Deep learning identifies accurate burst locations in water distribution networks. (1st December 2019)

• Record Type: Journal Article
• Authors: Zhou, Xiao; Tang, Zhenheng; Xu, Weirong; Meng, Fanlin; Chu, Xiaowen; Xin, Kunlun; Fu, Guangtao
• Is Part Of: Water research. Volume 166(2019)
• Publication Date: 2019-12-01
• Subjects: Burst localisation -- Deep learning -- DenseNet -- Pipe burst -- Water distribution network; Water -- Pollution -- Research -- Periodicals; 363.7394
• DOI: 10.1016/j.watres.2019.115058
• Languages: English
• ISSNs: 0043-1354

Abstract

Abstract: Pipe bursts in water distribution networks lead to considerable water loss and pose risks of bacteria and pollutant contamination. Pipe burst localisation methods help water service providers repair the burst pipes and restore water supply timely and efficiently. Although methods have been reported on burst detection and localisation, there is a lack of studies on accurate localisation of a burst within a potential district by accessible meters. To address this, a novel Burst Location Identification Framework by Fully-linear DenseNet (BLIFF) is proposed. In this framework, additional pressure meters are placed at limited, optimised places for a short period (minutes to hours) to monitor system behaviour after the burst. The fully-linear DenseNet (FL-DenseNet) newly developed in this study modifies the state-of-the-art deep learning algorithm to effectively extract features in the limited pressure signals for accurate burst localisation. BLIFF was tested on a benchmark network with different parameter settings, which showed that accurate burst localisation results can be achieved even with high model uncertainties. The framework was also applied to a real-life network, in which 57 of the total 58 synthetic bursts in the potential burst district were correctly located when the top five most possible pipes are considered and among them, 37 were successfully located when considering only the top one. Only one failed because of the very small pipe diameter and remote location. Comparisons with DenseNet and the traditional fully linear neural network demonstrate that the framework can effectively narrow the potential burst district to one or several pipes with good robustness and applicability. Codes are available athttps://github.com/wizard1203/waternn . Graphical abstract: Image 1 Highlights: A novel deep learning framework is developed to locate bursts from pressure data. New algorithm enables key features in pressure data automatically extracted. Bursts are narrowed down from a potential district to one or several pipes. The framework is robust with model uncertainty and applicable in practice.