No abstract

No abstract

The ability of lock-in vibrothermography to distinguish between different sizes, shapes and locations of vertical kissing defects, such as cracks or corrosion, is analyzed in this work both theoretical and experimentally. We have computed the oscillating, AC, surface temperature of samples containing inner modulated heat sources, representing the defect, and analyzed the sensitivity of this oscillating temperature to the defect geometric parameters. Moreover, we have prepared samples with calibrated vertical defects. Vibrothermographic data are usually recorded under non-steady conditions (i.e. while the surface temperature is still rising). As this transient temperature rise has a non-negligible component at the excitation frequency, the measured AC temperature is strongly distorted. By subtracting the Fourier component of the transient temperature rise from vibrothermographic data, we obtain the pure AC temperature (amplitude and phase) in very good agreement with the theoretical predictions.

A method to characterize vertical cracks which combines lock-in vibrothermography experiments and a stabilized inversion algorithm is presented. The capability of the method to retrieve the shape and location of square cracks buried at increasing depths is analyzed by inverting synthetic data with added white noise. Moreover, the ability of the method to resolve two square buried cracks as a function of depth is studied. A relationship between the depth of the cracks and the minimum distance between them in order to obtain separate reconstructions is found. The predictions of the inversions with synthetic data have been checked by inverting experimental data obtained from steel samples containing calibrated cracks. In order to prevent the sample from being damaged and to reach the steady state faster, low ultrasound excitation powers are used. The signal-to-noise ratio corresponding to the resulting weak signal is improved by processing a large number of infrared images using the lock-in method. The experimental results confirm that the method is able to reconstruct square heat sources whose depth is twice their lateral dimension, and that the resolution condition holds.

This work combines lock-in vibrothermography experiments and a stabilized inversion algorithm to characterize vertical cracks. The ability of the method to retrieve square cracks at increasing depths is analyzed by inverting synthetic data with different levels of white noise. Moreover, the minimum distance between square defects in order to be resolved is determined as a function of depth. Inversions of experimental data obtained at low intensity ultrasound excitation in samples containing calibrated defects confirm the predictions of inversions performed with synthetic data.