Bayesian-updated POD of ultrasonic tomography in the inspection of PT bridges

The present research aims to furnish a proper methodology for the reliability assessment of non-destructive tests in the particular field of inspection of post-tensioned (PT) structures. The focus is on the Ultrasonic Tomography used in the inspection of grouting defects inside the PT ducts. The presence of such defects is the most dangerous scenario for triggering the mechanism of corrosion in the prestressing steel and, therefore, for structural safety. Many studies exist about the application of ultrasonic tomography (UT) in the inspection of PT structures and bridges. However, the reliability of the technique is still not known when applied to field inspection of existing bridges. This paper explains the execution of a laboratory testing campaign on many specimens presenting different sizes and types of simulated grouting defects. The tested specimens are designed to represent the most typical source of grouting defects (voids, water) and the most spread typology of ducts (metallic and plastic) and post-tensioning elements (strands and bars). The inspection outcomes are modeled through the Probability of Detection (POD) curve, which gives information about the maximum defect size that the inspection method, UT, can miss, stating a certain confidence level. Since the number of tests is limited due to practical issues, as will be also the case in a real bridge, a Bayesian-updating approach is employed to better estimate the model parameters of the POD curve. In this way, the reliability of the inspection methodology can be derived, which is of great importance in the posterior capacity assessment of the existing structure.