Application of Wavelet Packet and Support Vector Machine to Leak Detection in Pipeline

Author

Na, Liu ; Yanyan, Zhao

Author_Institution

Beijing Inst. of Petrochem. Technol., Beijing

Volume

1

fYear

2008

fDate

3-4 Aug. 2008

Firstpage

66

Lastpage

69

Abstract

Acoustic emission (AE) technology is one of the promising methods for pipeline leak detection. But AE signal of pipeline leak carry the non-stationary feature information. Therefore, one major problem is feature extraction of AE signal from noise and complexity in transmission. This paper analyzes the characteristics of AE signal to acquire effective waveform signal, applies wavelet package to construct and reconstruct signal, extracts feature vectors to set up training sample, and establishes a support vector machines (SVM) fault classifier to diagnosis. Experimental results show that the identification rate can be up to 100% for pipeline leak.

Keywords

acoustic emission; leak detection; petrochemicals; pipelines; production engineering computing; support vector machines; wavelet transforms; acoustic emission technology; nonstationary feature information; pipeline leak carry; pipeline leak detection; support vector machine; waveform signal; wavelet packet; Acoustic emission; Acoustic noise; Feature extraction; Leak detection; Pipelines; Signal analysis; Signal detection; Support vector machine classification; Support vector machines; Wavelet packets; Acoustic Emission; Feature extraction; Pipeline Leak; Support Vector Machine; Wavelet packet;

fLanguage

English

Publisher

ieee

Conference_Titel

Computing, Communication, Control, and Management, 2008. CCCM '08. ISECS International Colloquium on

Conference_Location

Guangzhou

Print_ISBN

978-0-7695-3290-5

Type

conf

DOI

10.1109/CCCM.2008.346

Filename

4609470