Detection and quantification of corrosion defects in CFRP-strengthened steel structures based on low-power vibrothermography
Pengqian Liu,Changhang Xu,Yubin Zhang,Lina Chen,Yage Han,Rui Liu,Yi Qin
DOI: https://doi.org/10.1080/10589759.2024.2326603
2024-03-06
Nondestructive Testing And Evaluation
Abstract:For CFRP-strengthened steel structures, corrosion defects in steel are incredibly hazardous, but there is a lack of effective non-destructive testing methods. Based on the low-power vibrothermography (LVT) technique, this study proposes a quantitative detection framework for corrosion defects of CFRP-strengthened steel structures, including feature extraction algorithms for improving the defect detectability, and a feature enhancement method for reducing the quantification error. First, a finite element model is built to simulate the heat generation under ultrasonic excitation and to obtain the temperature distribution on the specimen surface. The processing results of the simulated thermal images show that feature extraction algorithms (FFT, PCA, and PLSR) effectively extract defect features. Subsequently, the LVT experimental setup was established to detect corrosion defects in CFRP-strengthened steel specimens. The experimental results demonstrate that the LVT technique can selectively heat corrosion defects, and processing results prove that the FFT and PLSR can further enhance the defect detectability. Finally, we propose a feature enhancement method based on gravitational calculation and double-threshold segmentation to reduces the quantification error. The quantitative evaluation relative errors of defects with diameters of 10 mm, 8 mm, and 6 mm are 5.26%, 5.26%, and 10.09%, respectively.
materials science, characterization & testing