Investigation on the Detection of Crack Defects in Metals Using Chirp Ultrasonic-Induced Infrared Thermography

Chirp ultrasonic infrared thermography is widely applied in material nondestructive testing, while its defect recognition effect is constrained by lateral thermal diffusion and image noise. In this article, a time--frequency domain transient features reconstruction (TFFR) technique has been proposed to solve these issues. First, a 3-D thermal-wave propagation model of metal cracks under a chirp excitation thermal source was developed to analyse the temperature field distribution and thermal-wave diffusion. Second, TFFR was proposed to extract defect characteristics and compared with other algorithms [fractional Fourier transform (FrFT), cross correlation (CC), dual orthogonal demodulation (DOD), principal component analysis (PCA), partial least squares regression (PLSR)]. In addition, the experimental setup of chirp ultrasonic-induced infrared thermography was developed, and the effects of image sequence window size and excitation parameters on TFFR signal-to-noise ratio (SNR) were investigated to determine the optimal parameters. Finally, the crack sizes were calculated and compared with actual measurements, showing that TFFR can effectively reduce lateral heat diffusion and noise, improving SNR. The TFFR phase map achieved minimal size measurement errors, with relative errors of 5.4% and 7.9% for defects 1# and 2#, respectively.