In order to study the principle and technologies of acoustic leak detection for natural gas pipelines

an acoustic method is applied. A simulation model is established based on the field conditions to analyze and verify the generation mechanism of leakage acoustics. The field experiments are carried out based on the simulation results and measured signals are processed by joint time-frequency analyses including short time Fourier transform

wavelet transform

Hilbert-Huang transform and Generalized S transform to extract the characteristics that propagate a long distance. The characteristics are applied to detect the leak of long distance pipeline with satisfactory results. Finally the characteristics are used to locate leakages. The results show that the characteristics extracted by signal-processing from leakage acoustic signals can be applied to leak detection for long-distance natural gas pipelines.