Integration of Pattern Recognition and Machine Learning with the Acoustic Emission Method to Locate and Assess Corrosion in Cable-Stayed and Suspension Bridge Post-Tensioned Cable Anchorages

The intensive development of transport infrastructure globally and in Poland has led to a rapid increase in the number of bridge structures. Prestressed concrete is currently the most widely utilized structural material, accounting for 43.4% of these structures. A primary advantage of prestressed concrete is its capability to achieve considerable span lengths; consequently, its percentage share in terms of total bridge surface area is even higher, reaching 58.2% by the end of 2017. Although visual inspections are feasible for exposed tendon components, evaluating the residual prestressing force and diagnosing internal cable degradation—such as corrosion, grout deterioration, and voids—in post-tensioned structures presents a significant technical and scientific challenge. This paper introduces a structural health monitoring (SHM) approach, utilizing either periodic inspections or continuous electronic monitoring, to evaluate anchorage condition. The proposed methodology employs a novel measurement system that identifies structural anomalies by utilizing pattern recognition algorithms applied to acoustic emission (AE) signals. Furthermore, the identified pattern classes have been correlated with crack opening widths. This correlation enables the tracking of crack propagation effects on structural stiffness, while simultaneously monitoring other degradative processes, including active corrosion and anchorage slippage.