Jensen, Jens SandagerFrangopol Dan M.Schmidt, Jacob Wittrup2024-11-282024-11-282024Naranjo Pérez, J., Jiménez Alonso, J.F., Muñoz Díaz, I., García Palacios, J.H. y Concha Renedo, C.M.d.l. (2024). Vibration-based NDT system for external tendons: Anomaly detection through machine learning classifiers. En 12th International Conference on Bridge Maintenance, Safety and Management (IABMAS 2024) (2303-2310), Copenhagen, Denmark: CRC Press.97810327704069781003483755https://hdl.handle.net/11441/165062This paper proposes a new prototype for the non-destructive testing of external post-tensioning tendons. The system is based on the dynamic response of the tendon from which a series of structural performance indicators are identified, allowing to assess them and to make maintenance or replacement decisions in case anomalies are detected. The prototype developed is economical, easy to use, scalable (more sensors or performance indicators can be included) and portable, since all the necessary equipment is placed on a trolley. The system has been used for the evaluation of the external post-tensioning tendons of a 12-span railway bridge, with a total of 202 tendon segments analyzed. For each test, the system generates a report summarizing the calculated indicators and pointing out any anomalies detected. Considering the results of all the segments, an unsupervised clustering technique is applied to create groups with similar patterns where the results obtained are classified and the presence of outliers, i.e. tendon sections with anomalous behavior, is checked.application/pdf8 p.engAttribution-NonCommercial-NoDerivatives 4.0 Internationalhttp://creativecommons.org/licenses/by-nc-nd/4.0/Vibration-based NDT systemMachine learninginfo:eu-repo/semantics/conferenceObjectinfo:eu-repo/semantics/openAccess10.1201/9781003483755-273