Guided wave signal processing and image fusion for in situ damage localization in plates

A spatially distributed array of single element piezoelectric transducers is considered for monitoring the integrity of plate-like structures over large areas. When these permanently attached transducers are driven with an impulsive excitation, broadband Lamb waves are generated and received signals are complex, typically containing multiple modes as well as many boundary reflections. The complexity of the signals, including dispersive effects, can be reduced by digital bandpass filtering, and changes in these filtered signals over time can be monitored to detect and localize damage. Results are reported here from an experimental study whereby artificial damage was introduced in an aluminum plate in the form of through holes drilled at various locations. Time shift averaging algorithms were applied to differential signals filtered at multiple frequencies, resulting in many images for the same structural state, and these images were fused to improve damage localization and background noise. Individual and fused images were analyzed to quantify their ability to localize damage. The results show the efficacy of tile imaging method as well as the significant improvement resulting from image fusion. (c) 2007 Elsevier B.V. All rights reserved.