Damage Detection in Thin-Walled Structures Using Broadband Air-Coupled Guided Wave Data

Abstract

Ultrasonic testing using guided waves is a powerful method for non-destructive testing of thin-walled structures, especially in maintenance-intensive industries like aerospace. The objective of this thesis was to systematically evaluate and compare advanced dam�age imaging algorithms using a novel robotic test stand equipped with an air-coupled optical microphone. To enable controlled testing, aluminum and composite specimens were specifically designed and manufactured with a range of artificial defects. Three main methods, Local Wavenumber Estimation (LWE), cumulative signal value mapping, and Wavelet-Transformed Ultrasonic Propagation Imaging (WUPI), were implemented, adapted, and tested on the recorded datasets. These systematic experiments demonstrate that, for this novel setup and specific imple�mentation, each method has distinct strengths and limitations. LWE enables quantitative defect characterization and precise mapping of larger or deeper defects, but is computa�tionally intensive and less effective for small or shallow damage. Cumulative signal value mapping is robust, simple, and efficient, providing reliable qualitative defect localization but lacking depth sensitivity. WUPI offers high spatial resolution and excels at detecting small or complex defects, though it is more sensitive to noise and produces more complex outputs. These findings provide practical guidance for selecting imaging algorithms for effectively extracting damage information from dense wavefield data.