EUVIMEDEUVIMED entdecken
Crossref

Framework for Event Detection and Unsupervised Clustering of Large-Scale SHM Data

Liv Breivik, Reza Allahvirdizadeh, John Leander

e-Journal of Nondestructive Testing · 2026-08 · 31 (8)

DOI: 10.58286/33959 ↗

Kurz gefasst

Fachartikel in der Zeitschrift e-Journal of Nondestructive Testing, 2026-08. Verfasst von Liv Breivik und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Structural Health Monitoring (SHM) systems produce large amounts of time-series data, where loading scenarios are of interest for evaluating bridge health. The prevalence of different types of loading events, such as different train types passing over a bridge, can hinder the reliability of data-driven event-based health evaluation methods due to local distribution shifts. Neglecting event classification poses a significant challenge to the reliability of data-driven SHM analyses, and this challenge is amplified by the scale of the data, emphasizing the need for low-cost, automated event detection solutions. To address this challenge, we develop an unsupervised event classification framework. This framework is based on using distances as linkage within unsupervised clustering methods, aiming to distinguish different train passage types from each other. A time warping method is used for getting distances between sensor signals of varying duration and alignment. Reduction of data in order to speed up computation time whilst retaining accuracy is taken into consideration. The developed framework is validated by applying it to a real bridge, demonstrating its practical utility and scalability.

Zitieren

APA

Breivik, L., Allahvirdizadeh, R., & Leander, J. (2026). Framework for Event Detection and Unsupervised Clustering of Large-Scale SHM Data. e-Journal of Nondestructive Testing, 31(8). https://doi.org/10.58286/33959

Vancouver

Breivik L, Allahvirdizadeh R, Leander J. Framework for Event Detection and Unsupervised Clustering of Large-Scale SHM Data. e-Journal of Nondestructive Testing. 2026;31(8). doi:10.58286/33959

BibTeX
@article{Breivik2026,
  author = {Breivik, Liv and Allahvirdizadeh, Reza and Leander, John},
  title = {Framework for Event Detection and Unsupervised Clustering of Large-Scale SHM Data},
  journal = {e-Journal of Nondestructive Testing},
  year = {2026},
  volume = {31},
  number = {8},
  doi = {10.58286/33959}
}
Kostenlos starten →