Quantitative Evaluation of Interfacial Defect Size and Pattern by Solving a 3D Inverse Problem Based on Step Heating Thermography

Active infrared thermography is proved to be viable and attractive for non-destructive evaluation of interfacial defects like delaminations in a coating-substrate system. But it is a challenging task to accurately quantify small and deeply buried defects from thermal images, due to the inevitable effects of lateral heat diffusion and measurement noise. The aim of this work is to estimate the size and pattern of defects at the interface of a two-layer system with high accuracy and high reliability based on step heating thermography. To characterize the effect of defect on the heat flow, a virtual heat flux is assumed at the interface, which is reconstructed from measured surface temperature by solving a three-dimensional inverse problem. The inverse solution is obtained using the Green’s function and regularization techniques, and then used for estimating the defect pattern by threshold segmentation. An improvement on computational efficiency is achieved by an iteratively substitution of nodal temperature. Simulations with synthetic data generated by a finite element model validate the feasibility of this approach. Results obtained from experiments for an Aluminum oxide/steel system show the robustness of this approach, when temperatures are contaminated with measurement noise. Both the performance on estimation of various defect shapes and the effects of regularization are discussed. This study show that the present approach brings an improvement in accuracy and reliability for the estimation of size and pattern of defects with various diameter-to-depth ratios, in comparison with conventional techniques.