Chebyshev Descriptors for SHM with Acoustic Emission and Acousto Ultrasonics. Issue 3 (28th August 2014)

• Record Type: Journal Article
• Authors: Crivelli, Davide; Eaton, Mark; Pearson, Matthew; Holford, Karen; Pullin, Rhys
• Editors: Rodopoulos, ChrisAlexander; de Jesus, Abi­lio
• Is Part Of: International journal of structural integrity. Volume 5:Issue 3(2014)
• Publication Date: 2014-08-28
• Subjects: Structural engineering -- Research -- Periodicals; 620.11205
• DOI: 10.1108/IJSI-04-2014-0013
• Languages: English
• ISSNs: 1757-9864

Abstract

Abstract : Purpose: The paper is a feasibility study on the use of alternative parameters for representing Acoustic Emission and Acousto Ultrasonic signals, using a wavelet based approach and the computation of Chebyshev moments. Design/methodology/approach: Two tests were performed, one on Acoustic Emission artificial signals generated on a CFRP plate and one on an Acousto Ultrasonic setup used for actively detecting impact damage. The waveforms were represented using a data reduction technique based on the Daubechies wavelet and an image processing technique using Chebyshev moments approximation, to get 32 descriptors for each waveform. Findings: The use of such descriptors allowed in the Acoustic Emission case to verify that the moments are similar when the waveforms are similar; in the Acousto Ultrasonic setup the correlation coefficient of the descriptors with respect to a reference dataset was found to be linked to the delimitation size. Practical implications: Such a data reduction while retaining all the useful information will be positive for wireless sensor networks, where power consumption during data transmission is key. With having to send only a reliable set of descriptors and not an entire waveform, the power consumption is believed to be reduced. Originality/value: This paper is a preliminary study that fulfills a need for a more reliable data reduction for ultrasonic transient signals, such as those used in Acoustic Emission and Acousto Ultrasonics.