EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A ResNet–TadGAN Hybrid Framework for Robust Detection of Imbalanced and Subtle Anomalies in Structural Health Monitoring

Mohammad Hossein Abbasi, Fatemeh A. Mehrabadi, Maryam Bitaraf

Structural Control and Health Monitoring · 2026-01 · 2026 (1)

DOI: 10.1155/stc/3434617 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Structural Control and Health Monitoring, online erschienen am 24.08.2026. Verfasst von Mohammad Hossein Abbasi und 2 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Addressing anomalies caused by sensor malfunctions, environmental influences, and operational variability is crucial for developing a reliable structural health monitoring (SHM) system. This study proposes a hybrid methodology that integrates a ResNet‐18 architecture with a time‐series anomaly detection generative adversarial network (TadGAN) to detect six types of imbalanced anomalies, namely, missing, minor, outlier, square, trend, and drift, in a cable‐stayed bridge. After preprocessing, the acceleration response signals are converted into grayscale images and fed into ResNet‐18 for initial classification. Segments, classified as normal, are subsequently re‐evaluated to detect potential misclassified anomalies. To this end, modification filters, including TadGAN, are applied to refine the detection process. The framework is validated using long‐term acceleration data from the Sutong Bridge, one of the world’s longest cable‐stayed bridges, achieving an overall detection accuracy of 98.1% and 90.94% F 1‐score for drift anomalies, substantially outperforming previous studies. The results demonstrate that the proposed approach is particularly effective in identifying subtle anomaly types, such as drift, and highlight its potential as a robust and scalable tool for real‐time SHM and improved structural condition assessment.

Zitieren

APA

Abbasi, M. H., Mehrabadi, F. A., & Bitaraf, M. (2026). A ResNet–TadGAN Hybrid Framework for Robust Detection of Imbalanced and Subtle Anomalies in Structural Health Monitoring. Structural Control and Health Monitoring, 2026(1). https://doi.org/10.1155/stc/3434617

Vancouver

Abbasi MH, Mehrabadi FA, Bitaraf M. A ResNet–TadGAN Hybrid Framework for Robust Detection of Imbalanced and Subtle Anomalies in Structural Health Monitoring. Structural Control and Health Monitoring. 2026;2026(1). doi:10.1155/stc/3434617

BibTeX
@article{Abbasi2026,
  author = {Abbasi, Mohammad Hossein and Mehrabadi, Fatemeh A. and Bitaraf, Maryam},
  title = {A ResNet–TadGAN Hybrid Framework for Robust Detection of Imbalanced and Subtle Anomalies in Structural Health Monitoring},
  journal = {Structural Control and Health Monitoring},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/stc/3434617}
}
Kostenlos starten →