What Is Smart Charging? A Look at How Electric Vehicles Fit In

, Kendall Science Fellow | January 7, 2016, 1:24 pm EDT
Bookmark and Share

Can electric vehicles allow the grid to accommodate greater levels of variable renewable energy? That’s what I’m looking to find out.

The grid of the past

The power grid has traditionally operated under the assumption that electricity demand simply happens, and that supply must be ready on command to meet it. This requires developing a system with several layers of electricity supply resources:

  • “Baseload” plants of high capital cost and low operating cost run as often as possible
  • Flexible “load-following” generators ramp up and down to match hourly swings in demand
  • “Peaking” units, cheap to build but expensive to operate, run for relatively few hours per year

Systems have target voltages and frequencies, but there is some tolerance for short-term mismatches of supply and demand. (See “How the Electricity Grid Works” for more details.)

As a 2013 paper notes:

“The operating principle of fossil generation is ‘burn when needed,’ a principle simple enough that it could be followed without computers, digital high-speed communications, or weather forecasting—precisely the conditions when today’s electric system was created, early in the 20th century.”

In this graphic, “Nuclear,” “Coal,” and “Geothermal” are baseload plants, “Hydro” and “Gas CC” are load-following plants, and “Gas CT” represents peaking units. The other resources do not fit neatly into those traditional classifications. Source: The Western Wind and Solar Integration Study Phase 2, National Renewable Energy Laboratory, 2013.

In this graphic, “Nuclear,” “Coal,” and “Geothermal” are baseload plants, “Hydro” and “Gas CC” are load-following plants, and “Gas CT” represents peaking units. The other resources do not fit neatly into those traditional classifications. Source: The Western Wind and Solar Integration Study Phase 2, National Renewable Energy Laboratory, 2013.

The grid of the future

Now, though, we do have computers, digital high-speed communications, and weather forecasting. We also have flexible loads such as air conditioners, electric water heaters, and electric vehicles chargers, alongside those that basically need power on demand, such as light bulbs and computers. These flexible loads can shift the times in which they draw power by a few minutes (or even hours in some cases) with no adverse impact on the quality of energy services delivered.

We’ve also developed technologies that, although cleaner than fossil fuels and cost-competitive, cannot be dispatched on command.

We have also greatly improved battery technology, and developed other options for energy storage. Storage can make renewable electricity available whenever needed, and can help baseload resources supply varying demand.  Even as costs of storage come down, it may be more economical to first use price signals and automatic controls to align flexible demand with supply as much as possible, and then use storage to help meet the remaining inflexible demands.

Diversity of loads and supply

Individual loads can be highly variable. Small appliances are turned on and off frequently (a microwave, for example, uses large amounts of power for a minute or two at a time). Air conditioners automatically cycle on and off to maintain the proper temperature, at times resulting in very large swings in power demanded.

But when you aggregate many individually variable loads, the variability diminishes considerably. This is the same effect seen with solar power, aggregated across a wide area. Isolated clouds passing over one photovoltaic (PV) array will take a few minutes to impact the production of another array a few miles away.

Source: “Implications of Wide-Area Geographic Diversity for Short-Term Variability of Solar Power,” A. Mills and R. Wiser, Lawrence Berkeley National Laboratory, 2010.

We can mitigate some of the short-term variability of PV systems just by having many of them spread out.

Time of use rates and smart charging

On the timescale of hours, variations in solar power output occur from large cloud systems and from Earth’s rotation. The second source of variation is entirely predictable—we know when the sun will rise and set. PV systems reach peak output around noon, with variations depending on the angle of the panels, the location within a time zone, and Daylight Savings Time.

Since peak power demand is generally mid-to-late afternoon or early evening, we’d want to shift some loads to run when abundant solar power is feeding in to the grid. One way to do this using PV and EVs is with a rate structure that offers low-cost electricity in the middle of the day. Electric companies in Hawaii have proposed this. Using “time of use rates” in this way has several benefits:

  • Retail prices paid by customers would reflect the wholesale market, in which mid-day power is already becoming less expensive in some locations due to abundant solar electricity.
  • Electric vehicles at workplaces could charge with low cost clean power.
  • “Smart chargers” could adjust the charging rate if needed to accommodate variations in power supply.
  • Flexible loads such as air conditioners could run at low cost during the mid-day hours to pre-cool homes and reduce the evening power draw.
  • The electric grid would see a more gradual “ramp” from the mid-day load to the evening peak.
  • Solar power systems selling at wholesale (such as utility-scale systems) would see increased revenues from their power sales due to greater daytime demand.

