Experimental Results of Acoustic Emission Attenuation Due toWave Propagation in Composites

Abstract

Recently, acoustic emission-based damage classification schemes gained attention for health monitoring of composites. Here, the reliable detection of different micro-mechanical damage mechanisms is important because of the adverse effect on fatigue life. It is well known that classical parameters obtained from acoustic emission measurements in time domain are strongly dependent on the propagation path and testing conditions. However, signal attenuation, which can be observed due to geometric spreading, material-related damping, and dispersion, is typically neglected. Therefore, it is generally assumed that frequency domain features are reliable descriptors of damage due to invariance of peak frequencies to the propagation path. Based on this assumption, several data-driven approaches for damage detection are developed. However, in contrast to metallic materials, where low attenuation is observed, acoustic emission signals are strongly attenuated in polymer matrix composites due to viscoelastic behavior of the matrix. For instance, it is reported in literature that at high frequencies most of the acoustic emission signal energy is attenuated after a propagation distance of 250~mm. Therefore, new experimental results of acoustic emission attenuation in composites are presented in this paper. Particular focus is placed on the frequency dependence of acoustic emission attenuation and the effect of different loading conditions. The specimens are manufactured from aerospace material. Carbon fiber reinforced polymer plates are used as a typical specimen geometry. Different acoustic emission sources are considered and the related attenuation coefficients are determined. Furthermore, full waveform data are analyzed in time and time-frequency domain using wavelet transform. From the experimental results it can be concluded that consideration of wave propagation-related signal attenuation is important for the interpretation of acoustic emission measurements for health monitoring of composites. Consequently, the impact on the detectability of different physical damage mechanisms using data-driven classification approaches has to be considered.