Use of Machine Learning for Leak Detection and Localization in Water Distribution Systems

This paper presents an investigation of the capacity of machine learning methods (ML) to localize leakage in water distribution systems (WDS). This issue is critical because water leakage causes economic losses, damages to the surrounding infrastructures, and soil contamination. Progress in real-time monitoring of WDS and ML has created new opportunities to develop data-based methods for water leak localization. However, the managers of WDS need recommendations for the selection of the appropriate ML methods as well their practical use for leakage localization. This paper contributes to this issue through an investigation of the capacity of ML methods to localize leakage in WDS. The campus of Lille University was used as support for this research. The paper is presented as follows: First, flow and pressure data were determined using EPANET software; then, the generated data were used to investigate the capacity of six ML methods to localize water leakage. Finally, the results of the investigations were used for leakage localization from offline water flow data. The results showed excellent performance for leakage localization by the artificial neural network, logistic regression, and random forest, but there were low performances for the unsupervised methods because of overlapping clusters.

Hybrid TDR-MI Based Wireless Sensor Network for Underground Water Pipeline Leakage Detection and Localization Using Pressure Residuals and Classifiers.

The pipeline leakage detection and leak localization trouble is a highly demanding and dangerous issue. Underground pipelines are a critical mode of transporting enormous fluid volumes (e.g., water) across extended distances. Solving this problem will save the country much money and resources, but it will also protect the environment. On the other hand, present leak detection technologies are insufficient for monitoring underground pipelines due to the extreme subterranean environmental conditions. This study proposes a hybrid wireless sensor network based on TDR (time domain reflectometry) and magnetic induction for monitoring underground pipelines to solve these problems. In this scenario, TDR is deployed beneath an MI-based wireless sensor network. TDR precisely locates the leak and dramatically decreases the amount of time required for inspection. We offer a wireless sensor network based on MI technology for low-cost, real-time leak detection in subsurface pipes. MISE-PIPE identifies leaks by integrating data from a range of different types of sensors installed within and around underground pipelines. Ad-hoc WSNs are used to measure pressure. (WDNs) is a hot topic that has piqued researchers' interest in recent years. Time and accuracy are critical components of leak localization, as they substantially impact the human population and economy. Statistical classifiers acting in the residual space are offered as a general method for leak localization. Classifiers are trained on leak data from all network nodes, taking demand uncertainty, sensor preservative noise, and leak magnitude on the account. Following leak identification and localization, all monitoring data is forwarded to the CH using the K-means clustering method, which serves two critical functions: optimal clustering and prolonging the Network Lifetime (NL) and preserving the QoS. The clustering method is optimized using the K-Means approach .

Accuracy and Uncertainty of Gradient Based Leak Localization Procedure for Liquid Transmission Pipelines

This paper describes issues of leakage localization in liquid transmission pipelines. It focuses on the standard leak localization procedure, which is based on the calculation of pressure gradients using pressure measurements captured along a pipeline. The procedure was verified in terms of an accuracy and uncertainty assessment of the resultant coordinate of a leak spot. An important aim of the verification was to assess the effectiveness of the procedure in the case of localization of low intensity leakages with a level of 0.25–2.00% of the nominal flow rate. An uncertainty assessment was carried out according to the GUM convention. The assessment was based on the metrological characteristics of measuring devices and measurement data obtained from the laboratory model of the pipeline.