| dc.description.abstract | **Please note that the full text is embargoed until 8/1/2025** ABSTRACT: Ground Penetrating Radar (GPR) has emerged as a valuable nondestructive testing technique for subsurface imaging and characterization in civil engineering applications. In particular, GPR plays a crucial role in the evaluation of reinforced concrete (RC) structures, enabling the determination of concrete cover thickness and the localization of reinforcement. However, the effectiveness of GPR data analysis is often hindered by inherent challenges, such as unknown time-zero, strong noise, and blurred signals, making the accurate and simultaneous determination of rebar horizontal location and depth a challenging task.
This Ph.D. research aims to address these challenges and enhance the accuracy and efficiency of GPR data processing for RC structures. The research consists of three main chapters, each focusing on a specific aspect of GPR data analysis.
Chapter 2 presents a novel time-zero (TZ) correction method specifically designed for RC structures. The proposed approach involves the identification of the first negative peak in the direct wave as the temporary TZ, followed by the application of an adjusting value to obtain the true TZ. This innovative approach results in more accurate depth measurements and facilitates reliable assessment of concrete infrastructure with GPR. To validate the proposed method, GPR scanning of 32 RC specimens with different rebar depths, sizes, and spacings was conducted, leading to the determination of a specific adjusting value of 0.14299 ns. The validation experiments confirmed the accuracy and reliability of the proposed TZ correction method, opening new avenues for improved GPR data analysis in RC structures.
Chapter 3 introduces a nondestructive algorithm for accurately estimating GPR's electromagnetic wave velocity. The algorithm leverages hyperbolic fitting and travel-time analysis, offering a practical solution for velocity estimation without the need for core drilling. Various factors, such as subsurface media type, moisture content, temperature, and antenna frequency, affect GPR wave propagation velocity. Traditional methods for velocity estimation often rely on empirical models or assumptions, limiting their accuracy in complex subsurface conditions. In contrast, the proposed algorithm utilizes advanced techniques that account for the variation of wave velocity with depth and consider the effects of subsurface heterogeneity, resulting in more precise velocity estimation. Laboratory experiments successfully validated the algorithm's accuracy and robustness, demonstrating its potential to enhance subsurface imaging capabilities and improve data interpretation.
Chapter 4 presents a comprehensive data processing algorithm for rebar localization in RC structures. The automated algorithm addresses challenges such as unknown time-zero, strong noise, and blurred signals, which are common in GPR data. By eliminating the need for manual interpretation or rebar-picking, the algorithm achieves full automation, enhancing efficiency and accuracy in data processing. Additionally, the proposed algorithm corrects time-zero and calculates electromagnetic wave velocity without the requirement for core-drilling, further streamlining the data processing workflow. Validation on various datasets, including lab-made reinforced concrete blocks, bridge decks, and a culvert, demonstrated promising performance in determining the rebar's location. Compared with existing methods, the proposed algorithm proved to be cost-effective, practical, and efficient, providing accurate and reliable rebar localization.
In conclusion, this Ph.D. research makes significant contributions to the GPR-based assessment of RC structures through the development of innovative TZ correction, velocity estimation, and rebar localization methods. The validated approaches improve the accuracy, reliability, and efficiency of GPR data processing, paving the way for more informed decision-making in civil engineering and geophysical applications. The proposed methods open avenues for further research, such as exploring noise reduction techniques, machine learning-based approaches, and real-time GPR data processing integration. Overall, this research empowers civil engineers and researchers with robust tools for effective GPR data analysis and structural assessment in concrete infrastructure evaluation scenarios. | |
