Carnegie Mellon: “Optimally Varying the Charging Speed” for EVs Could “Cut Electricity Cost in Half”


Recently, researchers at Carnegie Mellon University put forth a study with this conclusion: optimally varying the charging speed of plug-in vehicles can cut the electricity cost in half.

Charging a Prius PHEV

Charging a Prius PHEV

The full study won’t be released until sometime later this month.  The study, funded by the CMU RenewElec project and the National Science Foundation, will appear in the journal Applied Energy.

For now, we have these few statements on the study from Carnegie Melon University:

“Although plug-in electric vehicles currently make up a small percentage of the vehicle fleet, federal and state subsidies are promoting electric vehicle sales, and they may become a significant portion of the fleet in the future. If owners regularly plug in these electric vehicles when they get home from work, this would add to greater demand, when the most expensive power plants are running.”

“Controlled charging can shift loads later in the night when cheaper power plants are again available.  Controlled charging could also help to manage fluctuations from renewable energy sources like wind and solar power, which change their output as the wind changes and as clouds pass by.”

“It is already cheaper to charge an electric vehicle than fill up a gasoline vehicle.  But allowing grid operators to control electric vehicle charging speed could reduce these costs further. We see additional savings up to $70 per vehicle each year or even higher for systems that expect new power plant construction and systems with a lot of wind power.”

Carnegie Melon University says that additional studies will need to be conducted to fully understand all of the implications of controlling EV charging.


Category: General


28 responses to "Carnegie Mellon: “Optimally Varying the Charging Speed” for EVs Could “Cut Electricity Cost in Half”"
  1. Johnny GT says:

    I wonder how hard this would be to accomplish at a residential level. Currently I’m building a house and have purposely created a large expanse of south-facing roof that will be blanketed by solar. My utility does not charge more for peak power than off peak, but they do pay less for power put back into the grid for net-metering than I pay to take it back out. I’m OK with that, it’s their wiring I’m using. How difficult would it be to tell a car to absorb any excess power from the PV system instead of dumping it back into the grid:?

    1. Dr. Kenneth Noisewater says:

      If you’re charging overnight, you’re getting grid power anyway, unless you have a battery backup.

      That said, I wouldn’t worry too much about it, if you’re grid-tied just try to size your installation so that you’re providing as little surplus as possible. I’d also consider getting an inverter that can take multiple source inputs, such as battery or natural gas backup generator, even if you add such things later.

      1. Alaa says:

        A man from IRAN

        He showes how to use a Grid Tied without a Grid. He makes his own pure sine wave to get the micro inverters to start.

          1. Mark H says:

            Microinverters like Enphase can be AC-coupled but it is expensive. Microinverters are extremely easy to deploy with no string calculations and easy to add to the system though

    2. Dave R says:

      The difficult part is getting the car to accurately respond to requests. By using a quick charge protocol like CHAdeMO it should be possible, but using J1772 many cars have idiosyncracies that would prevent effective management.

      Without car-side control, you can only limit J1772 to 6A minimum which is 1.4kW, so if you want to reduce draw less than that from outside the car, you have to cut off charging completely. And when it’s time to ramp back up, not all cars will ramp up to a faster rate.

  2. Djoni says:

    Not a difficult task anyway.
    When it come to manage power, it’s alway’s a lot easyer to do it with electricity than any other source.
    It’s more a software problem than a harware one, althout a bit of hardware might be needed.

  3. Anderlan says:

    Duh. Slap your EV dissing friends with a knowledge stick.

    Fundamentally, the centralized fossil baseload power architecture is just as imperfect as any variable distributed generation network for the opposite reasons. And time-shifting helps them both become more perfect.

  4. vdiv says:

    That’s exactly what we need, slower charging or big brother telling when you can/cannot charge.

    Cut the cost to whom exactly? Most EVs already have some sort of timers to schedule charging if time of use metering is in place.

