UND Study to Automate Wind Turbine Inspections

150425 UND Study to Automate Wind Turbine Inspections picture

 

Rather than hanging from a rope 80 meters in the air, wind turbine technicians may use drones to inspect the blades.

University of North Dakota’s Technology Accelerator research center is conducting a two-year study to develop unmanned aircraft technology to automate inspections for the wind power industry and others, according to director Kevan Rusk.

Rusk, who worked for Grand Forks-based turbine blade manufacturer LM Wind Power for seven years, says he is aware of the challenges the industry faces.

Because of the time and money it takes to do inspections, a large percentage of turbines are not inspected on as regular a basis as the manufacturer suggests. With rain and dirt and other elements striking the spinning blades, whose tip speeds can reach 100 miles per hour, the leading edge of the blade can be eroded within a few years.

“If the shape of the blade is not maintained, it doesn’t make as much power,” Rusk said.

A drone could conduct the inspections more often, and increased inspections could help extend the life of blades. Companies could benefit financially by optimizing the amount of power produced by the turbines and the safety risk for the technician is minimized.

“They’ll have an opportunity to do inspections faster and more economically. It’s an added tool for the toolbox,” said Rusk, pointing out, that at it’s very simplest, the project would duplicate images already being taken manually. “If the project did nothing more than that, you would have a huge win.”

Beyond reducing inspection time, a large part of the study will be dedicated to data analysis, Rusk said.

“The UAS is just a tool to gather data,” he said. “The bigger value is to be able to automatically process the data.”

For example, a 100-turbine wind farm could leave a technician with six terabytes of video to review. With a data analysis tool, a technician could use that to analyze the video and hand the company a short report listing which blades have defects in need of repair.

EdgeData, a Grand Forks-based data analysis company, matched the $450,000 Research North Dakota grant, bringing the grant total to $900,000. EdgeData is partnering with LM Wind Power on this research project to develop a data analysis tool for the images gathered by the drones. LM Wind Power will be the customer for the product when it is commercially deployed.

“This technology is currently being offered at some maintenance services companies, and we believe this is very promising,” said Kevin Tschosik, manager of distributed generation for Basin Electric Power Cooperative, which has several wind farms in North Dakota. “We have discussed trying this at Basin Electric’s wind projects, as this technology may provide a greater detail of the condition of the blades in a lower risk environment.”

Tschosik said the cooperative is worried that the technology can only be used on very low-wind days, which could be challenging as wind projects are typically built in known high-wind areas and could lengthen the time to complete blade inspections.

“However, we still believe that using unmanned aircraft systems for blade inspections will provide a better inspection result,” he said.

The study will be performed in a two-story laboratory. The Tech Accelerator will bring in large portions of blades and do test flying indoors. The lab will have the ability to mimic different elements, including sunlight and dust.

If the data analysis portion is successful, the program will be able to take the drone’s imagery and tell the difference between dirt and other buildup versus erosion or cracks, Rusk said. After the data analysis technology is proven for wind, it can be used by other types of companies.

“You can almost name the industry that processing could be applied to,” he said, including pipelines, precision agriculture, salt water disposal wells in the oilfield and power lines.

Photo: Tom Stromme

About this article

Add Comment

Your email address will not be published. Required fields are marked *

nine − four =

error: Alert: Content is protected !!