If a car is plugged in at work for nine hours, and needs only two full-power hours to charge, there is a great amount of flexibility in when the charger is used and what rate it charges at.

Shifting the charging time to align with a period of surplus solar power, through “time of use” rates, helps address solar power variation on the timescale of hours. Starting, stopping, or varying the rate of power transfer (“smart charging”) can help address shorter-term variation.

My research: how PV and EVs can best work together

There are dozens of pilot projects on “smart charging,” and hundreds of research efforts in this area on topics ranging from consumer behavior to battery lifetime impacts to communication protocols. Some companies already earn real revenue from providing ancillary services to the grid (including both one-way “smart charging” and two-way “vehicle-to-grid” arrangements).

Part of my work at UCS is identifying how these options can increase value for EV owners, for electricity grid operators, for renewable energy providers, and for the public. Right now it is possible to integrate variable energy resources into the grid at a low cost. Can flexible loads reduce this integration cost further as variable renewable generation grows?  If so, how can we align incentives so that the market prioritizes lower-cost solutions for grid integration of variable renewables over more expensive ones?

Stay tuned.

Posted in: Energy, Vehicles

Support from UCS members make work like this possible. Will you join us? Help UCS advance independent science for a healthy environment and a safer world.

Show Comments

Comment Policy

UCS welcomes comments that foster civil conversation and debate. To help maintain a healthy, respectful discussion, please focus comments on the issues, topics, and facts at hand, and refrain from personal attacks. Posts that are commercial, self-promotional, obscene, rude, or disruptive will be removed.

Please note that comments are open for two weeks following each blog post. UCS respects your privacy and will not display, lend, or sell your email address for any reason.

  • solodoctor

    Thanks for this general analysis. I look forward to more specific ideas in coming articles as to how individual consumers like me can adapt our use of power to make it more efficient/less costly.

    I know,for example, that PG&E here in N Calif where I live offers different rates/programs for those who can shift more of their power use to late evenings/very early AM’s. Ie, they are trying to incentivize customers to charge their EV’s overnight when the cost of electricity is lower.

    Is this the kind of thing you mean?

    • Pete O’Connor

      Thanks for your comment! I will try to do an article on how this would impact consumers, and how other appliances can contribute. OhmConnect is a company with some interesting ideas, I might write up a piece on them.
      Yes, “time of use pricing” is one of the concepts I am looking at. There might (or might not) be value in having even more sophisticated and automatic controls, where the EV charger alters the power flow a little bit to help balance shorter-term variations on the power grid (this is called “regulation”).
      Overnight power has traditionally been the least expensive. But California is starting to see periods of very low wholesale power prices at mid-day, due in part to solar power. Sometimes even negative — they are paying people to use electricity. It takes some time to adjust the “time of use” retail rates to reflect new wholesale market conditions. Maui has changed its rates so that mid-day electricity is the cheapest.

  • Ken Cooper, P.E., P.S.

    Smart charging/flexible-time use is a well-reasoned strategy to reduce peak electrical demands and optimize the use of environmentally sound generation methods. Case in point, my HVAC technician suggested that during the summer AC season I supercool the house beginning at 11 PM and ending at 4 AM. Using this method over the past two summers the AC, which is set at 75 degrees F, has kicked-on before 11 PM no more than 5 times. In addition, my summer electrical bills are approximately 2/3 of the previous years’ bills.

    • Pete O’Connor

      Thanks for your comment! That’s actually a much longer period of time than I was envisioning for pre-cooling; your home must be well insulated. I was just anticipating pre-cooling from, say, 1-4 pm with a low-cost mid-day TOU block.

      • Ken Cooper, P.E., P.S.

        The thinking is that the lowest energy costs are overnight plus after dark is a cooler part of the day than when the sun is up. My house is well insulated…however, it was build in 1960 and the windows are the weak link, but they are architecturally significant in a house that is architecturally significant, so I kept them and had some single-pane storm windows built for the house. Not great for insulation value, but the storm windows reduce most of the drafts. Anyway, even though the pre-cooling period extends for 5 hours, it rarely takes that long for the house to reach the target temperature. The 4 AM time was just to ensure a pre-cooled structure before the daylight period. Lastly, my house is a one-story ranch with 30-in deep soffits that keep the walls shaded throughout most of the day.

