Pipeline leak fault feature extraction based on wavelet packet analysis and application

Wavelet packet analysis provides a more subtle approach for signal analysis, in the signal processing, through multiple levels division of the signal band, which made the high frequency part a deeper level decomposition. There is also according to the nature of the signal itself choose the corresponding band adaptively, so as to consistent with the signal spectrum, and thus make the time frequency resolution has been greatly improved. Thus we could extract fault feature in a more fine-grained band, so this method is suitable for non stationary signal analysis. This article describes the pipeline leak fault feature extraction based on wavelet packet analysis and application.