Discrete-Wavelet Analysis of Acoustic Emissions During Fatigue Loading of Carbon Fiber Reinforced Composites

Abstract

Wavelet transform decomposition was used to gather time-frequency-based information from acoustic emission signals generated during fatigue loading of unidirectional Carbon Fiber Reinforced Composite (FRC). The acoustic emissions were detected using a resonant sensor and were digitized for analysis. The sensor response was de-convolved from the acquired signal using a point-by-point divisional spectral method. The analysis of the collected signals revealed that friction-related emissions due to the fretting of fractured surfaces are of very high frequency and can mask emissions due to actual damage. Through utilization of wavelet transforms, it became possible to present the spectral composition of a transient signal (AK signal) in a time-frequency map which is not easily achieved through conventional spectral analysis techniques. It was determined that most of the acoustic energy (95%) was localized in levels corresponding to central frequencies of 120, 250, and 310 kHz. Results indicate friction-related emissions are associated with levels 8 and 9 and have a frequency range of 250-300 kHz. There are indications that matrix related emissions are of high frequency and high acoustic energy. The results indicate that wavelet analysis would be an effective tool in the analysis of AE by providing information relative to the frequency of the emissions.