Do utilities need edge computing to get ready for the EV influx?

More electric vehicles on the grid means yet another load growth challenge for the grid: how to manage their charging. And as the issue becomes an increasingly urgent one for utilities, the industry has a decision to make about the best tools for the job.

Edge computing — putting an AI chip directly into electric meters to get visibility down to the customer level — is one answer. But while that solution promises to offer improved reliability and real-time data processing, it also poses a costly overhaul of how most utilities currently operate.

• The top line: In a small study released in late April, the AI grid edge startup Utilidata highlighted some potential challenges with electric vehicle charging including “unexpected rapid swings” in current draw and lower power quality in local grids. But while those are challenges the utility sector will need to manage, exactly how it will do so is still up for debate.
• The current take: David Groarke, managing director at Indigo Advisory Group, told Latitude Media that the value proposition for edge computing to help manage the rise of EVs isn’t yet as clear as it may become. But as the cost of edge computing comes down and the number of pilot-scale deployments go up, he added, “the value could become apparent.”

There’s no doubt that managed charging programs will be “essential” as more EVs hit the grid, Groarke said, and the startups and utilities that are already experimenting with the potential for edge computing to offer deeper insights into grid operations getting ahead of any problems. 

By anticipating challenges at scale and testing edge computing solutions, he added, “the power sector is demonstrating digital maturity and preparedness to meet the customer, reliability, and infrastructure challenges of an increasingly complex network.”

But what’s less clear is whether edge computing specifically, especially an AI chip in a meter, is needed.

The findings

Utilidata’s study with the University of Michigan Transportation Research Institute tracked data from AI electric meter adapters at six charging stations on the university campus. It found that the type of data those meters can provide may be key to designing programs that “fully capture the potential benefits of managing EVs.”

And while the study’s scope was narrow, its granular findings are significant for utilities, Utilidata’s VP of product solutions Yinchen Zhang told Latitude Media. 

“While it is expected that EVs will impact power quality, the specifics are unknown,” he said. “Most programs focus on using EVs as a resource to serve system-wide generation and capacity needs while failing to capture the additional value of managing local distribution grid constraints.”

That’s a sentiment echoed by Suncheth Bhat, chief business officer at commercial charging developer EV Realty. There’s still a lot to be figured out about charging at scale for EVs to evolve to the kind of grid asset offered in something like vehicle-to-grid charging, said Bhat, who previously led PG&E’s clean energy transport team.

The potential for more granular data to optimize charging and therefore allow a utility to circumvent circuit upgrades is intriguing, Bhat added, but edge computing is different from what the utility sector is used to.

“This is just sort of a dramatic kind of change that's going to have to happen,” he said. “The problem is going to have to be solved with a different approach than it has historically.”

A utility’s alternatives

Cyril Brunner, who leads Vermont Electric Cooperative’s innovation and technology group, said that several of the power quality issues raised in the report — like current harmonics and variability in local voltage — are news to VEC. 

That said, service-side (or behind-the-meter) issues like harmonics and variability are beyond VEC’s purview, Brunner said, adding that using solutions like member-owned batteries or device configurations is ultimately the customer’s responsibility.

VEC is already anticipating voltage issues at primary lines and even substation transformer or transmission line overloads. 

“These issues are caused by the [kilowatt] output of the EV charger at any one given time,” Brunner said. The cooperative is utilizing telematics from FlexCharging and AI-backed grid planning services from Camus to get the visibility necessary to eventually conduct peak management and energy arbitrage.

“We believe we can mitigate these upstream infrastructure concerns through member programs,” Brunner added. “To do so we are working with partners to develop software tools that understand near real-time grid conditions and adjust charging speed to mitigate infrastructure impacts.”