A distributed water pipeline leakage monitoring system based on smart audio

Abstract A water supply network relies on multiple intertwined and complex pipelines, which are often buried underground and will inevitably leak during long-term usage, resulting in serious water waste. In this paper, a distributed pipeline leakage detection system is proposed, which comprises two identical sets of low-power terminals and a central host computer responsible for executing the detection algorithms. These terminal units can be placed on the ground above the pipeline, enabling leak detection and accurate positioning without the need for excavation. The host computer performs signal preprocessing using algorithms such as Short-Time Fourier Transform (STFT), Wavelet Transform, and Empirical Mode Decomposition (EMD) to enhance signal quality. The location of the leakage point is then determined by the cross-correlation method. Additionally, to enhance the reliability of detection, reducing the influence from neighboring peaks, an energy centroid method is proposed and tested. Results show that algorithms are effective in detecting leakage points, and the location is relatively accurate.