The AE signal characteristic parameters extraction and optimization of initial crack during deep drawing of metal

Abstract

The paper extracts acoustic emission signals between the normal condition and the crack state through the micro crack experiment of drawing parts. The local wave decomposition of acoustic emission signals was achieved to get the nine intrinsic mode functions and a tendency by the MATLAB software. The energy value of the former 8th-order IMF (Intrinsic Mode Function) were extracted as the initial characteristic parameters, which would be optimized by genetic algorithm to obtain the optimal characteristic parameters. The results of research states that energy of acoustic emission signals changed with different frequency when the metal micro crack occurred, while each IMF of local wave decomposition respectively contained from low to high different frequency to extrude the local characteristics of the crack signal. Thus it was viable to extract the energy value of the IMF as the initial characteristic parameters of analyzing the crack characteristics, and solve the extraction and optimization to the characteristic parameters of acoustic emission signals.