For generation and transmission cooperatives, timely maintenance is critical to safety and reliability. That’s where an artificial intelligence tool built by Dairyland Power Cooperative can be a hero.

Built on computer-entered data and machine learning, VoltWrite is the La Crosse, Wisconsin-based G&T’s secure in-house AI platform capable of generating data, text, images and even video specific to Dairyland’s needs within minutes.

The tool had a notable success during scheduled maintenance at the co-op’s 1979 John P. Madgett Station.

The co-op had recently digitized some 2 million pages of printed policy and standards documents and fed them into VoltWrite. When an operator came across an unknown piece of equipment at the power plant, he was able to upload a photo of the component into VoltWrite with the prompt: “What is this?” The AI not only instantly identified the part but noted that it was in poor condition and needed replacing.

“This is a whole new level to interact and to support plant staff in those types of maintenance windows,” says Nate Melby, Dairyland’s vice president and chief information officer.

Cybersecurity and Reliability

The VoltWrite project started last year as a conversation among Melby’s team about the rapid rise of AI and how to harness its benefits and avoid its pitfalls, particularly as the G&T’s grid generation uses increasing amounts of data from things like smart devices, behind-the-meter energy resources, renewables and storage.

Dairyland tapped into its partnership with Microsoft and gained ungated access—a privilege shared by only 4% of companies worldwide at that time—to some of the tech giant’s AI software, applications and platforms.

The team prioritized cybersecurity, reliability and the electric power industry’s critical infrastructure in building VoltWrite.

“If we look at our ultimate mission, which is safe, reliable, cost-effective electricity, we want to make sure that that we’re holding true to that in every way that we can so we’re protecting the data privacy and we’re protecting the resilience of our grid,” Melby says.

“But we’re also trying to create more efficiency so that our employees and our members’ employees can have the advantages of using this new superhighway of information, leverage it effectively and be as productive as possible.”

Because it was built uniquely in-house, VoltWrite provides the benefits of AI without the cyber risks associated with publicly available open models like ChatGPT.

“In a space where there’s a lot of concern about privacy of data and the critical infrastructure that we provide, we were able to take the benefits of AI and make a private environment that leverages language models,” says Melby, who has a Ph.D. in information systems. “Cybersecurity was top of mind for my team when we were developing this.”

'Like a Wildfire'

After successful pilot tests in 2023, VoltWrite was introduced in February across the G&T. Following a couple more pilots this year, Dairyland opened it to all 24 of its member co-ops in May.

“It’s like a wildfire,” says Melby. “Now they’re coming up with ways to use it that we hadn’t thought of. And they’re getting benefits from it, too.”

Topping the list is new “augmented worker strategies” to help staff accelerate tasks.

In one instance, a co-op SCADA engineer used VoltWrite to process a month’s worth of backlogged system data in less than three days.

“SCADA engineers are really hard to find in terms of their talent and recruitment. Their time is very valuable to us,” says Melby. “So all of a sudden we’ve got someone who’s that much more productive.”

Vladimir Tsoy, Dairyland’s supervisor of business systems applications, said seeing plant staff wield the tool to improve operations was not foreseen during VoltWrite’s development.

“When we talk about winning back time … imagine looking for a specific piece of information in five-inch-thick binders,” Tsoy says. “There’s really a lot of enthusiasm and adoption at the plant.”

Safety is also improving, according to Melby and Tsoy. Tailgates talks, safety briefings and incident reports at Dairyland facilities produce reams of information that can now be scanned into VoltWrite to identify safety patterns and conduct incident analyses.

“Making this connection by leveraging AI, we’re able to get to that next level of analysis,” says Melby.

The team is now looking at using AI to allow engineers to guide technicians remotely through site inspections.

“Engineers can be in multiple places on the same day instead of taking a four-hour trip one way to one plant and a four-hour trip back,” Melby says.

They’ve also begun testing how VoltWrite can ingest and analyze Dairyland’s volumes of rules and standards and identify any conflicts or redundancies.

“As a G&T, we have a lot of policies,” says Melby. “If you think about the number of hours it would take a person to do that, especially in a company as large as ours, I mean, that’s a winning use case right there.”

And Dairyland’s human resources team is feeding data on workforce demographics and existing talent, talent management and talent acquisition into the AI for insights on talent gaps as generations of employees retire. Analyses also can be run on what skillsets need to be updated or acquired.

Katie Thomson, Dairyland’s manager of strategic communications, says she was recently pressed to develop a survey that “was a little complicated.” VoltWrite helped her organize the questions and make the content more comprehensive.

“Honestly, it probably cut out two hours of my time,” she says.

As Dairyland readies VoltWrite 2.0, the G&T is eager to share the value of the AI tool with other co-ops.

“The notion is that instead of thinking about AI as the thing that replaces people, we’re thinking about it as the thing that makes people faster, more productive,” Melby says. “That benefit alone lets us do more with the resources we have, and it helps everyone.”