Davide Crivelli, Mark Eaton, Matthew Pearson, Karen Holford and Rhys Pullin
The purpose of this paper is to study the feasibility on the use of alternative parameters for representing acoustic emission (AE) and acousto-ultrasonic (AU) signals, using a…
Abstract
Purpose
The purpose of this paper is to study the feasibility on the use of alternative parameters for representing acoustic emission (AE) and acousto-ultrasonic (AU) signals, using a wavelet-based approach and the computation of Chebyshev moments.
Design/methodology/approach
Two tests were performed, one on AE artificial signals generated on a CFRP plate and one on an AU 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 AE case to verify that the moments are similar when the waveforms are similar; in the AU setup the correlation coefficient of the descriptors with respect to a reference data set 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 fulfils a need for a more reliable data reduction for ultrasonic transient signals, such as those used in AE and AU.