The grid and how EVs will affect it

This discussion may not deserve its own topic (separate from the blog’s plug in article), but I think there is a big potential for the grid to crumble with increased EV usage.

Take my house: I am on natgas for cooking and water heating (22 year old water heater will give out soon, right?). My demand kW is usually 4.5-5.3 kW (based on the elec dryer’s ~3kW and whatever else is operating at the time). Now I’m considering whether to add a NEMA 14-50 at 40A (32A usable) or NEMA 14-30 (24A usable) 240V receptacle for my mother-in-law’s Tesla (that may get delivered this year). The 32A is the max current a mobile connector will take, probably so it doesn’t trip on a 40A-breakered 14-50 receptacle.

But that’s 7.7kW. My house will more than double its instantaneous load on the grid, but only for a week when she comes to visit. As EV sales double by the year, this ramp will start to impinge on the available capacity of plants, transmission lines and distribution infrastructure. I think this is manageable, but the framework to manage it isn’t really being implemented. Tesla’s have wifi; couldn’t they check the local utility’s stress level and cut to zero amps when something happens to perturb the grid? The system exists in hardware for AC; it’s a voluntary cutout that shuts off your compressor and condenser fan when the utility feels like it (micro rolling blackouts, but only for AC).

If it makes sense for AC it makes sense for EV charging, though EV’s haven’t reached nearly as high market penetration. And I think the implementation could be much less brute force than simply opening a switch. Perhaps it would limit charge amperage smoothly. My guess is most Tesla drivers simply plug in their car when they get home at 6, and most don’t limit charging to after 9pm, at least in Idaho where Time Of Use rates are rarely used. And even if most EVs started charging at 9pm, that’ll start to become a new surge once EV penetration grows higher. So if people would stagger their start times and feather in the amperage over 5 or 15 minutes, it won’t be a problem. But is this how it’s done? I drive a 20 year old van less than once a week, so I dunno how the current crop of earlyish adopters initiate charging.

And I’m not even talking about Vehicle to Grid, just “smart” charging so the public good of the grid isn’t as stressed as it otherwise would be. $/kWh incentives or demand charges are probably the way to solve this imminent problem, but maybe there are gentler nudges that’ll keep our grid stable and allow us to get from 2% EVs to 50 or 75% EVs in 30 years.

The answer is “it’s complicated”. I lean toward the more skeptical side, but I will play devil’s advocate here.

As has been pointed out to me, in areas that aren’t too dense, you can deploy localized solar and charge during the day. Since the solar is localized (e.g. in the store’s parking lot or at the office), the transmission lines aren’t a factor. It’s also relatively cheap, albeit not free, to add the panels and wiring.

In Texas where I live, it is not atypical to see a peak on a hot summer day at twice the overnight low. ERCOT is forecasting a peak >80GW for this summer. Overnight load drops to 40~45GW. If you restrict EV charging to 10PM-6AM and assume 30GW of available transmission capacity, that’s 240GWh that you can deliver on existing infrastructure. If you also assume 3mi/kWh, that’s 720 million miles. Divided by a population of ~30 million, you get 24 miles/day/person.

Now, I don’t know whether the generation infrastructure could add an extra 240GWh without crumbling. It seems that our infrastructure relies heavily on natural gas plants, so as long as you can source the natural gas, perhaps it’s fine.

Yeah, I think the real answer is ‘it depends’.

Wind? Solar? Hydro? What’s the current power profile and mix look like? Does V2G exist? Is there a utility program to do demand shifting? What’s the current grid in that area rated for?

This topic has come up often enough, and with enough different people, that I’m working on a blog post on it. However, the short answer is “I don’t think it’s likely to take the power grid out,” and people who know far more about power grids than I do have come to similar conclusions. One may reasonably raise answers about how one is going to generate that power without a lot of fossil generation capacity, but I think that’s mostly solvable as well, if you don’t mind the mining and such that go along with it…

It boils down to the fact that while individual EVs may pull a lot of power briefly for charging, on average, there just isn’t that much of an increase in draw over what we have now, and it’s likely to be spread out enough, and growing slowly enough over time, that it’s no massive step change in the power grid function and should be generally absorbed with normal grid growth. This may require newer substations than planned, but it’s a gradual increase, not a sudden overnight change.


So, just to demonstrate the ballparks, I’m going to spitball pretty hard here with local numbers, for no particular reason beyond “I’ve done the math before while playing around.”

Idaho Power serves about 500,000 residential customers.

If you look at their stats for 2020, you find the following (and, yes, I know 2020 was a weird year, but it doesn’t matter for the point I’m trying to make:

  • Total Annual Sales: 16,715,399 Megawatt-Hours
  • Peak Summer Load (August 18, 2020): 3,392 Megawatts
  • Peak Winter Load (Dec. 29, 2020): 2,219 Megawatts
  • All-time Summer System Peak (July 7, 2017 at 5 p.m.): 3,422 Megawatts
  • All-time Winter System Peak (Dec. 10, 2009 at 8 a.m.):* 2,528 Megawatts
  • Nameplate Generation: 3,482 Megawatts

I’m going to, again for the sake of spitballing and easy math, consider that every residential customer drives 35 miles, every day of the year. This is high for some people, low for others, some people have two commuters, some have a short commute and stay at home parents, but it’s “something about average” to work from. And, for the sake of making the math easy, I’ll call it 10kWh/day for that 35 miles. It will be a bit higher in the winter, a bit lower in the summer (potentially), and it depends on the vehicle, but, again. Ballpark region here.

