A pipeline leak localization method based on SVMD and ASPE denoising
作者
Zhenhuan Song,Xiaofeng Zhou,Liang Shi,Chong Luan,Changwu Xin,Zhongji Yu,Chao Yuan,Jiapeng Duan,Tenglin Li
出处
期刊: [SPIE] 日期:2026-07-02卷期号:: 106-106
标识
DOI:10.1117/12.3113294
摘要
The leakage signal of the water pipeline is affected by noise interference, resulting in low accuracy in leak location. A novel denoising technique for pipeline leak detection is introduced, combining Successive Variational Mode Decomposition (SVMD) and Amplitude-Sensitive Permutation Entropy (ASPE). Initially, the leakage signal undergoes decomposition through SVMD, yielding a set of Intrinsic Mode Functions (IMFs). Then, the extracted IMFs are classified into two types: those primarily influenced by noise and those dominated by the leak signal, using ASPE. The leak signal-dominated IMFs are then reconstructed to obtain the denoised signal. This denoised signal is used to calculate an accurate time difference, which in turn helps determine the precise leak location. The proposed technique significantly enhances the precision of leak detection, as validated by experimental data, with an average positioning error of only 4.33%.