Feature extraction of acoustic emission signals based on wavelet decomposition

Abstract

In this paper, the theory of wavelet decomposition with its Mallat algorithm is introduced and the method of features extraction based on the wavelet decomposition for the acoustic emission (AE) signals is presented. Additionally, a database of wideband acoustic emission signals is used to examine the method. The characteristics of two typical AE signals sources both crack and lead snap are abstracted. Experimenting by several typical acoustic emission sources signals, we validate its feasibility and validity.