What question did this study set out to answer?

This research aims to compare deep learning architectures for fault detection in small wind turbine blades.

July 2, 2026International Journal of Energy Studies

A comparative study of CNN and vision transformer architectures for fault detection in small wind turbine blades

View Full Paper

Ask AI

Bookmark

Authors

MAMuhammet Fatih Aslan BABusra Aslan SBSelami Balcı

Discussion

Member takes

Overview

Randomized trial evaluates fault detection accuracy in wind turbine blades using CNN and vision transformer models, indicating different effectiveness.

Key Points

This research aims to compare deep learning architectures for fault detection in small wind turbine blades.
Systematic evaluation of five deep learning architectures on the CAI-SWTB dataset.
Utilized binary classification for healthy versus faulty blades using 6,000 RGB images.
Employed transfer learning, AdamW optimizer, and cosine annealing for model training.
EfficientNetV2-S achieved 99.75% accuracy, the highest among all evaluated models.
ResNet-50 followed with 99.42% accuracy, showing strong performance.
Swin-Tiny outperformed the other ViT models with 88.25% accuracy, supporting its effectiveness for fault detection.

Cite This Study

Aslan et al. (2026) studied this question.

synapsesocial.com/papers/6a45ffa29ed13430313100ae https://doi.org/10.58559/ijes.1936625

View Full Paper

Ask AI

Bookmark

July 2, 2026International Journal of Energy Studies

A comparative study of CNN and vision transformer architectures for fault detection in small wind turbine blades

View Full Paper

Ask AI

Bookmark

Authors

MAMuhammet Fatih Aslan BABusra Aslan SBSelami Balcı

Discussion

Member takes

Overview

Randomized trial evaluates fault detection accuracy in wind turbine blades using CNN and vision transformer models, indicating different effectiveness.

Key Points

This research aims to compare deep learning architectures for fault detection in small wind turbine blades.
Systematic evaluation of five deep learning architectures on the CAI-SWTB dataset.
Utilized binary classification for healthy versus faulty blades using 6,000 RGB images.
Employed transfer learning, AdamW optimizer, and cosine annealing for model training.
EfficientNetV2-S achieved 99.75% accuracy, the highest among all evaluated models.
ResNet-50 followed with 99.42% accuracy, showing strong performance.
Swin-Tiny outperformed the other ViT models with 88.25% accuracy, supporting its effectiveness for fault detection.

Cite This Study

Aslan et al. (2026) studied this question.

synapsesocial.com/papers/6a45ffa29ed13430313100ae https://doi.org/10.58559/ijes.1936625

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark