This week I attended the 50th Allerton Conference on Communication, Control, and Computing. Although the conference has not historically focused on electric power, the leadership of Professor Alejandro Dominguez-Garcia of the University of Illinois at Urbana-Champaign has resulted in a well-attended track on electric power and electricity markets.
I presented research co-authored by UT PhD student Mahdi Kefayati and me about strategies to avoid peak load being exacerbated by electric vehicle charging. (Kefayati’s research is funded by UT’s Center for Electric Vehicles.)
Kefayati’s research asks how we can avoid overburdening the grid with rechargeable cars. It’s part of a larger research agenda to facilitate the integration of renewables into the grid.
In our paper, we assess three policies for charging electric vehicles in typical circumstances where charging takes place over many hours at home or at work. (These policies are all “local” in that they do not require any communication with the grid.)
Policy one, or “immediate charging”: charging at the maximum rate immediately when plugged in until the battery is fully charged. Policy two (so-called “grid friendly”): charging at the maximum rate starting just enough hours prior to the car’s next use to charge it fully by the time of the next use. Policy three (“average rate charging,” developed by Kefayati): charging at a uniform rate from plug-in to departure.
Immediate charging is really the default case, just like plugging in any typical rechargeable battery. Unfortunately, we find that immediate charging tends to result in cars synchronizing their charging times, thus increasing peak loads on the grid.
Although the second policy is called “grid friendly,” we found that it can actually result in even higher loads on the grid than immediate charging.
In contrast, average rate charging tends to smooth out the charging profile over time, lessening peak loads on the grid and even matching times when demand is lowest on the grid.
Our conclusion: average rate charging is the policy to best manage overall peaks and best utilize grid capacity.
Currently, these conclusions are based on traffic and electricity consumption statistics rather than actual measurements. To strengthen the accuracy of our results, ERCOT employee and UT masters student Michael Legatt is assembling a testbed for charging cars located at ERCOT in Taylor, Texas. We will be testing out average rate charging as well as more sophisticated communication and control of electric vehicle charging, including strategies that involve communication and remote control of vehicle charging rates.