Identification of debonding in CFRP stiffened panels nsiog pattern recognition

Composite materials have been increasingly employed in primary aerospace structures; particularly in modern unmanned aerial vehicles (UAV). Unlike metals for which direct quantitative paths between microstructure and overall mechanical behavior have been established and are continuously refined, the activation and evolution of failure mechanisms in composite materials represents a hierarchically more complex mechanics problem, especially in field applications. To address this issue, advanced inspection techniques and nondestmctive methods have been implemented to identify signs of early damage and prevent catastrophic failure. Among them, Acoustic Emission (AE) has shown perhaps the most potential of identifying the onset of the different failure mechanisms in composite materials in laboratory conditions. Due to recent advances in hardware/software, AE has the potential to be applied in actual aircrafts. In this study, AE was used to identify and distinguish signals arising from debonding in carbon fiber reinforced polymer (CFRP) stiffened panels, primarily used in Unmanned Aerial Vehicle (UAV) wings. To this aim, the authors demonstrate an approach to identify and classify these AE signals through pattern recognition (PR). Specifically, CFRP stiffened panels were loaded under both monotonie and fatigue bending conditions and critical feature descriptors sensitive to debonding were identified. The K-means pattern recognition approach was then attempted to identify the class of signals associated with debonding. The overall approach revealed that the class of signals associated with debonding has similar feature characteristics across the different tested panels, and thus a framework is presented suitable for future robust damage detection for onboard applications.