network can not only notify relevant maintenance personnel in time when the water pipe leaks or even bursts,accelerate the progress of pipeline maintenance,and improve the efficiency of water supply,but also reduce the water waste caused by troubleshooting.This paper focuses on the leakage detection of water supply network and the construction of smart network,and analyzes the overall design of smart network.First,the framework of the intelligent pipe network system is constructed,and the composition and design ideas of the intelligent pipe network system are described in detail.At the same time,starting from the actual water supply pipe network of the subject,the hardware foundation is understood,and the structure is familiar so as to carry out the system architecture.In particular,the actual water supply pipe network is modeled through the GIS system to facilitate the construction of the intelligent pipe network system.Then,based on spectral clustering algorithm,it provides theoretical guidance for DMA zoning of water supply network,and constructs hydraulic model of water supply network.With the actual pipe network data of the subject as the data support,we build the water supply pipe network model through calibration,divide the pipe network into DMA.Based on the model,the main body of intelligent pipe network system is built to realize computer control of water supply pipe network and receive relevant data.Finally,considering the influence of pipe diameter parameters and pipe materials on the leakage of water supply network,the leakage detection algorithm of water supply network is improved.Based on the idea of Transformer model,the attention mechanism is introduced into the neural network algorithm,and different dimension weights of each data parameter are assigned according to the attention mechanism.Using LSTM(Long Short Term Memory)neural network to mine the water pressure data of each node can not only improve the accuracy of missing calculation detection,but also improve the operation efficiency and eliminate garbage data.Then this model is applied to the intelligent pipe network system.Through the training of the data of the intelligent pipe network system,it is proved that the accuracy of this model in practical application reaches 90%,which is of use value and can realize the leakage detection of the water supply pipe network timely and accurately.