    1. Dan Hue says:

      That is the price to pay for living in a civilized country. Sorry that it is so inconvenient to you.

    2. Dave R says:

      Cut the cost to the consumer, of course. Of course, if you decide that you want a charge NOW, you won’t pay the lowest price.

      It’s just like how TOU pricing works now, except in real time instead of averaged out over entire seasons and time-periods.

  5. Brian says:

    I would be happy to sign up for a program like this. I would be much more comfortable with tapering charge rates than actually taking energy out of the pack (i.e. V2G). As it is, even L1 charging is faster than I ever really need. As long as there is some reasonable lower limit (e.g. L1, or 1.44kW), and presumably a way to override it if necessary (for a fee of course), I don’t see any problem with this at all. We already have an aging grid. As more renewables come online, stability will become an ever more pressing issue. Let’s make EVs part of the solution rather than part of the problem.

    1. Dave R says:

      Yes, essentially what you’d do is say “I want to be fully charged by 8AM” and the car/grid would then figure out the cheapest way to get that done.

  6. ffbj says:

    Just running the numbers from the article.
    The headline reads say up to 50% on electricity.
    The article says $71 a year saved. So assuming that is their 50% savings doubled is $142 dollars per year/on average per pev. Is that accurate?

    1. Brian says:

      The 2013 Leaf is EPA rated at 3.4 miles / kWh (from the wall, not the battery).

      At $0.12/kWh (AFAIK, the national average for electricity), $142 would buy you 1183 kWh.

      1183 kWh * 3.4 miles / kWh = 4022 miles.

      I would say that $142 / year is pretty low unless you’re a light driver (or already on dirt-cheap Time-Of-Use electricity)

    1. ggpa says:

      The link is calculating $45 from using “5,400 kWh of charge per year” … which seems high to me …

      I think Leaf/Fiat/Fit etc can usually get 4mi/kWh, so for such vehicles $30/month is a better estimate.

  7. SeattleTeslaGuy says:

    I’m having a hard time discerning the insight of this paper. Isn’t all this well known?

    And, I don’t believe the 50% savings. Based on 12K miles a year, 330 WH/mi and the national average electricity cost of $.114/kwh works out to $450 a year. $71 ain’t half.

    My provider, Seattle City Light, doesn’t have late night rates. Not sure how many others are like that. Could it be that hydroelectric power is easier to adjust to peak demand?

    1. Brian says:

      Hydroelectric power is very easy to ramp up/down quickly, but you can only do so to a point. Eventually, you have to even out to an average power production unless you somehow manage to stop the whole river. Otherwise, you simply spill excess water over the dam instead of generating power.

  8. ffbj says:

    Yeah I was saying it’s more like a 25% or so not 50%.

    If your average monthly electric bill is Your estimated annual savings on the TOD rate* Your estimated annual savings with reduced on-peak consumption**
    $280 $0 $75
    $300 $17 $100
    $325 $35 $125
    $350 $60 $160
    $400 $100 $215
    $475 $165 $310
    * Based on 25 percent on-peak and 75 percent off-peak monthly consumption.
    **Based on 20 percent on-peak and 80 percent off-peak monthly consumption.
    Another way to show this would be to compare average rates across the board for off peak vrs peak. The article would indicate up to 50% difference between the two.

  9. ffbj says:

    @Brian @SeattleTteslaGuy. Yeah and also why Wahsiningyton has the cheapest electricity in the nation under 9 cenrts an hour. Boo Hoo, no off peak pricing. j/k

    1. SeattleTeslaGuy says:

      Well, Seattle City Light (supplier for Seattle and a fair amount of King County) top tier is 11.4 cents per kwh. I think all EV owners will be charging at the top tier. So, not exactly a boohoo situation, eh?

      My fiance’s sister lives near one of the dams on the Columbia and they pay something like 4 cents a kwh.

