Leakage detection algorithm integrating water distribution networks hydraulic model

Abstract

Water loss through leaking pipes is inexorable in water distribution networks (WDNs) and has been recognized as a major challenge facing the operation of municipal water services. This is strongly linked with financial costs due to economic loss, environmental issues, and water resource savings. Water distribution systems are complex in nature with a large number of pipes. Therefore, monitoring long-range pipelines for leaks is a challenging task. This problem is aggravated when there are simultaneous leakages at various points in more than one pipe in the networks. Consequently, there is an urgent industrial need for a reliable approach to provide the capabilities for loss reduction through the detection of leaking pipes in WDNs. Nevertheless, a large number of techniques for detecting leakages have been proposed in the literature. These methods include the use of acoustic correlation techniques, measurements and statistical analysis of the abrupt changes in pressure at the leak points, and the transient based approach to mention but a few. However, their performance in detecting background leakage is quite low. In WDNs, background leakage is often hidden, small, and continuous, posing the biggest threat to water utilities. Therefore, its detection and estimation is vital for effective water service. For effective detection of background leakages, a hydraulic analysis of flow characteristics in water piping networks is indispensable for appraising such type of leakage. A leakage detection algorithm incorporating the hydraulic model of flow in water piping networks should prove worthwhile in detecting and estimating background type leakages. In this work, a leakage detection algorithm integrating the hydraulic model of the flow in water piping networks is proposed. The hydraulic model takes into account some important parameters of pipes such as pipe deterioration due to ageing, which was not considered in previous leakage detection studies. The deterioration in pipe materials and changes in its diameters over time can lead to unsatisfactory results. Furthermore, in large-scale networks, most leakage detection algorithms only cover a specific area such as the district meter areas (DMAs) of the network. The authors are optimistic that the proposed algorithm for solving network leakage outflow will help improve the leakage detection accuracy while providing the capability to cover a wide area in the large-scale water piping networks.