Leakage Monitoring And Early Warning Technology For The Small/Medium DMA Water Supply Network

Leakage monitoring of urban water supply network has important practical significance for saving water resources and ensuring normal use of water by urban residents.There are many disturbance factors in the actual water supply production,which cause large fluctuations in data from real-time monitoring hydraulic parameters,and it is difficult to judge whether there is an abnormality in the pipeline network.Focusing on this practical problem,the following research work is carried out in this paper on the basis of existing researches:(1)A small DMA short-term water demand prediction method based on water consumption pattern is proposed,aim to solve the problem of inaccurate in water demand prediction caused by climate and social factors.The methods of isolation forest,correlation analysis and factor analysis are used to analyze the main factors affecting daily water consumption,and utilizes the extracted main influencing factor information and historical monitoring data to establish the Elman water demand prediction model.Furthermore,extract the daily water consumption pattern through correlation analysis,matching the pattern of predicted day,and combined with the daily water demand prediction value to allocate water demand at all times.The method is improved the short-term water demand prediction accuracy.(2)A small DMA leakage monitoring method combining DTW timing series distance and conventional deviation is proposed,aim to solve the problem of misreporting caused by the deviation of normal water time and water consumption fluctuation.First of all,DTW was applied to calculate the time series distance between the predicted water demand and the actual water consumption,which reduced the influence of the water time offset on abnormal judgment.Then,conventional deviation was applied to judge each time point in the timing series group with abnormal DTW distance to reduce false positives.The comparative experiment proves the method has better effectiveness and real-time performance.(3)A medium DMA pipe burst monitoring and location method considering water disturbance is proposed,aim to solve the problem of many false positives caused by local water disturbance.First,by extracting the ideal fluctuation curve of the pressure signal without noise,the disturbance that causes the abnormal pressure of the measuring point is analyzed.Secondly,EPANET is applied to simulate the operating state of the pipeline network,and the coloured noise generated by the disturbance is quantified by adding the emitter coefficient and adjusting the hydraulic head,to improve the simulation results and reduce the leakage false positives.Finally,the centre-of-gravity method is used to calculate the pressure differential and coordinates of the strong abnormal measuring point to realize the pipeline burst location.The experiment demonstrates that the method has a lower false positive rate and better positioning capabilities than the monitoring method that does not consider water disturbance.