EUVIMEDDécouvrir EUVIMED
Crossref

Non-destructive experimental fault diagnosis of wind turbine drivetrain components using tower-base accelerometers

Francesco Castellani, Davide Astolfi, Alessandro Canali, Alessandro paolo Daga, Emiliano Sisinni, Alessandra Flammini

e-Journal of Nondestructive Testing · 2026-07 · 31 (7)

DOI: 10.58286/33501 ↗

At a glance

Journal article in e-Journal of Nondestructive Testing, 2026-07. Written by Francesco Castellani and 5 co-authors.

RésuméRéférences bibliographiquesCiteRelated studies

Abstract (original language)

In wind energy, the development of condition-based maintenance (CBM) strategies is of paramount importance in the effort to minimise the levelised cost of energy (LCOE). In this context, vibration-based approaches are widely regarded as the most reliable indicator of component health. Recent studies have demonstrated that vibration signals collected at the tower base can effectively and safely reveal drivetrain faults, even when the drivetrain geometry and vibration transmission paths are not known in detail. As demonstrated by the preceding findings, the present work presents the results of an industrial measurement campaign that was conducted on an operational wind farm. The proposed methodology is a two-step, data-driven approach in the frequency domain. Firstly, an unsupervised autoencoder neural network trained using healthy data achieved over 99% accuracy in identifying faulty turbines via reconstruction-error analysis of selected Fourier features. Secondly, a semi-supervised feedforward neural network was employed to refine the diagnosis by estimating the probability of a weak faulty turbine belonging to the same class as a severely faulty turbine. This yielded values below 50%, thus enabling diversification of fault severity. This combined methodology demonstrates high diagnostic capability with minimal supervision, paving the way for practical, cost-effective, condition-based monitoring (CBM) of wind turbines.

Cite

APA

Castellani, F., Astolfi, D., Canali, A., Daga, A. P., Sisinni, E., & Flammini, A. (2026). Non-destructive experimental fault diagnosis of wind turbine drivetrain components using tower-base accelerometers. e-Journal of Nondestructive Testing, 31(7). https://doi.org/10.58286/33501

Vancouver

Castellani F, Astolfi D, Canali A, Daga AP, Sisinni E, Flammini A. Non-destructive experimental fault diagnosis of wind turbine drivetrain components using tower-base accelerometers. e-Journal of Nondestructive Testing. 2026;31(7). doi:10.58286/33501

BibTeX
@article{Castellani2026,
  author = {Castellani, Francesco and Astolfi, Davide and Canali, Alessandro and Daga, Alessandro paolo and Sisinni, Emiliano and Flammini, Alessandra},
  title = {Non-destructive experimental fault diagnosis of wind turbine drivetrain components using tower-base accelerometers},
  journal = {e-Journal of Nondestructive Testing},
  year = {2026},
  volume = {31},
  number = {7},
  doi = {10.58286/33501}
}
Start for free →