Anomaly Detection in Targeted Pipe Sections in Water Pipe Systems Using Hydroacoustic Signal Deconvolution

Abstract

Detection of anomalies, such as leaks, blockages and deteriorated sections, in underground water pipe networks and long transmission mains is a challenging problem. This paper proposes a novel method for anomaly detection in targeted pipe sections embedded within any complex pipe system. In the proposed approach, persistent hydroacoustic waves generated simply by opening a side-discharge valve are sent into the pipe system through existing access points, such as hydrants and air valves. Pressure measurements are required only at existing access points that bracket the pipe section of interest. A signal deconvolution process was developed to transfer the complex waveforms of the measured hydroacoustic waves into a deconvolution trace, which consists of impulse response functions (IRFs) of the pipe. Mathematical models that link the spikes in the deconvolution trace to the anomalies existing in the pipe system were derived to identify and localize these anomalies. Numerical validation was undertaken on three different pipe configurations: two single-pipe systems, and a pipe network. Experimental validation was conducted on a laboratory copper pipe network connected to the municipal water distribution system in which a simulated leak was localized accurately. The results demonstrate that the proposed technique (1) is easy to implement (it uses only valves and single pressure transducers connected to existing access points), (2) is able to detect and locate anomalies accurately for targeted pipe sections in complex pipe systems, and (3) is tolerant of background pressure fluctuations and noise that naturally occur in real water distribution systems.