AI-Assisted thermographic and visual classification of leading-edge erosion of wind turbine blades
- The wind industry is crucial for carbon neutrality, with turbines featuring blades over 100 meters long. Regular inspections, often manual and visual, struggle to capture subsurface damage or airflow dynamics. Leading-edge erosion, caused by rain and hail, significantly reduces turbine efficiency.
The Federal Institute for Materials Research and Testing (BAM) in Berlin is working with industry partners to classify leading-edge damage and estimate yield loss using ground-based thermographic images. These images visualize airflow disruptions caused by erosion. AI models, trained on 1500 thermographic images, can detect and classify this damage.
BAM aims to create a reference dataset by 2024, using data from 30 wind turbines. This dataset will include simultaneous thermographic and high-resolution visual images. The project also explores predicting stall and calculating performance loss due to erosion. A secure data platform facilitates data exchange and federated learning,The wind industry is crucial for carbon neutrality, with turbines featuring blades over 100 meters long. Regular inspections, often manual and visual, struggle to capture subsurface damage or airflow dynamics. Leading-edge erosion, caused by rain and hail, significantly reduces turbine efficiency.
The Federal Institute for Materials Research and Testing (BAM) in Berlin is working with industry partners to classify leading-edge damage and estimate yield loss using ground-based thermographic images. These images visualize airflow disruptions caused by erosion. AI models, trained on 1500 thermographic images, can detect and classify this damage.
BAM aims to create a reference dataset by 2024, using data from 30 wind turbines. This dataset will include simultaneous thermographic and high-resolution visual images. The project also explores predicting stall and calculating performance loss due to erosion. A secure data platform facilitates data exchange and federated learning, enhancing AI systems with diverse data.…

