In 2024, more than one in five cars sold is an electric vehicle (EV). Intergovernmental agencies estimate that by 2035, half of all new cars sold globally will be EVs.
While more EVs on the road sounds like great news for the environment, it could lead to complications. The electric grid is not yet ready to support the EV influx, and unaddressed capacity limitations could threaten the future of the EV industry.
Researchers at the Georgia Institute of Technology have developed a device to help avoid grid overload: a revolutionary EV smart-charging system.
Their optimization-based approach would not only help reduce grid stress but also enable users to customize the charging process to minimize cost, prioritize the use of carbon-free energy, or adjust the charging speed. The technology might even help promote the EV industry.
The EV revolution is happening so quickly that electrical utilities will find it difficult and costly to update the grid fast enough to accommodate these vehicles,” said Professor Michael J. Leamy. “We wanted to investigate a way of smart charging that considers EVs as they converge on the electric grid, which we already know is near its maximum capacity. We need to buy time for the grid operators, so they aren’t overwhelmed when more EVs enter the market.”
In the southeastern U.S., during the summertime, power use peaks from late afternoon to early evening. By then, the heat of the sun has warmed up homes and buildings, and people have cranked up their air conditioners. This is also when most people come home from work and plug in their EVs, which immediately start charging. When plugged in, an EV can easily be the most significant consumer of power in a home.
A bunch of EVs charging during the peak power load can be problematic for utilities, exacerbating electricity demand and overloading the grid,” said Dr. Kartik Sastry, a co-inventor of the technology. “But you don’t need to start charging as soon as you come home. In most cases, there’s ample time overnight to fully charge electric vehicles.”
The team set out to address the issue of grid overload by taking advantage of overnight surplus charging time. They developed cutting-edge optimization algorithms to distribute charging over time, thereby minimizing grid stresses while offering benefits to consumers.
Rather than charging at full speed as soon as an EV is plugged in, the team’s smart-charging algorithms allow for gradual charging, or charging at several intervals over time rather than all at once. The algorithms also decrease the maximum power draw of a home during charging, achieving what the researchers call “peak shaving.”
The algorithm itself plans the charging and when it will occur, based on the state of the grid and consumer preferences. The system makes these decisions without the need for any centralized control, and without communicating with other vehicles in the neighborhood or city.
“By using random timing and predictions of household power usage, our algorithm can effectively distribute the load,” said Professor David G. Taylor, a co-inventor of the charging system. “This means that the number of EVs in a community can increase without the local utility incurring huge expenses for infrastructure upgrades, provided these EVs use the smart charging algorithm we've developed.”
Using an app the team designed, a consumer can control four different facets of charging: cost, carbon-free energy, speed, and battery health. The app’s slider widgets allow a user to weigh their preferences for each; based on the user’s selections, the algorithm determines how best to charge the vehicle while minimizing grid impact.
For example, to charge their vehicle quickly, a user would move the charging-speed slider all the way to the right. Prioritizing carbon-free charging, on the other hand, programs the charger to only charge the car when renewable resources such as wind or solar are available on the grid. Choosing to charge at the lowest cost would prompt the system to charge at low-demand times based on a utility’s price schedule, allowing the user to avoid any demand charges.
The team has published several research articles about the project in major journals. They have also built prototype hardware, completed charging case studies with individual EVs, and submitted a patent for the technology.
Expanding beyond the system’s applications for individual consumers, the team has also developed this technology to function in commercial settings. They hope to collaborate with an EV charging station company and a public school system that uses EV buses to demonstrate how their algorithms automatically reduce fleet charging costs.
The goal, Leamy says, is for companies to license their smart-charging algorithm technology.
The researchers are also reaching out to auto companies to have their technology incorporated into EVs themselves, which would allow users to set their charging preferences right from their dashboard.
“We hope to demonstrate that smart charging is feasible on a large scale — and also measure the benefits of doing so,” Leamy said. “We plan to show how much money consumers, companies, and school systems can save with smart charging rather than simply charging in an ad hoc fashion.”