Research On Leakage Detection And Location In District Metering Areas Based On Water Balance

As one of the most precious natural resources,freshwater has drawn great attention worldwide.However,a large amount of water is lost during water distribution process,which can be largely attributed to bursts and leaks.It is of great significance to explore an effective method for leak detection.However,only based on the abnormal flow or pressure caused by leak,the traditional leak detection methods suffer a high false alarm rate,and have difficulty in detecting small flow leaks.Based on water balance,this thesis proposed a leak detection and location method in district metering area(DMA).The main work and innovations of this thesis are summarized as follows:(1)To improve the precision of the pressure-based burst detection method,an information fusion method is proposed to fuse the abnormal possibilities of pressure data and flow data.The paper introduces unmetered water consumption,predicts the abnormal probability based on an autoregressive model.And the logistic regression model is introduced to fuse the abnormal possibilities of pressure data and flow data.Based on the two-dimensional information of pressure and flow,the precision of burst detection is improved and the false alarms are reduced.Experimental results show that the proposed method can effectively reduce false alarms and improve the performance of burst detection.(2)To improve the performance of leakage detection in DMA,an integrated bottom-up approach for leakage assessment is proposed in this thesis.The amount of water loss caused by leaks is used to detect non-sudden leaks such as small flow leaks.Firstly,the water balance model is constructed based on the discriminative behavior of apparent and real losses.Then,based on the hydraulic information(e.g.pressure,inlet flow,water usage)gathered by SCADA system,the Adam algorithm is employed to perform least square estimation of the model parameters.Then the support vector machine is introduced to detect leaks based on the parameters of water balance model.By comparing the principal component analysis method and the feature extracted by the water balance model,it is showed that the feature extraction method based on the water balance model can effectively extract the features of the leakage event,thus improve the performance of leak detection in district metering area.(3)The hydraulic model-based leakage zone identification method relies heavily on the accuracy of the pressure sensor and the accuracy of the hydraulic model.Thus,it is difficult to apply in complex pipe networks.To solve this problem,a leakage zone identification method is proposed based on the pressure distribution of the pipe network.First,clustering algorithms are used to aggregate water nodes with similar water supply pressures to divide the water supply network into multiple leak identification zones.Then,the gradient boost decision tree model is used to predict the average zone pressure of each zone.Finally,based on water balance model,the average zone pressure of each zone is used to fit the flow data,and the best fitting zone is identified as leakage zone.The experiments result shows that under the appropriate number of leak identification zones K,this method can effectively identify the leak zone.(4)Investigated the needs of the online monitoring platform for water supply network leakage and the selection of major technologies,the SSM framework in Java Web technology and the MySQL database are used to establish an online monitoring platform for water supply network leakage.The database module,algorithm module,Background management module,interactive interface and other modules are developed.Finally,the monitoring data of a certain DMA in eastern China is used as an application case to verify the effectiveness of the leak detection method studied in this paper.