2024-11-262024-11-262024Ghiasi, R., Jiménez Alonso, J.F. y Malekjafarian, A. (2024). Data-driven framework for damage detection in structural system. En Bridge Maintenance, Safety, Management, Digitalization and Sustainability - Proceedings of the 12th International Conference on Bridge Maintenance, Safety and Management, IABMAS 2024 (2327-2333), Copenhagen: CRC Press.978-103277040-6https://hdl.handle.net/11441/164950© 2024 The Author(s)This paper presents an automated methodology for data-driven damage detection (DD) based on feature extraction from the measured vibration responses of civil engineering structures subjected to ambient/operational conditions. To do so, after recording the dynamic response of the structure, feature extraction is performed in the time domain and then the best subset of measurement features that characterize the state of the structure is selected using Analysis of Variance (ANOVA) algorithm. Finally, DD is performed using a machine learning (ML) algorithm trained with selected features. To test the proposed framework, a dataset from a laboratory steel structure is used, and two supervised ML algorithms, K-nearest neighbor classification (KNN) and artificial neural network (ANN) were applied in order to classify damage states. The results showed that both classification techniques were able to successfully classify damage. In addition, the impact of using different feature extraction/selection methods on the accuracy of the proposed approach is studied in detail.application/pdf7 p.engAttribution-NonCommercial-NoDerivatives 4.0 Internationalhttp://creativecommons.org/licenses/by-nc-nd/4.0/info:eu-repo/semantics/conferenceObjectinfo:eu-repo/semantics/openAccess10.1201/9781003483755-276