Extreme weather


DocuFly – Autonomous inspection of offshore wind turbines for defect identification

DocuFly is a drone company concept with the idea of specializing in machine learning image technology.

Long-term goal – Autonomous release of drones.

DocuFly utilizes drones to be deployed from the nacelle or another location of the wind turbine where it is protected from weather and charged between missions. When the weather conditions are favorable it autonomously deploys and scans the turbine blades for defects.

After returning, the pictures are uploaded to the cloud. This requires no on-site personnel and provides end-users with the ability to inspect the offshore turbines easily and safely, reducing labor costs associated with maintaining the offshore wind farms.

Figure 1: Even a small, light and affordable drone like this one from DJI can capture images of sufficient quality

Intelligent image analysis – Track the degradation of offshore wind farms at scale.

Machine learning allows us to automatically detect and classify potential damage on the offshore wind turbine blades. Cutting-edge techniques are used to train a neural network to detect and classify faults in images of wind turbine blades. Any detected issues can quickly be verified by a human, who also decides whether further action is required.

The image data and any corrections made by the human expert are used to train the neural network to further improve its accuracy. Data from different wind farms across the industry can be combined to provide better predictions to all customers.

Figure 2: Example of image crack detection and classification describing the concept.


Leave a Reply