Estimation Of Leakage Quantity Of Urban Water Supply Pipeline Network

The leakage of pipeline network is a major challenge for urban water supply network system.Due to the increasing complexity of water supply network,the tradi-tional leakage detection is becoming more and more difficult to work.The industry expects to establish a pipeline network leakage model on the basis of IoT sensing,and to determine the leakage point and leakage volume through the system monitoring method.Based on the existing research,this paper studies the leakage problem of pipeline network as follows:1.Estimation of leakage volume of water supply network based on FastICA.The water consumption signal in the regional flow is separated by FastICA algorithm,the separation effect is evaluated by the correlation coefficient of Pearson,and the balance equation is established,and the component of the blind source separation is restored by genetic algorithm.Combined with the example network,the leakage of water supply network is simulated and the leakage is estimated under the condition of single leakage point of single water source,multiple leakage point of single water source,single leakage point of multi-water source and single leakage point under background leakage.Through experimental analysis,the leakage separation model based on FastICA can accurately separate the total leakage in the water supply area,the relative error is between 1%and 7%,and the separation effect is better.Experime-nts show that the leakage degree,the distribution of leakage points and the amount of water source have little effect on the separation of leakage in FastICA.2.The initial location of water supply network leakage based on spectral clustering and pressure anomaly analysis.The sensitivity analysis of the pressure monitoring point and the joint in the pipe network is carried out,the sensitivity coefficient matrix is constructed,the correlation area of the pressure monitoring point is divided by spectral clustering,and the pressure anomaly change is observed by comparing the measured pressure of the pressure monitoring point with the average pressure of the previous week to determine the leakage area.It is proved by an example that this method can reduce the detection range of network and determine the leakage area of network quickly.3.Accurate location and leakage estimation of leakage based on pressure dependent leakage detection model.The PDLD model is used to check the running data of the time network with representative time,such as the minimum flow time during the night,the peak time of daytime water and the daytime trough period,and the artificial bee colony algorithm is used to optimize the leakage coefficient of the leakage point and determine the location of the leakage point,and then estimate the leakage amount.The experiment of Tube network shows that the multi-variable solution speed of artificial bee colony algorithm is faster than that of genetic algorithm and particle swarm algorithm,and the method of optimizing checking PDLD model based on artificial bee colony algorithm is feasible in leakage location and leakage estimation,and the leakage location of non-water node in pipe network is more accurate than that of water node.In conclusion,this paper studies the leakage in the area of pipeline network,the initial location of leakage area,and the precise location of leakage&the estimation of leakage quantity,in order to provide guidance for the leakage detection of the actual water supply network.