Burst Detection And Location In Water Distribution Systems Based On Hydraulic Model

Pipe burst is a sudden and hazard accident in water distribution system which is often occurred.The discovery of pipe burst accidents in major cities of our country is still in the way of artificial alarm.It is difficult to find a pipe burst in some special places because the pipeline is buried under the road,especially in the drainage ditch place.It is not only results in the waste of water resources,but also affects the normal operation of the city’s life and work.It is a common problem in the water industry to locate such emergencies of pipe burst quickly.Therefore,it is of great significance to find an economical and effective method to locate the pipe burst accident,which can ensure the safety of the water industry,the normal operation of the city and reduce the economic loss.Although major water supply companies have installed the SCADA systems to monitor the running state of pipe network,the systems are mostly not optimized and the arrangement is relatively sparse.The monitoring range is not clear,and they are can not locate the accident area timely even if appear the abnormal data.In this paper,the main contents are the location of pipe burst and the optimization of pressure monitoring points.And the research is developed from the following aspects:1.Pipe network hydraulic model and SCADA system are the necessary means to realize the modern management of water supply system.It is also a necessary tool to carry out the research of pipe burst location,optimized dispatching and water quality early warning events.The concept,composition and steps of the establishment of the two techniques were introduced,which lied the foundation for the following research.2.In this paper,the program was written to get the pressure change data of the pipe network based on the calculation of hydraulic adjustment of the network.And the fuzzy C mean clustering algorithm was used to partition the pipe network.3.In this paper,the hydraulic characteristics in the condition of pipe burst was carried out,and the two kinds of burst accidents of complete and non complete burst were simulated by the method of "virtual short pipe".Combined with the results of FCM clustering,the pipe burst early warning accident database was established.4.Based on the above research results,the auxiliary locating model of pipe burst was established using the the advantage of BP neural network in pattern recognition,and applied to the example pipe network and the actual large-scale urban pipe network.Its rationality was verified.5.Under the premise of existing pressure monitoring points and actual installation difficulties,this paper proposed a method for optimizing arrangement of pressure monitoring points.The results showed that the method can determine the position of pressure monitoring points more reasonably compared with the traditional layout method.