Damage localization based on the reconstructed Green’s function from a diffuse noise field on a thin rectangular aluminum plate

Abstract

The extracted Green’s function from ambient noise cross-correlation is applied to localize a damage on an aluminum plate using a passive sensor array. Damage was investigated by drilling holes at various locations with different sizes. The localization process uses a subtracted Green’s function of an undamaged plate from that of a damaged plate, which can reveal the scattered wave from the damage boundary. Localization algorithms used in active structural health monitoring, including the time of arrival, the time difference of arrival, the energy arrival and the Rayleigh maximum likelihood estimation [Flynn et al., Proc. R. Soc. Lond. A. Math. Phys. Sci. (2011)] methods are applied to the subtracted signal according to the number of sensors and that of noise sources. The subtracted signals are band pass filtered with various center frequency resulting in multiple images which are used in image fusion to increase the accuracy of localization [Michaels et al., Wave motion. (2007)]. The performance of the algorithms demonstrates the efficacy of the passive sensing-based localization method even when using small number of sensors and sources.