Answers are grounded in this paper and cite their sources.
Pipeline leak detection must be both accurate and quick to ensure the operational integrity, safety, and sustainability of fluid transport systems. This work introduces an innovative integrated methodology that merges advanced acoustic signal analysis with geostatistical spatial modeling for efficient leak identification and localization. Under controlled leakage conditions, we collected acoustic emission signals and used the Maximal Overlap Discrete Wavelet Transform (MODWT) to remove noise and make the signals more distinct. The Time of Arrival (TOA) was precisely estimated by the Hilbert envelope of the denoised signals, a critical characteristic associated with leak-induced acoustic propagation. We then used both one-dimensional and two-dimensional (space-time) ordinary kriging to spatially interpolate TOA. Variogram modeling with exponential, spherical, Gaussian, and Matérn models was tested, and the Gaussian and spherical models performed the best in terms of RMSE. The kriging results assessed uncertainty with 95% confidence intervals and produced interpretable spatial maps of the leak's propagation. This hybrid acoustic-geostatistical technique provides a scalable and data-driven leak detection system, marking a significant step forward in the integration of signal processing and spatial modeling for pipeline integrity monitoring.
No takes yet. Share an insight, caveat, or question.
Ali et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: