Abstract
An inversion procedure is presented to reconstruct buried heat sources (revealing defects) from surface temperature data obtained by multifrequency lock-in vibrothermography. The severe ill-posedness of the problem is overcome by regularizing the minimization of the squared differences between experimental and calculated data. Two regularization functionals, zero-order Tikhonov and total variation, have been tested by inverting synthetic data. For added uniform white noise levels as high as 20%, total variation has proven to give more accurate inversions. This procedure is applied to reconstruct heat sources from experimental lock-in vibrothermographic measurements performed on metallic samples containing calibrated inner heat sources. The results are very promising regarding the characterization of hidden defects using lock-in vibrothermography.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call