Research on wavelet analysis for pipeline pre-warning system based on phase-sensitive optical time domain reflectometry

This paper proposes a strategy based on wavelet for phase-sensitive optical time domain reflectometry(Φ-OTDR) signal processing, which is used in pipeline early-warning system. This strategy can pick up vibration signal near the pipeline without frequency loss and sharpen the signal in time domain against shot noise, thermal noise and environment noise. This paper introduces the principle of Φ-OTDR and researches wavelet thresholding method and wavelet information entropy for it. Compared with the moving differential method, the wavelet method will avoid frequency loss. According to the noise analyses through wavelet, this paper finds an effective way to set thresholds for every point along the fiber adaptively. Through field experiment, the method is proved to be effective in reducing noise and sharpening signal with a SNR improvement from 4.23dB to 12.22dB. The wavelet information entropy can evaluate vibration in a comprehensive view and reach 40m alarm positioning accuracy along 23km distance.