In the field of tunnel structure inspection, non-destructive testing techniques have replaced core drilling as the primary means of tunnel disease detection due to their non-destructive nature and operational convenience. Currently, the most widely used non-destructive testing methods in tunnel engineering include ground-penetrating radar (GPR) and machine vision technology, corresponding to hidden defects and surface defects, respectively. However, there is limited systematic research on the interference of tunnel lining structures with the electromagnetic waves emitted by ground-penetrating radar. Important parameters such as central frequency, polarization direction, and height above the ground lack comprehensive studies on their impact on the detection of hidden defects. The research team participated in the design and development of a multi-offset ground-penetrating radar system and applied it to the detection of grouting layers behind tunnel walls. The applicant proposed a method for detecting internal defects in concrete based on ultrasound imaging, clarifying the mechanism behind the formation of pseudo-images of structural hidden defects and significantly improving imaging accuracy. Using deep learning algorithms, the applicant proposed an automatic identification and localization method for underground targets detected by ground-penetrating radar. Through on-site experiments, it was verified that the absolute error in estimating the depth was less than 0.04 meters, and the average relative error was less than 4%. This method can achieve high-precision detection of buried pipelines around tunnels. Addressing the issue of electromagnetic wave scattering caused by densely embedded steel bars in concrete structures severely affecting the penetration capability of GPR waves, the applicant found through numerical and experimental studies that steel bars that are perpendicular to the polarization direction of GPR waves and have diameters much smaller than the wavelength are almost transparent to incident GPR waves. Additionally, steel bars parallel to the polarization direction of GPR waves cause scattering and interaction, leading to a shielding effect, manifesting as a blind zone in the low-frequency range of the transmission spectrum. This contradicts the commonly used empirical rule that lower frequencies have greater GPR penetration depths. This achievement provides a theoretical basis and technical support for the detection of steel bar corrosion in the internal structures of tunnels.
Collaborator: Prof Hai Liu (Guangzhou University)