Speaker
Description
Detection and characterization of delaminations and adhesion defects remains a challenge in many industries, applications and materials: from manufacturing defects and low impact damage in composites, to quality control of new bonding technologies such as magnetic pulse welding of dissimilar metals. Optically excited lock-in infrared thermography (IRT) has shown a high potential to size the depth and thickness (thermal resistance) of ideal uniform delaminations. In this work, we show the potential of lock-in IRT to size the depth, lateral dimensions and thickness of more realistic delaminations featuring smooth variations of the parameters along the defect. The methodology is based on solving a parameter estimation problem that relies on a two-dimensional model of the delamination. The surface temperature is calculated semi-analytically by applying the cosine Fourier transform together with the quadrupoles method. The model parameters (depth, length and thickness of the delamination) are determined by fitting the model to experimental data obtained on samples containing calibrated delaminations. For this purpose, we have manufactured artificial delaminations with smoothly varying length, depth or thickness and combinations of two of them in AISI-304 stainless steel. The experimental campaign includes lock-in thermography experiments on a variety of specimens at several frequencies. The real and imaginary part of the complex surface temperature, obtained from experimental amplitude and phase data, are the inputs to fit the two-dimensional theoretical model to estimate the length, depth and thickness of the artificial delaminations. The consistency of the results was checked on each sample by fitting data taken at different frequencies. It has been found that in AISI 304- stainless steel, the depth of delaminations down to 3 mm can be sized accurately. The length of the delaminations has been obtained with accuracies better than 10% and the thickness remains the most challenging parameter to be determined. In AISI-304, delaminations up to 50-60 microns thick can be sized accurately but sizing thicker delaminations is difficult because the contrast between damaged and sound regions saturates and the sensitivity decreases. This maximum sizable thickness would be higher (lower) in materials of lower (higher) thermal conductivity. We think that these findings boost IRT as a reliable technique for the characterization of real delaminations.