  10. Anthony says:

    I already do this! I just switched away from flat-rate pricing to Time-of-Use pricing. I went from flat 12c to 5c at night in the winter, spring, and fall and 7c in the summer. I saved $500/yr on my regular electric bill and then some more compared to my original cost expectations of owning an EV.

  11. Joe Huber says:

    This proposal is way more than just ToU incentivized time shifting.

    Since consumption and generation must always be precisely balanced in the utility grid, there’s a lot of value in being able to ramp an EV’s charge rate up and down in realtime to balance the ever changing consumption and generation patterns. Currently utilities PAY for ancillary services to balance the grid as needed. This EV SmartCharging approach could PAY EV owners for being able to control their charging speed in real time to keep the grid balanced.

    This payment might take the form of a monthly credit on your utility bill or might be an even lower electric rate than the nighttime ToU rate.

    And since EV charge rates can respond quickly (vs ramping a power plant up or down) this would be a valuable service (i.e. worth $$$) to help compensate for the variability of integrating renewables like solar and wind.

    1. SeattleTeslaGuy says:

      I’m not saying it’s a dumb idea, just that I’ve heard it before. Not just for EV charging. There have been similar schemes proposed for water heaters and air conditioners. Those didn’t go anywhere due to the requirements that the in home devices be smart and connected. The problem is that you need to get a critical mass of people with smart-connected chargers. I just don’t see it happening anytime soon and will probably require a government mandate.

    2. Bill Howland says:

      Actually schemes for water heaters used to be in common use years ago. The billing meter (totally mechanical) also had a time switch in it, along with using the 5th jaw in a meter socket, to shut off the lower element during high peaks. This was Nyseg, and the reason they HAD to do this is decades ago they only had really tiny generating plants and had to watch their peak usage very carefully since they didn’t have much to spare. The Genius of this solution was, the entire water heater was the ‘energy storage system’. The element at the top ran to keep the discharge water hot, while the entire rest of the tank got cold. The lower element would turn on at either 11 or midnight and run the whole time to heat the tank for the next day, and at 7 am the process would repeat. They still have the plan in effect, but you now have to supply your own timer/contactor.

      As EV’s are similiar loads, TOU plans are really all that is necessary.. Although National Grid’s rates for TOU (my utility) are far less compelling than NYSEG’s, I would convert to TOU if I had anything more than charging 2 EV’s (which in the winter time draw much more juice than the 2 ev’s in summertime plus the central air conditioner, really the only other large electrical device I have). All other large energy consumers (including a huge hot tub (8′ x 10′),) I heat with natural gas, so I’m not quite a large enough electric consumer to benefit from TOU in National Grid Territory. If I had no natural gas service, then it would be a no-brainer.

      Seeing as people last year in Naperville, Illinois got arrested if they refused Smart Meter installation, there is obviously governmental public/private partnership going on in those locales. So much for the customer’s “Implied Consent”. Hopefully, Juries (and brave judges) will eventually shoot much of this down, but the EV angle is just another Faux Justification, when much cheaper simpler methods in accordance with Occam’s Razor have withstood the test of time.

  12. shawn marshall says:

    charging intervals can be controlled by the car – no outside interference necessary – the car divides the charge time into segments over the available interval and randomly charges. This will spread the charging load all across the night time charge interval leveling the load on utility generation. This means less generation(expensive) is needed and that in fact, the cheapest generation can supply the load as long as possible. This is a real efficiency and would result, after a long time and many EVs, in savings for the utilities. This savings should eventuate in slower rate increases at some point, some point afar. The car can also be programmed to avoid morning, evening, low temp and high temp peaks if the interval length will allow these periods to be skipped. Not well stated but you know what is intended hopefully.

  13. Jouni Valkonen says:

    I wrote a feature suggestion at Tesla forum that would do just that. Not to save money (Tesla owners typically are not in need for saving few bucks), but to do altruistic good for the society as a whole. And to my knowledge JB Straubel hinted in Norway or Nederland that this kind of smart charging feature is coming to Model S.

    Using Tesla car as free storage for renewables