Mechanism of Pipeline Leakage Sound Generation and Leak Detection Technology Under Multiple Operational Conditions

Area of Science:

• Engineering
• Acoustics
• Fluid Dynamics

Background:

• Acoustic emission detection is crucial for water supply system leak detection.
• Current methods struggle with diverse operational conditions, leading to inaccuracies.

Purpose of the Study:

• To enhance acoustic emission-based leak detection accuracy and generalization.
• To overcome limitations of field data in model development.

Main Methods:

• Large Eddy Simulation (LES) for analyzing leakage flow fields.
• Computational Aeroacoustics (CAA) for studying acoustic radiation.
• Feature extraction using wavelet packet energy ratio, centroid frequency, and frequency entropy.
• Support Vector Machine (SVM) model development.

Main Results:

• Established correlations between operating conditions and flow field characteristics.
• Clarified sound generation mechanisms of leakage acoustic signals.
• Achieved 98.6% accuracy in leak detection across varied conditions.

Conclusions:

• The developed model significantly improves high-accuracy leak detection with limited field data.
• Provides insights for developing cost-effective, low-computation leak detection systems.