Acoustic-based approach for micro-leakage detection and localization in water supply pipelines

Abstract

Acoustic detection is a widely used method for investigating leaks in water supply pipelines. However, improving the accuracy of acoustic detection techniques is crucial, especially in scenarios with low pipeline pressures (≤0.2 MPa) and small leak apertures (≤2 mm), where micro-leakage detection poses challenges. In this study, a pipeline model is constructed based on acoustic leak detection principles, and numerical simulations are performed using FLUENT software. The occurrence and propagation of sound are simulated using the Foutz-Williams–Hawkins (FW–H) equation, to generate sound signals by micro-leakage in pipe sections with adjacent tee pipe fittings. The results indicate that the average sound pressure amplitude caused by pipeline vibration varies with changes in pressure. In particular, upstream locations exhibit a higher degree of variability compared to downstream locations. An increase in both pipeline pressure and leak aperture leads to an amplified power spectrum across different frequency bands at various detection points. However, the energy generated by water leaks and vibrations in tee pipe fittings is relatively low and heavily distorted by ambient signals. To mitigate these challenges, a combination of the empirical mode decomposition (EMD) method is utilized to extract leak sound signals and eliminate interference information. Additionally, the cross-correlation time delay estimation method is used to determine the time difference between upstream and downstream sensors when receiving leak sound signals. This approach successfully identifies and localizes leakage points in pipe segments with tee pipe fittings. This study provides evidence of the effectiveness of this approach in detecting and localizing micro-leakage points in water supply pipelines, achieving a remarkable localization result with a relative error of only 1%.