Lashed by rain and hail, subjected to freezes and thaws, and tested by the sheer girth of towers and blades stretching ever skyward. Much is asked of the wind turbines expected to generate power for two decades or more.

Over time, the relentless exposure to the elements and high-stress loading conditions can damage components.

Now, researchers at Sandia National Laboratories are testing autonomous inspection technologies intended to catch cracks, erosion and other flaws before repair or replacement is needed for wind turbine blades. Researchers are sending drones outfitted with infrared cameras and crawling robots equipped with ultrasonic scanners to search for damage on the surface of, and even inside, blades.

“Wind blades are some of the largest, if not the largest, composite structures that are built in the world. They undergo enormous numbers of fatigue cycles in operation,” said Joshua Paquette, a mechanical engineer and blade reliability project lead at Sandia National Laboratories.

“Defects that maybe weren’t caught during the inspections at the manufacturing facility can grow under fatigue loading and manifest themselves as damage," Paquette told Greentech Media.

Weather, lightning strikes, damage occurring as components are transported to, or assembled in, the field: “All of those things need to be inspected on some interval — it’s not precisely defined — out in the field." 

"The difficulty with a wind turbine rotor or blade as compared to other composite structures like airplanes or helicopters is you can’t just bring the rotor down into a hangar facility or bring a blade down and look at it on the ground,” Paquette added.

Beyond the human eye

The most common methods used by wind project owners and operators to inspect wind turbine blades today are less than ideal.

Technicians often perch a camera with a high-powered telephoto lens atop a tripod at the base of a turbine and take pictures of both sides of individual blades, from root to tip, with a focus on the leading edge on the wind-facing front of the blade. The photos are searched, typically manually, with inspectors looking for cracks or missing paint or separations between the skin layers — “anything that doesn’t look right visually,” said Paquette.

The hope for project owners and operators is to avoid having to send human inspectors down on a rope or up in a basket on a crane to inspect blades up close. Such work is even more expensive and dangerous for turbines located in remote locations far offshore.

Paquette said computers can be trained using learning algorithms to sift through a batch of photos and spot damage. And some project operators have started using drones instead of human technicians to take pictures of turbine blades.

But Sandia researchers think new inspection methods they are testing that employ autonomous vehicles are even better than these incremental improvements.

In one project, researchers sent crawling robots to methodically traverse the length and width of turbine blades. The robots use an on-board camera to capture real-time, high-fidelity images of the surface and a phased array ultrasonic imaging scanner to search for subsurface damage. Sandia’s partner in the project, dolphitech, is a Norwegian ultrasound imaging company.

The other project deployed drones outfitted with infrared cameras to detect damage up to a half inch below the surface of turbine blades. The drones also employed a lidar sensor to capture super-resolution surface images of the blades.

Sandia’s drone partner, the Ann Arbor, Michigan-based firm SkySpecs, was awarded a $254,000 small business voucher from the U.S. Department of Energy for the project.

“From infrared you can see some things you don’t necessarily see from visual inspection,” said Paquette. “You might be able to see damage just below the surface that has not manifested itself in a crack yet.”

The goal, he said, is to detect damage before it triggers urgent, costly repairs. “That could be a much more low-cost repair because it’s smaller and less invasive, or it possibly allows the wind plant operator to schedule maintenance in such a way that they can have multiple turbines worked on sequentially rather than doing some sort of emergency action.”

Industry interest in autonomous inspections

Paquette said the research projects are based on the labs’ internal thinking, as well as feedback from industry.

“We have close contact with owner-operators and repair companies. We communicate with them about what they’re seeing in the field and what challenges they see,” he said.

In an email, Vestas spokesperson Chanté Condit-Pottol said the turbine maker is pursuing autonomous inspections.

“Vestas has been using versions of autonomous drones for blade inspections for the last few years, and we continue to invest in research and development for future technology solutions like robotic capabilities and inspection software,” she said.

“Development and deployment of these types of technologies are a key part of advancing [operations and maintenance] capabilities and enhancing the life and performance of wind turbine assets.”

According to Shashi Barla, a global wind supply chain and technology analyst with Wood Mackenzie Power & Renewables, “Robotic inspection on blades is gaining prominence in the industry. The costs of blade inspection are escalating primarily due to high labor costs for rope technicians, as these tasks require special skills and training to perform the inspection and repair on the blades.”

Wind manufacturers "are working on ways to reduce these costs. The drones and autonomous robots can be leveraged to conduct blade inspection at a minimal cost compared to rope technicians,” Barla said.

Sandia’s Paquette is likewise optimistic that automation can reduce inspection costs.

“Our hope with automation is to be able to bring the cost of doing the inspection down to a level where more operators are willing and able to try it and innovation can take over from there and bring the cost down even more,” he said.

Coming next is using robots not just to identify problems but to perform repairs.

Over the long term, said Paquette, instead of having a human technician rappel down a blade for repairs, a robot could be sent up to the blade to grind out composites and re-adhere repairs or patches.

“A few startup companies are moving one step ahead and are working on robotic solutions that not only inspect the blades but also perform repair activities like lead edge repair on blades,” confirmed WoodMac’s Barla.