<title>Thermographic detection and quantitative characterization of corrosion by application of thermal line source</title>

Abstract

Recent advances in thermal imaging technology have spawned a number of new thermal nondestructive evaluation (NDE) techniques that provide quantitative information about flaws in aircraft structures. Thermography has a number of advantages as an inspection technique for aircraft. It is a totally noncontacting, nondestructive, imaging technology capable of inspecting a large area in a mater of a few seconds. The development of fast, inexpensive image processors have aided in the attractiveness of thermography as an NDE technique. These image processors have increased the signal to noise ratio of thermography and facilitated significant advances in post-processing. The resulting digital images enable archival records for comparison with later inspections thus providing a means of monitoring the evolution of damage in a particular structure. NASA Langley Research Center has developed a thermal NDE technique designed to image and quantitatively characterize the thickness of thin aluminum sheets. The technique involves the movement of a thermal line source across the outer surface of a sample followed by an IR imager at a fixed distance behind the line source. Images of the material loss due to corrosion are reconstructed from measurements of the induced surface temperature variations. This paper will present a discussion of the development of the thermal imaging system as well as the techniques used to reconstruct images of flaws. The application of the thermal line source coupled with the analysis technique represents a significant improvement in the quantification of flaws over conventional thermal imaging. Results of laboratory experiments on specimens with fabricated material loss region swill be presented to demonstrate the capabilities of the technique. An integral part of the development of this technology is the use of analytic and computational modeling to optimize the technique and reduce the data. The experimental results will be compared with simulations to demonstrate the utility of such an approach.