Vermont Electric Cooperative, which serves a small, dense region in the northern half of the Green Mountain State, has been working with grid management platform Camus Energy since early last year to fine-tune visibility into its present and future grid.
The pair recently concluded what the utility’s innovation and technology leader Cyril Brunner described as the “painful” process of integrating internal and external data sets — like advanced metering, electric vehicle charging, and real-time control room data — into an integrated platform. That platform will now allow them to leverage more sophisticated artificial intelligence and machine learning to optimize VEC’s asset management.
- The top line: VEC is partnering with Camus to dig deep into the state’s EV charging trends and projections, running a detailed analysis that will ultimately help the utility understand just what the growth of electric transport will mean for the grid.
- The market grounding: Despite some EV industry growing pains, the additional load of plug-in cars is still expected to be immense, up to nearly 2,000% growth by 2040, by some accounts. In VEC’s territory, that growth is heavily segmented by region — just ten of its 72 feeders are experiencing the bulk of growth in both solar installations and EVs.
- The current take: According to Camus founder and CEO Astrid Atkinson, the impact of the growth of EV adoption is expected to show up roughly two to three years from now. “The utilities I’ve seen that have run fine-grained analysis on that have found it to be really informative as to where they might see challenges,” she said, adding that more serious grid challenges are “on the horizon” according to the models.
When it comes to pilots, VEC is doubling down on VPPs, running anywhere from ten to 15 “little tiny experiments” with different vendors at a time, Brunner explained. VEC’s electric vehicle pilot with Camus plays directly into the utility’s goal of maximizing its visibility into the future of the grid, which Brunner hopes will ultimately save millions of dollars otherwise spent on building for new peaks.
Pilot planning
The program, which is still in its early stages, integrates state of charge data from telematics provider FlexCharging, which provides access to information from a vehicle, rather than from its charger.
That means VEC will have visibility into things like how much charge a vehicle needs on the average day, when it needs to be charged, and how long it takes to fill up, he said, down to a specific model.
To start, the utility is targeting Tesla, Ford, and Kia owners — those are the automakers that have decided to give access to telematics data for free. Brunner said VEC has around 550 EVs and plug-in hybrids on its grid today, largely Teslas. Of those, he’s hoping the pilot will reach 100 enrolled EV owners by the end of the year.
VEC is ultimately looking to leverage those vehicles for things like peak management and energy arbitrage, Brunner added, and needs the ability to schedule different chargers at different times of day. But as more EVs come online, doing so becomes increasingly complex, which is why they’re looking to automate.
Broader visibility into how much flexibility exists, down to a substation level, will enable VEC to build “much more sophisticated” machine learning models on charging behavior, Atkinson explained. Ultimately, that will help the utility better manage those EV assets, she said.
Modeling future load
The pair are set to begin working on the control aspects of this integration in the second quarter of this year — whether that’s leveraging EVs for peak shaving or for distribution services — and then expand into a more automated program.
EV charging loads tend to range between six and 18 kilowatts of instantaneous usage, Atkinson said. That’s quite a large load compared to a home’s usual demand, she added. While a single EV’s load is not a big deal for the grid, once there are four or five EVs behind a single transformer, or a hundred on one circuit, it can start to create problems.
The ability to manage EVs using AI will ultimately save VEC money it would otherwise have invested on building additional capacity to manage the potential peak created by EV chargers, Brunner said.
That flexibility ensures the utility won’t create a new peak when EV owners move to charge in off-peak hours, Atkinson said, and will allow it to spread charge load out in a way that doesn’t negatively impact customers.
“We can automatically give everyone a schedule on a day-to-day basis,” she said. That schedule is based on a model that figures out timing that works for everyone, based on charge data unique to their car.
However, it’s still something of a long game. Atkinson said that with these types of automations, operators typically like to “push the buttons” for at least six months before turning the tech loose.
“Building those models will take time,” she added. “Building out automation will take time, particularly building the operational confidence that you could trust the model to protect the grid really takes some time and work, but that’s where we’re at.”
Brunner, who sits in a unique role designed to conduct tech experiments that could someday be used by the wider utility, said there are still “mixed feelings” about automation on VEC’s small team, but that there’s recognition of the value a flexible EV program could bring.
The traditional utility approach to load growth would be for VEC to build twice as much capacity as needed, because until now there haven’t been many viable alternatives to managing peaks, he said. But that’s no longer the case.
”These technologies have evolved in a way that’s now enabled this kind of creative idea,” he said.
