Localization of Multiple Leak Sources Using Acoustic Emission Sensors Based on MUSIC Algorithm and Wavelet Packet Analysis

Abstract

Multiple leak sources may occur in a large pressure vessel that contains corrosive materials or has been in use for a long period of time. Although, a variety of leak localization methods have been proposed in previous studies, they are capable of locating only a single leak source. Methods for simultaneous localization of multiple leak sources are desirable in practical applications. To address this issue, a novel method using acoustic emission (AE) sensors in conjunction with MUltiple SIgnal Classification (MUSIC) algorithm and wavelet packet analysis is proposed and experimentally assessed. High-frequency AE sensors are assembled into a linear array to acquire signals from multiple leak sources. Characteristics of the leak signals are analyzed in the frequency domain. Wavelet packet analysis is deployed to extract useful information about the signals from the frequency band of 50–400 kHz. The MUSIC algorithm is applied to identify the directions of the leak sources through a space spectrum function. Leak sources are located based on the directions identified by the AE sensor array placed at different locations. The performance of the proposed method is evaluated through experimental tests on a stainless steel flat plate of 100 cm $\times 100$ cm $\times 0.4$ cm. The results demonstrate that the method is capable of locating two leak holes. In addition, the localization accuracy depends on the leaking pressure. It is demonstrated that the two leak holes are located within two small areas, respectively, which are 25.12 cm2 for leak hole 1 and 1.96 cm2 for leak hole 2.