Continuous Leak Location in Gas-Filled Steel Pipe Based on Optimum Wavelet Analysis of Leak AE Signal

Abstract

Abstract This study is investigates how to reconstruct the continuous leak acoustic emission (AE) signal to precisely locate the leak source on a gas-filled steel pipe based on cross-correlation location theory. Thus, in this article, the determination method for the optimum mother wavelet was proposed. The frequency distribution of the continuous leak AE event was firstly extracted through wavelet transform (WT) with optimum mother wavelet. In the considered frequency range, the WT coefficients were further used to reconstruct the raw continuous leak AE signal. The correlation-based location experiments were finally carried out to evaluate the performance of the proposed method. The results show that the location errors of the leak source were decreased by some 10 % compared with those achieved by only using the raw AE signals. Thus, improvement verifies the validity of choice of the optimum mother wavelet. Meanwhile, it is also founded that, whether the reconstructed signals or raw signals were used, the localization accuracy of the leak source using the speed of the T wave is greater than that using the speed of the L wave and theoretical acoustic speed under the same symmetry of the AE sensors with reference to the leak source. Therefore, defining the acoustic speed with the speed of the T wave was recommended to use on-site for detection of the continuous leak source on a gas-filled steel pipe.