Passive flaw detection and localization in thin plate from ambient noise cross-correlation

Abstract

Passive structural health monitoring (SHM) is an emerging technology. More than 10 years ago, it has been shown that transient response between two sensors can be passively estimated from cross-correlation of ambient noise. The work presented here is an experimental application of this approach dedicated to detect the occurrence of flaws on a thin aluminum plate. The detection sensitivity is directly related to the fidelity of the estimation of the transient response. Using a laser vibrometer, we show that it strongly depends on the number of uncorrelated noise sources. An artificial damage is detected from the difference between the cross-correlation matrices measured before and after defect appearance. To localize the defect, a beamforming array processing is applied on the matrix. The resolution can be as small as half a wavelength.