The Calibration and Sensitivity Aspects of a Self-Referencing Routine When Applied to Composites Inspection: Using a Pulsed Thermographic Setup

This text analyzes the effects of calibration settings on a self-referencing processing routine, when applied to inspecting a set of Carbon Fiber composite structures via pulsed thermographic apparatus. The Self-referencing algorithm operates on acquired thermograms by tracking each pixel location contrast-temporal cooling curve, to expose any potential deviation and acquire quantitative assessment of the host material subsurface condition. The calibration factors in this study include; the kernel size that represents the sound (or defect-free) or reference, temperature-time history, the threshold in each center pixel signal relative to its surrounding (average of the kernel pixel population). The study quantifies the effect of each factor using a pulsed thermographic test campaign; first applied to a set of artificially designed samples; in addition to actual blind samples. The controlled coupons are 3D printed out of plastics; while the defects are embedded (via 3D printing) to represent different and controlled shapes and configurations. The 3D printed samples are used to help represent different fill-rates to imitate honeycomb structures, and to embed defects of finite thickness. The study findings highlights the effect of each of the calibration settings and demonstrate tangible improvement in the processed frames.