High-resolution pipeline leak localization using the side-lobes of the defect detection functional of matched-field processing

Matched-field processing (MFP) has recently been utilized to detect leaks in pipelines. The defect detection functional (DDF) of MFP features a primary lobe linked to the actual leak alongside secondary lobes. Such side-lobes have, until now, been considered as a nuisance and source of false identification of defects in pressurized pipelines. In this study, we turn this nuisance into a defect identification advantage. Specifically, we theoretically and experimentally demonstrate that the spatial positioning of the side-lobes of the DDF of MFP varies with the candidate leak location and the probing wave frequency, and that the paths or "trajectories" of the undesirable side-lobes converge to the accurate leak position as the frequency increases. By leveraging this property of the side-lobes, this paper introduces novel broadband MFP techniques for high-resolution leak detection in the presence of white, colored, and impulse-type noise. Both numerical and experimental data are used to demonstrate the accuracy, stability, and resolution of the proposed defect detection techniques and their ability to suppress the influence of undesirable side-lobes.