Zero-shot learning for wind turbine blade defect detection via symptom description transfer

Key Points

• Effective zero-shot learning allows for accurate defect detection without specific training data for each defect type.
• The proposed method achieves detection accuracy of over 85% in identifying various blade defects.
• Analysis using symptom description transfer highlights the potential of machine learning in industrial applications.
• This approach may enable cost-effective inspections, reducing downtime for wind energy systems.