Damage detection and localization via cross-correlation on metallic panels under ambient loading

Detecting and localizing damage of aircraft structures using naturally occurring air flow provide a viable means of developing more efficient structural health monitoring systems. Such passive sensing techniques eliminate the need for acquiring power either from the battery or ambients to generate controlled excitation. In a recent study, high-pressure air was randomly sprayed on metallic panels via air blow gun to mimic nonstationary and randomly distributed nature of turbulent flow occurred during operation of the flight. Self-Green's functions (SGF) were reconstructed using auto-correlation (AC) at 169 (13 × 13) sensing points covering a 150 mm ´ 150 mm area. Mean square error (MSE) of SGFs between damaged and pristine metallic panels was able to map the damage location in the panel. To inspect a larger area with fewer sensing points, a damage imaging technique based on cross-correlation (CC), reconstruction algorithm for probabilistic inspection of damage (RAPID), and image fusion is proposed in this paper. A flat aluminum panel was tested to validate the proposed idea, and an integrally stiffened aluminum panel was studied. Green's functions (GF) were extracted between sparsely distributed sensor arrays. Eight sensors were attached to each of the aluminum panels to cover an enclosed area (up to 200 mm ´ 200 mm) to record the structural response for the pristine and damaged panel. Compressed air was manually swept the panel in arbitrary directions to create nonstationary and random excitation. GFs were reconstructed using cross-correlation, and RAPID and image fusion were used for generating the image of the damage location. The result indicates that the stability of GFs is the key to accurately pinpoint the damage. At least 20 seconds of long-time recording from sensor guarantees such stability, and well-distributed air impact locations results in symmetric GFs. Damage images are in good agreement with the real damage locations. The proposed technique may detect and localize damage utilizing turbulent flow, and it enlarges the inspection region while significantly reduces the sensing points.