Leakage Sensor Placement Optimization Using Acoustic Attenuation Features in Water Mains

Leakage detection using integrated wireless vibration sensors has been minimizing labor in water pipe networks. Specifically, vibration sensors with a built-in low-noise micro-electro-mechanical systems (MEMS) accelerometer can detect leakage-induced acoustic waves at long ranges. Automatic sensor placement optimization provides sufficient coverage and minimizes the required number of sensors. This can save weeks of manual labor selecting the installation positions of 1,000 or more sensors in a city-scale pipe network. However, the detection distance of a vibration sensor varies based on the piping conditions around the sensor installation position, often causing mismatching between estimated and actual coverage. Therefore, we propose a novel sensor placement optimization method that utilizes the acoustic attenuation features of water piping structures. We successfully utilize it with a city-scale pipe network and optimize placement of 2,997 sensors in one day. We achieve coverage of over 95% regardless of the piping conditions, indicating our method would be sufficient for practical use.