Aircraft Fastener Defect Detection through Polar-Wavenumber Filtering of Full-Field Guided Wave Data

Abstract : Scanning laser systems are a class of ultrasonic guided wave systems that have shown great promise for the inspection of complex structures because of the high spatial resolution of their resulting data. In this study, the focus is on applying full-field signal processing techniques to automatically detect and characterize fastener damage in aircraft components with realistic structural complexity. The particular testbed being considered is a composite tail section with several rivet lines. Artificial damage was introduced into several of the rivets with different magnitudes and orientations. While trained users of the laser scanning system can often detect the damage modes by simply viewing the resulting images, a more automated and quantifiable approach is desired to increase the consistency and quantify the confidence of the inspection system. In order to characterize the state of each rivet, a polar wavenumber processing technique is proposed. First, the full-field data is transformed to a new polar coordinate system centered on a particular rivet. A 3D Fourier transform then transforms the data to the frequency-wavenumber domain. Finally, the data is windowed to isolate the waves propagating in the positive radial direction. Analyzing the scattering pattern from each rivet in turn provides information on the health of each rivet. Damage detection may be accomplished either by comparison of the extracted scattering pattern with a model of the expected scattering pattern, or by an absolute reference-free method comparing rivets to one another to identify outliers. The results show that there is the potential to provide quantified damage detection and characterization for fastener damage using laser-excitation guided wave inspection systems.