Source identification and classification of acoustic emission signals by a SHAZAM inspired pattern recognition algorithm

Abstract

This study focuses on the development of a source identification algorithm inspired by the SHAZAM music app. The algorithm makes use of a spectrogram analysis technique for distinguishing different Acoustic Emission (AE) events. The peaks of the spectrogram are used to obtain a constellation map generating a "fingerprint" like pattern for each acoustic emission source. The fingerprints are then used within an artificial intelligence algorithm as part of a Knowledge Discovery database. The database is then able to link the AE signal to a specific source type. An experimental program was developed to test the methodology. The results of this study demonstrate that signal sources can be classified and linked to specific emission types with a high level of accuracy.