The Research on Leak Detection Technology of Natural Gas Pipeline Based on EMD

The acoustic emission signal of pipeline leakage is characterized by nonlinear and non-stationary. It is not feasible to extract the leakage feature signal in traditional signal processing methods. The leak locations can be detected by employing the improved empirical mode decomposition (EMD) to decompose the acoustic emission signal into several intrinsic mode functions (IMF), choosing IMFs containing leakage characteristics to be reconstructed, and doing correlation analysis. Experimental results show that the positioning accuracy of leakage detection is improved obviously.