Adaptive Acoustic Leak Location of Long Buried Pipe via Transient Optimization Statistical Analysis

Recently, nondestructive leak detection of buried pipe by acoustic approaches has attracted extensive attention of operating units. Although the practical cost of hydrophone is far lower than that of optical and magnetic flux devices, there are still some problems such as short detection distance and large positioning error. Aiming at hydrophone application, a robust leakage location strategy—statistical analysis of transient reconstruction (SATR) with adaptive parameter optimization is proposed. Unlike previous methods that only capture the first arrival transient component, this paper splits the reconstructed signal and estimates the leakage position by statistical characteristic distribution. Additionally, minimum envelope entropy is exploited as a constraint to optimize the parameters of maximum correlation kurtosis deconvolution, thus ensuring transient detection. Through the statistical probability distribution of the time delay of all deconvolution wavelets, the peaks higher than the threshold tolerance are weighted as the position estimation. Experiments at varying distances show that the relative positioning error (0.32%) of the proposed method is superior to other state-of-the-art approaches.