Reliability Assessment for Time-Slice Enhanced Bidirectional Thermal Wave Radar Thermography of Hybrid C/GFRP Defects

Abstract

In this article, a bidirectional thermal wave radar thermography (BTWRT) is used for the inspection of hybrid carbon/glass fiber reinforced polymer (C/GFRP) laminates with subsurface defects, and the reliability of BTWRT is deeply investigated based on the probability of detection (POD) analysis. Three classes of C/GFRP laminates are prepared for experimental study, and the hybrid volume ratio of glass fiber and carbon fiber is 1:4, 1:1, and 4:1, respectively. Each class of C/GFRP laminate includes a total of 72 artificial flat-bottom holes with different diameter-to-depth ratios. An 808-nm laser is used as external excitation and modulated by a bidirectional chirp signal. Thermal wave radar feature images are obtained by time-slice imaging algorithms [dynamic component extraction, Hilbert transform (HT), and analytical chirp correlation (ACC)] and conventional algorithms (cross-correlation, chirp lock-in, HT mean, and fast Fourier transform). The imaging quality and reliability of different algorithms are compared and evaluated through the signal-to-noise ratio (SNR) and POD assessment using a hit/miss method, respectively. The SNR comparison and POD assessment results indicate that the ACC time-slice algorithm is more suitable for C/GFRP delamination defect inspection using BTWRT in view of the optimal detectability and reliability. This work provides a valuable reference for BTWRT inspection of hybrid C/GFRP subsurface defects, and the ACC time-slice imaging algorithm has the potential to develop into a highly reliable vision inspection method for the industrial applications out of laboratory.