A much smaller study by the University of Michigan Transportation Research Institute (UMTRI) and startup Utilidata suggests that new AI tools can give utilities real-time data to make the power grid and EV charging more reliable.
Researchers are using AI to analyze EV charging behavior, hoping the insights can improve the experience for drivers and help utilities prepare for increased electricity demand. So far, they've found that EV charging can draw power erratically and degrade power quality, leading to wear and tear on charging equipment.
Those fundamental problems waste energy and can lead to exploding EV chargers that have become a bane for drivers. So the ability to quickly detect and predict these issues with AI can be a game changer. AI models can guide utilities on how charging might affect the power grid, the authors write. And they can also advise drivers on where and when to charge and help EV charging companies better maintain their equipment.
Those fundamental problems waste energy and can cause EV chargers to explode.
UMTRI initially approached Utilidata for this pilot study, intended to inform the design of a larger research project investigating the same issues. UMTRI says it is already working with the North American Electric Reliability Council to address their preliminary findings.
For the study, researchers installed electric meter adapters in conjunction with Utilidata's AI platform Kerman at six EV charging stations at the University of Michigan. Kerman analyzed voltage, current, power and other dynamics between March and June last year. The study authors also installed devices on the vehicles of 10 drivers who frequent college campuses to monitor their charging habits.
Although the project is still in its early stages, researchers hope it can help prepare people for the challenges that come with electrifying a fleet of vehicles. In the U.S., aging power grids are already strained to meet increasing demand for electricity from AI data centers, crypto mining, and clean energy technologies. But compared to a data center, utilities have a harder time predicting when and where EVs will plug into the grid.
Utilities have to cope with this unpredictability without real-time data to help them adjust. Those blind spots are becoming a bigger problem at the “grid edge,” where consumers are increasingly connecting their devices to the grid, such as batteries for EVs and solar panels.
“There is a big role for AI at the edge of the grid,” says Siobhan Powell, a postdoctoral researcher at ETH Zurich who was not involved in the research. “It didn't used to be like that, right? There wasn't much interesting going on and now that we have control, there's more opportunity and more value to find out what's going on.
“There's a Big Role for AI at the Edge of the Grid”
One problem the researchers noticed in the study was short-cycling, the inconsistent draw of electricity from vehicles that would stop and start charging even after the battery was completely drained. Not only does it burn through energy inefficiently, it can also overheat wires and transformers. They also found that EV charging degrades power quality when power deviates from ideal voltage and frequency ranges. Flickering is a clear sign of poor power quality, which can also cause more wear and tear on equipment.
“I think the biggest benefit is that we've confirmed that there are a lot of behaviors of electric vehicles that nobody knows about — car owners don't know, grid operators don't know, charger OEMs don't know. “, says Yingchen Zhang, Utilidata's vice president of product solutions. “So there's really a great need to open up all that data.”
The study's authors cautiously make the case that locations with highly unregulated EV charging could see large impacts on the power grid. In a worst-case scenario, they say it could affect power supply to other customers. But Zhang was quick to say that the likelihood of a power outage as a result is very low.
“It's good to know how these new charges affect voltage, power quality issues locally, but I'm not going to go into outages,” Powell says, because to prevent outages. Utilities can take a number of steps. And again, this is a very small study of unpredictable charging behavior, so it's too early to make clear statements about grid-wide implications from these preliminary results.
Both Powell and Zhang want to avoid creating unnecessary alarm about the impact of EV charging on the grid — especially when EV adoption faces partisan attacks. “A lot of the fear is because people don't know the actual EV behavior,” Zhang says. “So actually revealing that information will reduce a lot of those fears.”
The rise of AI has also raised concerns about energy-hungry data centers that stress the grid. Zhang says his company is thinking about that, too, using custom-designed chips from Nvidia to use less power than more common AI chips. And using machine learning to analyze data in this way is typically much less energy-intensive than creative AI models that spit out text and images.
It comes down to preparing as the key to scaling the power grid against new technologies that change the way we live, work and move around. Fleets of EV batteries can also help strengthen the grid by acting as virtual power plants that feed electricity into the grid when needed. Automakers are already testing it, in part to make EVs more affordable for consumers. “We need EVs. We need that transition. And there are things we have to do to develop the grid, but we can do that,” Powell says.