  • JRT256

    Your supposed demand chart is certainly not representative of either AC or non AC season demand. With non AC demand, the peak starts at sunset and continues to about 10:00 PM. Demand reaches a minimum at about 4:00 AM and returns to normal at about 8:00 AM with a slight morning peak at 10:00 AM. AC demand would be added to that. We probably also need to consider that as we convert to heat pumps for heat instead of burning fossil fuel that this will also change the demand curve.

    I note that later you presumed that people would have electric water heaters but not electric heat.

    You make a serious mistake by presuming that in the future, people will continue to have old obsolete ACs rather than new efficient ones with variable speed compressors (VRF — Variable Refrigerant Flow). These do not cycle on and off but rather run all the time at a variable output when cooling is needed. These could be interrupted but much less power would be saved and they would be slightly less efficient when they came back on for a a period of time.

    We should also consider that people will start installing desuperheaters on their AC to partially heat the water so the water heater savings will be reduced in the AC season. It is also possible that home size ACs will start to come equipped with an accumulator and receiver so that more heat can be used for hot water.

    As I said above, we will have to switch to electric (heat pump) heat (possibly solar assisted) if we are going to stop using carbon containing fossil fuel. This will also make a large difference in the shape of the demand curve. it will tend to flatten it out. So, some of the time of day ideas are not going to be applicable in the future

    The idea of using solar to charge electric cars sounds nice. However, have you really thought this through? It takes a lot of power to charge an electric car. This should immediately raise issues regarding this source of solar power, its alternate use and its backup when it wasn’t available. I presume that the large amount of power required would mean that solar PV would have to be installed specifically for this purpose. Then, could it be used all day for charging cars? If not, what other use could it be put to? Power can’t just be dumped on the grid because the gird is not a battery. And, with all of these cars to charge, what would backup the solar when there wasn’t enough sun?

    Storage could solve these issues. But, now we have only pumped hydro storage. And, we could upgrade conventional hydro to use for backup and storage. If you are counting on having other storage available in the near future at reasonable prices for general grid storage. You will be very disappointed. Batteries now cost 50 to 100 times what they would need to cost to be practical for grid level storage. It will be at least Two decades before we see anything but specialized battery storage.

    • Pete O’Connor

      The dispatch stack results from a model of the Western Interconnection, using PLEXOS. The discussion of those results begins at p. 84 of the linked report if you would like more details about the methodology. I use the image solely to illustrate the basic concepts of baseload, load-following, and peaking generation. There are other dispatch stacks I might have used, but nevertheless it still seems to peak in the afternoon as I am looking at it.

      I do not make any claims about what future AC systems will look like.

      Some areas do in fact use electric heat as a flexible load to integrate variable renewable energy resources. There was a pilot project in New Brunswick doing this. Electrically-powered heat pumps are an efficient option as you note.

      PUCs would be most interested if technology in new homes could flatten out the load curve. But load shapes have gotten worse in recent years as EPRI has noted. Given the slow turnover of the building stock, many states are at least considering expanding time of use rates.

      While PV systems installed specifically for charging EVs do exist (i.e. standalone systems with batteries), that is not at all what I am discussing here. Consider that states like California and Hawaii already have a lot of solar feeding in to the grid. The reductions in mid-day wholesale power prices due to this zero-marginal-cost resource argues for
      increasing power demand in the hours when the solar is generating. Use price signals to shift demand to when supply is abundant. A low-cost mid-day TOU block would encourage workplace EV charging. Flexible
      loads that can use this power (and therefore *not* need to draw as much power in a late afternoon peak period) would offer economic benefits for the system as a whole compared to a system in which those loads do not shift in response to price signals.

      The solar’s already here and on the grid. Solar energy production was about 37 million MWh last year. An EV uses roughly 4 MWh per year. So we currently have enough solar power for 9 million EVs (FAR more than are on the road). With the right price signals, flexible loads can help integrate this resource at a low cost. The other option, the one we are currently using, is the suite of ancillary services that back up all of the generation on the grid (all of which is subject to unplanned outages). Flexible loads can not only help integrate solar but also help respond to conditions like plant outages, transmission congestion, and so on.

      I think it will be at least two decades, probably three, before we have enough variable energy resources (as in 80%+ of generation) that we even need large amounts of grid-scale storage. There’s a great amount of flexibility in the existing supply (through the ancillary services markets), and there can be in demand as well.