An Improved Method for Pipeline Leakage Localization With a Single Sensor Based on Modal Acoustic Emission and Empirical Mode Decomposition With Hilbert Transform

Abstract

Due to its great convenience for inspection of pipelines under special conditions, leakage localization with a single acoustic emission (AE) sensor has attracted increasing attention. However, the reported study achieves a good accuracy of leakage localization only for short source-to-sensor distance. In this work, an improved method for pipeline leakage localization with a single sensor is proposed based on modal acoustic emission and empirical mode decomposition with Hilbert transform (EMD-HT). The pipeline is considered as a cylindrical shell to compute the velocity dispersion curves of different guided wave modes. A criterion to select the specific wave modes for leakage localization is presented with the consideration of frequency matching, dispersive properties, and the modes’ confusing possibilities. Integrating the empirical mode decomposition (EMD) with wavelet transform (WT), the framework of the proposed method is developed. A number of field experiments were performed on a steel pipe with a continuous leakage source. The results show that using the proposed method, a maximum relative leakage localization error of 7.32% for the source-to-sensor distances from 0 to 33 meters (m) can be achieved, which is much improved than those of the reported method. With the capacity of precisely locating the leakage source within the range of 33m for each side of the sensor, the proposed method provides a promising way to make the AE technique more suitable for in-situ inspection of pipelines.