Gas pipeline leakage detection and localization method based on VMD-DTW

Abstract

The identification of signal anomalies at the moment of leakage is crucial for acoustic-based gas pipeline leakage detection and localization. To address this, a leakage detection and localization model based on variational mode decomposition and dynamic time warping (VMD-DTW) is proposed. First, the leakage signal is decomposed using the VMD method, and an adaptive mode selection and reconstruction is performed based on the low-frequency characteristics of the leakage signal for denoising. Next, a sliding window combined with DTW is used to design an algorithm that automatically identifies continuous time windows containing transient acoustic leakage signals. The algorithm improves leakage localization accuracy by estimating the time delay within multiple windows and automatically removing anomalous values. The principles and steps of the algorithm are presented, and its effectiveness is verified through experiments. The experimental results demonstrate that the proposed method offers better denoising performance and higher localization accuracy, with a localization error of only 0.247%.