Application of machine learning to a novel analytical method of AE waveform signals

Abstract

Acoustic emission (AE) is an elastic wave generated by the release of energy when the damage occurred in a material. In this research, we proposed a new signal processing technique by applying a neural network to analyzing the AE waveform in CFRP laminates. Through verification experiments, we demonstrated that the proposed method was able to clarify the complex relation between the AE singles with different waveform features and the damage-induced AE sources with individual mechanisms in exciting AEs. Hence, the new technique is potentially used to identify the types of damages, such as transverse crack and delamination, in CFRP laminates.