Efficient Pipe Burst Detection in Tree‐Shape Water Distribution Networks Using Forward‐Backward Transient Analysis

• Published in: Water resources research 2022-11, Vol.58 (11), p.n/a
• Main Authors: Pan, Bin, Duan, Huan‐Feng, Keramat, Alireza, Meniconi, Silvia, Brunone, Bruno
• Format: Article
• Language: English
• Subjects: Analysis; Defects; Detection; Distribution; Energy conservation; Leak detection; Localization; Methods; Networks; Optimization; pipe burst detection; Pipes (defects); Robustness (mathematics); Signal processing; Transient analysis; Trees; tree‐shape network; Water distribution; Water engineering; water leakage; Water management; Water shortages; Water supply; water supply pipe
• ISSN: 0043-1397; 1944-7973
• DOI: 10.1029/2022WR033465

Transient‐based leak/burst detection has received more and more attentions in the operation and management of water distribution networks. However, with the increase in pipe network complexity, it is hard to detect and locate the potential pipe defects and water losses efficiently. Recently, we proposed an efficient leak/burst localization strategy in single pipelines based on the forward and backward transient analysis, and in this work, this effective approach is customized for the more realistic branched (tree‐shaped) pipe networks. Unlike traditional transient‐based methods which rely on tedious optimization procedures, the proposed method uses a mismatch property between forward and backward analyses of transient signals for the pipe burst detection during the water supply process. Compared with other developed methods, this method can directly calculate the defect location, so it is expected to have high efficiency and wide applicability. To fulfill this goal, a framework for burst detection in a tree‐shaped pipe network is developed in this paper. Using the mass conservation and energy relations at the branched junctions, the forward and backward transient analysis is extended from one pipe to another until the burst is found and then located in the network. Both experimental and numerical tests show the effectiveness of the proposed method, and further analysis has shown that the method is also valid for leak detection in both transient and steady conditions, confirming the robustness and practicality of this proposed method. At last, this paper discusses the influence of transient sources and the potential combination of the method with other methods. Key Points Burst detection in tree‐shape pipe networks using forward and backward transient analysis An effective and efficient framework for burst identification, and localization based on pipe end pressure transients Successful burst detection in pipe networks under steady and unsteady flow states