LAUSR

Smart water networks, created using Internet of Things (IoT) technologies, have been increasingly adopted by water utilities across the world. This research focuses on the use of smart water technologies for detecting newly formed through-wall pipe cracks and leaks in water distribution systems for the purpose of pipe break early warning and prevention. This research develops easy-to-implement algorithms for analysing the hydro-acoustic vibration wave files regularly collected by permanently installed accelerometers in a water network, and for generating alarms when small leaks from developing pipe cracks are detected. Descriptive features of the historical wave files are investigated for five sites where newly-formed pipe cracks were detected nearby. It is found that the median frequency (MF) and the root-mean-square (RMS) values derived from the wave files are useful indicators for detecting new cracks and leaks. The confidence of pipe crack detection in noisy city environments can be significantly increased by windowing each already short-duration wave file into a number of frames with even shorter durations, and analysing the behaviour of the MF and RMS values of all the frames to look for the persistence of the leak feature even when part of the signal is contaminated with interference. The results show that the developed technique can robustly detect new through-wall cracks and leaks in a timely manner, and is tolerant of interference from customer water use and transient environmental noise.