Wind farms are becoming a key part of the energy landscape in many different ways. Both on-shore and off-shore, wind farms are providing more and more renewable energy, particularly for data centres that are handling petabytes of data every day.
The ownership and optimisation of these wind farms is also changing. For example, Apple has invested in a turbine in Denmark for wind power to a data centre there, and wind farms are regularly changing hands between owners.
Machine learning is already used for monitoring the performance of turbines and blades, providing preventive maintenance and reducing the number of times people have to climb up a tower to check on a turbine blade. But AI has distinct limits in the optimisation of wind farms, says Blair Heavey, CEO of energy consultancy Windesco in Boston, Mass.
“Outside of some unique spaces, AI hasn’t delivered,” said Heavey. “In some areas such as battery life there are standards where the data can be fed into the machine learning models and can provide almost human-like responses. What we found in the wind space is there isn’t anything that’s a standard and there’s no consistency across the data patterns that come in – different turbines, gearboxes, even different climate models, locations with a ton of different variables.”
Windesco is a consultancy in software analytics, helping wind power OEMs and farm operators develop more value from the assets. The human experience is essential for these highly complex systems, says Heavey.
“Our teams has been working on the algorithms for six years that we believe can tell the human expert what we should do but it requires an expert to oversee what the AI and the models are driving us towards,” he said.
“Because we don’t have all the standards, we could interpret that data correctly for one turbine and adjust the pitch and blades but that would damage a turbine at another locations. We always believe in AI and physics based models, but before we implement these models we oversee the unique data of that particular turbine in that particular farm,” he said.
A digital twin is one approach that is being considered by some operators to handle the complexity of a wind farm, but it has limitations, says Heavey. Siemens Energy is a major OEM of wind turbines, and digital twin technology is a strong technology capability for Siemens.
“It’s difficult to put these things together,” he said. “Some of the additional modelling you will see us bring out to the market next year is wake steering, that’s something that we are modelling and have proof of concept algorithms put together. The added complexity of those models is not only the blade turbine and gearbox, but in a single area you have turbines communicating with each other and that can have downstream turbines do a bunch of different things.”
This can extend the life of the turbines, avoiding running them when the wind is too strong or out of range. “We already use fluid dynamics, physics and a broader ecosystem of academics to challenge, test and implement algorothms on smaller wind farms to make sure we do no harm,” said Heavy. “Our customers are increasingly comfortable with the AI and algorithms that are running and sharing them back with us to improve the algorithms. The trust factor that customers have is applying the human oversight to ensure that the system works and delivers what is anticipated.”
This helps to build the models that help to improve the performance and efficiency of wind farms.
“Customers are sharing more data with us long term, constantly working to improve that efficiency as long as it is anonymised to help the overall industry. I think we are still educating people on this,” he said.
The problem is there is no standard for this data. “There is no standardisation across that today,” he said. “We do work with operators and OEMs to try to drive that standardisation, and we continue to push for that.
The support comes from perhaps an unexpected direction – private equity funds. These are buying up wind farms to provide energy to suppliers and to large data centre operators. They have a vested interest in boosting the performance and reliability of such farms.
“The most leverage is the private equity investors and it is in their best interest to help them drive those things to increase the revenue production and efficiency,” said Heavey.
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.