Corrosion damage detection of anchorage zone of steel cables using ultrasonic guided waves

Steel cables are important load-bearing components of cable bridges, and they are also vulnerable to damage. The cable part near the anchorage zone often suffers the most serious corrosion damage, which is difficult to detect and quantify due to the structure complexity of the multiwire cable with the anchorage. In this study, the ultrasonic guided wave (UGW) based on magnetostrictive effect, which has high damage sensitivity and large detection range, was applied to detect the corrosion damage. The dispersion curve of UGW in the cable was first calculated by the semi-analytical finite element (SAFE) method to understand the propagation characteristics of guided waves and select optimal excitation frequency. Simulation of corrosion damage was also given to analyze the signal characteristics. Above results guided the following experiments. The accelerated corrosion test was designed to produce the different corrosion degree of the steel cable anchorage zone. The guided wave signals obtained from the experiments were analyzed first by the synchronous compression transform, for denoising and extracting the amplitude of the reflected UGW signal. Though the corrosion reflected signal was submerged in the front-end reflected signal, its amplitude is extracted by constructing pure front-end anchorage amplitude curves to study the variation of corrosion damage. Finally, machine learning classification models were selected to classify the anchorage-reflected signal and the corrosion-reflected signal when no baseline data are available. In practical engineering, utilizing machine learning classification models can swiftly identify corrosion damage at the front-end of anchorages, thereby enhancing the detection efficiency and meeting the requirements of routine maintenance assessments.