A year and a half ago, Siemens Gamesa Renewable Energy transformed the inspection of wind turbine blades replacing people, who were doing a risky activity, by autonomous drones and a digital solution called Hermes. Through this proposal, the aerial device captures high-resolution images quickly, while the solution analyzes the photos to detect possible damage to the blade, a combination that results in safer, faster and more accurate inspections.
The company, headquartered in Spain, is further improving this project with the migration to Microsoft Azure and the incorporation of Azure AI to process image recognition. These digital advances will allow Siemens Gamesa to further speed up blade inspections.
“Artificial Intelligence, the cloud and big data allow us to make a quantum leap in terms of innovation”
“Hermes has taken a big step forward with the collaboration with Microsoft,” said Christian Sonderstrup the director of Digital Services at Gamesa a company that has installed wind energy technologies in 90 countries. “Artificial Intelligence (AI), the cloud and big data allow us to make a quantum leap in terms of innovation and reduction of the levelized cost of renewable energies.” The levelized cost of energy refers to the lifetime cost of an asset divided by the amount of electricity produced.
The drones, which will inspect 1,700 turbines this year, are fast, accurate photographers who capture about 400 images of the three turbine blades in 20 minutes. The images have the capacity to form an overview of the state of the blades and the necessary repairs, but the need to classify them and put them together manually has been a challenge. The burdensome task recently became apparent in a large inspection project with 100,000 photographs.
By using the recognition of images, the Azure AI services can put together photos of an entire rotor in a precise model in 34 seconds
“We had a person dedicated to analyze each of the photos taken, and then every severe failure needed to be assessed again by an engineer,” said Anne Katrine Karner-Gotfredsen, Manager of Product Integrity and Guarantee Management of Siemens Gamesa in the company’s blade program.
The integration of the Azure AI services will greatly accelerate the process, thanks to the recognition of images, which can put together photos of an entire rotor in a precise model in 34 seconds. The same work done manually would take four to six hours and could lead to errors.
AI tools can differentiate the blades from water, sky and other irrelevant elements; distinguish cracks and defects of, for example, bird droppings. They can also integrate the location of the drones and the zoom data of the camera for a precise union, as well as classify faults by type and severity.
Faster, more accurate inspections mean less turbine downtime, earlier fault detection, and better predictive maintenance
“Reviewing all the photos is a huge task,” said Karner-Gotfredsen. “Before Hermes, it was very complicated to categorize and store all the data in a place that we could all access,” Karner-Gotfredsen said, adding “the more we manage to automate the process, the easier it will be to work with the data”. Faster, more accurate inspections mean less downtime for turbines, earlier detection of faults, better predictive maintenance and less costly repairs, which contributes to more affordable wind power.
For Karner-Gotfredsen, the cloud will also help to optimize projects: “the fact that we can now have the data sorted and consolidated automatically in Hermes through the cloud, saves us many hours of work by not having to manage hard drives”, says Karner-Gotfredsen. “Artificial Intelligence is increasing the productivity of our employees, allowing them to concentrate on their core competencies,” she concludes.
“We want to be the digital leader in renewable energy,” says Sonderstrup. “Artificial Intelligence, cloud and Big Data will be the facilitators of that journey.”
