In-pipe leak detection by Minimum Residual complexity: A Robust time Delay Estimation method against correlated noise

TL;DR

This paper introduces a robust time delay estimation method using residual complexity for in-pipe leak detection, demonstrating superior robustness against correlated noise compared to traditional measures.

Contribution

The paper presents the application of residual complexity as a new similarity measure for time delay estimation in leak detection, outperforming cross correlation and mutual information in noisy conditions.

Findings

RC shows greater robustness against correlated noise.

RC reduces the number of samples needed for accurate TDE.

Experimental results confirm advantages over CC and MI.

Abstract

In this work, the residual complexity (RC) similarity measure, is employed for time delay estimation (TDE) in gas pipe leak localization. The properties of TDE by RC is compared with those of Cross Correlation(CC) and Mutual Information(MI) similarity measures on our experimental data. The comparison confirms the advantages of RC relative to CC and MI, in robustness against both correlated noises and reduction of number of samples. These advantages originate from not only its mathematical nature of RC similarity measure which considers inter-dependency but also from broadband frequency of acoustic waves propagating during gas pipes.