Performance analysis of several wavelets in pipeline leak detection and location using Colpitts chaotic signal

In this paper, the performance of wavelet transform on the detection and location of pipeline leak are explored. The effects of three wavelets on different thresholds and different decomposition layers are studied. Firstly, the UWB(Ultra Wideband) chaotic signal generated by Colpittscircuit is used as the signal detection signal, and the leak point of the pipeline is detected and located by analyzing the cross-correlation coefficient between the detection signal and the echo signal after the pipeline reflection. Because there are some questions such as how to choose the optimal threshold function, wavelets and optimal number of decomposed layers, three wavelet thresholds (hard threshold, soft threshold and given threshold)which has different signal-to-noise ratio (5dB, 10dB, 30dB) are used in signal threshold denoising. The paper explores the denoising effects of three wavelets(sym4, db4, Haar) after three, four and five wavelet decomposition, and finally gets the optimal threshold method, wavelet function and decomposition layer. This is important to achieve high-precision detection and location of water pipeline leak.