Cuzner, a professor of electrical engineering at the University of Wisconsin-Milwaukee (UWM), has noticed the trend in Southeast Wisconsin, where Microsoft Corporation is building a data center that will soon be largest individual consumer of electricity in Wisconsin. As for the more garden variety of outages, Cuzner said, the future looks just as worrisome. “In the last year, I’ve kept track at my home, and I’ve been without power for at least 100 hours,” he said.
How can the U.S. fix its infrastructure and boost reliability without starting from scratch?
The answer, Cuzner said, is microgrid technology. Microgrids are power sources for a limited area, such as a military base. They integrate different kinds of energy, such as diesel generators, solar cells, wind turbines, fuel cells and battery banks, whether connected to the main grid or operating as an “island,” serving as backup power for the immediate vicinity it serves.
Cuzner views microgrids as pathways to transform the old grid into an automated modern system. Because of microgrids’ systemwide communication – the components “talk” to each other – they are much quicker at detecting defects before they lead to a blackout.
However, because microgrids are smart, they are complex, making them expensive to operate. Cuzner has pioneered an idea that would clear the way to for microgrids to thrive commercially. He proposes breaking them down into “building blocks,” or smaller units of microgrid components, called nanogrids.
The background on microgrids
One reason microgrids aren’t widely used yet involves equipment compatibility, Cuzner said.
“You’re trying to merge the old infrastructure with the new equipment of the microgrid, where no uniform standards exist,” Cuzner said.
Finding the equipment needed to integrate renewables is one example, said Mark Vygoder, a doctoral student and longtime lab member. Cuzner’s lab members have been working with large U.S. military bases in Europe that already use microgrids to address grid insecurity but are grappling with costs related to knitting together unstandardized equipment.
“It’s a bit like the Wild West where you can buy devices from different vendors and all the products are a little bit different,” Vygoder said. “So, it becomes quite costly when you have to hire a service provider to sort that out for you. When the microgrid operates independent of the grid, all those different components need to coordinate and communicate.”
Cuzner and his lab members are working with researchers Giovanna Oriti, Douglas Van Bossuyt, Ron Giachetti and Dan Nussbaum at the Naval Postgraduate School in California with support from a series of grants from the U.S. Naval Facilities Engineering and Expeditionary Warfare Center (NAVFAC EXWC) to simplify the architecture, employ artificial intelligence, and bring down the cost of microgrids.
UWM expertise in power distribution
Cuzner, an expert in power controls, conversion and distribution – the areas of vulnerability in microgrid technology – stands at the center of UWM’s reputation as a leader in both energy storage and electric grid technology.
Cuzner’s lab is a lead partner in the GRid-connected Advanced Power Electronic Systems (GRAPES), a national industry/university research center that aims to accelerate insertion of power electronics into the national grid.
Cuzner, Vygoder and Andrew Eggebeen, a recent PhD graduate in computer engineering, who worked in Cuzner’s lab, visited three U.S. bases in Europe in the summer of 2023 to get a first-hand look at how these microgrids are being implemented and their limitations.
“One thing we found was that these bases are very large and spread out, leading to transmission problems,” Cuzner said. “In one case, the solar array is several miles outside of the base.”
To solve the problem, Cuzner and his colleagues at the Naval Post Graduate School developed a “zonal distribution concept” – essentially breaking microgrids down into less complex units, called nanogrids.
What is a nanogrid?
Cuzner’s background is in the conversion of Navy shipboard power generation to electrical distribution. Such architecture features damage-control zones: When the power goes out in one part of the ship, the system reroutes itself using smart switchgear and continues to operate with only the affected zone shut down.
Nanogrids can be strung together within the microgrid itself, improving overall smart capabilities. And they can be added one at a time, easing the cost burden of a microgrid.
Cuzner and his team are researching the best ways to standardize components supplied by commercial vendors and ensure “grid-edge inter-compatibility,” which means that even components supplied by different vendors can work and play well together.
“If smart components of a nanogrid are standardized,” Cuzner said, “it can become a ‘plug and play’ building block that can be produced cost-efficiently.”
UWM’s microgrid ‘sandbox’
To work out the details of nanogrids, Cuzner’s lab members have built a fully functional microgrid at the University Services & Research building near UWM’s Kenwood campus. Since 2021, the lab has been building an energy distribution system with smart metering and controls, giving them an experimental sandbox.
The researchers can now observe how a microgrid responds under varying conditions, quantify how commercial components monitor microgrid data, and then simulate in real-time a full-scale system that interacts with real control hardware.
“With our microgrid, we can simulate equipment that is on the grid, test it at scale, quantify the ‘grid-edge’ where everything comes together, and figure out how to improve performance,” Cuzner said. “That’s something no one else has done yet.”
In 2023, they worked with a local company, Badger Technologies, to install, test and integrate a battery energy storage system with the UWM microgrid. In addition, UWM’s microgrid includes a solar array, one wind turbine, two natural gas generators and a smart switch that could connect it to the national grid.
The college is currently exploring ways to secure federal funding to turn the UWM microgrid into an industry-collaborative lab with a 1,000-13,000-volt testing facility. Such a facility would attract industry, quicken the pace of new microgrid technology and would include research on electric ships and aircraft.
Power, energy and controls in focus
Nanogrids also improve control of the flow of electricity if connected to the grid.
Control refers to how the existing grid meets demand. Higher than normal demand for electricity could cause a blackout, but so could a glut of power to the grid from renewables.
Another benefit of nanogrids is that the controls can be built from the “bottom up,” Vygoder said, giving nanogrids the ability to speedily compensate for power disturbances.
Southeast Wisconsin is the perfect place to develop this commercial potential with its cluster of companies related to energy, power and controls. UWM is at the epicenter, with longstanding research partnerships with industry leaders including Eaton, Rockwell Automation, Leonardo DRS, and Johnson Controls.
Most recently, Office of Naval Research has funded Cuzner to develop a “digital twin” approach to study how nanogrid components respond to a wide range of both normal and damaged scenarios.
AI and nanogrid decision-making
Digital twins rely on artificial intelligence to improve communication among the smart components of both micro and nanogrids. It’s the next step in the integration research.
When paired with machine learning, AI could potentially allow microgrids and nanogrids to teach themselves what leads to a breakdown and autonomously decide what to do when they detect trouble brewing.
For this research, Cuzner has consulted with Zhen Zeng, UWM assistant professor of computer science, who is an expert in digital twins and cybersecurity. Zeng is co-advising computer engineering students who are helping in Cuzner’s lab, bringing together power/energy and computer engineering in the college.
“When we feed a lot of information into an AI model, the model can quickly tell you what is going on in the system,” said Zeng. “We try to understand which situations we would need to consider when building cyber-protection into the design.”