So, that works out to an annual power use of:

500000 drivers * 10 kWh/day * 365 days = 1825000000 kWh = 1825000 MWh

EV charging use:  1 825 000 MWh
Annual Sales:    16 715 399 MWh

So, it works out to an 11% increase in annual system power if every house is driving an average amount. Plus commercial trucks and such, which I’m not including here because I intend to dig into these numbers more at some later point, but handwave here It’s probably fine.

And, again, for “scope of the problem” estimates, 10kWh/day is only an extra 415W, spread over the 24 hour day. Yes, some people will be whomping away at 9600W, but not everyone, and those people won’t be doing it all day long.

I also tend to think it both “likely” and “a really good idea” to do a lot of EV charging during the solar induced mid-day lull. A lot of solar is behind the meter and manifests as “demand reduction” on the grid out here, so it’s a bit hard to tell what exactly it’s doing, but there’s a typical morning peak around 7-9AM, an evening peak from 5-8PM, and then a mid-day minimum and an overnight minimum.

If, as I’ve argued elsewhere, places put “slow, dumb charging” in (say, 16A/240V/3.8kW unmetered), you can plug in at work in the morning, charge for a while during the day, and make up any surplus in the evening/at night - and be just fine. This neatly shifts load into the morning and early afternoon period when solar is doing a lot and the grid demand is low. It also leaves late afternoon/early evening production that could go to charging substation batteries for the evening peak.


32A of charging is 7.6kW, or a 100kWh Tesla battery from empty in about 13 hours. Do you really need that much charging for her? Would a 3.8kW 16A circuit (27 hours empty to full to full) be easier to deal with? Unless, of course, you’re looking for the EVSE tax credit and coincidentally ending up with a usable 40A 14-50 in the garage. :wink:

If only they had wifi only… they have a full cell data connection that’s always live. However, the short answer is “It doesn’t matter and the power company doesn’t find your Tesla worth bothering with.” Not even on a reasonably large scale. If they need to shed load, they talk to the big industrial customers who are pulling 30-50MW and have them back off, via established contracts, with plenty of warning.

We might see some sort of EV charging cutoff option along the lines of the air conditioner credit ($5/mo for three months a year, IIRC), but it’s not needed any time soon, and is unlikely to really matter much.

If you’re looking at “100% EVs overnight,” the grid would handle it fine for most of the year. There would be a few days where it would matter, but… otherwise, it would be fine. Looking at 50-75% Evs in 30 years, it’s just a rounding error on general electrical use growth. Remember, the average vehicle age in the US is something like 11 years old. Cars don’t get changed out overnight (though, lately, it doesn’t look like anyone can even buy new cars… they use a lot of unavailable chips).

But, first order guesstimate, 10-20% increase in use over baseline, slowly creeping in over the years.

1 Like

I wonder what percentage of Idaho’s power usage is currently taken by gas stations?

I don’t think an EV-caused grid collapse is going to occur simply because it’s a slow moving problem, and utilities are pretty good at dealing with those. It’s the fast moving ones that get you.

It is interesting to consider what happens when ALL of your personal energy usage is dependent on the grid, as we can see from Texas it can be nice to have non-grid tied heating devices, even if it’s just idling a (gas) car.

It can’t be that high. Pumping a liquid from 10’ underground to aboveground at a decent flow rate just doesn’t take much energy.

Agree. No matter how fast you try to do it, it can’t be that fast.

Yeah… that’s a very real risk. I’m working on some solar power trailers with a friend, we just can’t get the fleeping parts for it. :confused:

How much power can your reasonably get from a trailer? I assume it’s not bad considering the size of your a-frames.

Less than we’d like. Think about 1000W of panel, the larger versions are on hold due to some engineering challenges. So, 5-8kWh/day depending on how you feel like moving it. It’s enough to be useful, certainly, though would require additional power to fully handle a typical home’s use.

I really want to use my dual meter setup to create a “necessary” power grid inside the house alongside a “nice to have”. So things like the furnace blower, fridge, would be on one set of breakers and everything else on another.

All you need for that is a critical loads subpanel. No need for a separate meter. It’s a common enough thing.

And then a transfer switch to isolate the panel?

Yeah, you’ve got a manual or automatic transfer switch, and that critical loads panel is fed with grid power when the grid is up, then your other power source (isolated from the grid) when you want to run on generator/solar/etc.

The problem is that unless you design the house for it, it’s usually really hard to do that for more than a few circuits.

One more thing to keep in mind whenever I buy a house. Hopefully it’ll be enough of a fix-er-upper that I’ll just want it stripped back to the studs, and in that case it’s relatively cheap and easy at that time to get a critical loads panel, run all the network conduit/etc, and so on. Pre-supposes I’ve got the money to keep living in the current place for another 6-12 months (I’m being pessimistic here, but better to be pessimistic) while that’s all going on.

So I don’t have to worry about timing her charging, though it’d be best to do it after 9 or 10pm. Doing my best for the grid is inconsequential except, maybe, on the absolute hottest days, right at peak.

And nationwide, even if gas cars stopped being sold tomorrow, it would still take decades before every car was an EV. It’s a slow-moving target that the utilities will manage.

To me it seems that long-term large amounts of EV charging will be a net win for grid stability, especially if they’re connected in such a way that the utility can easily shed load.

How do EVs improve stability? Do you mean that they necessitate an expansion of capacity, which presumably leads to better ability to handle load swings? I think this depends entirely on correlation. If EV charging is correlated to peak usage (4~9PM), and unless it can be shed, then it exacerbates things. If it’s negatively correlated, then perhaps it narrows the gap between peak and overnight low, which means less need for dispatchable sources. But there are a lot of assumptions here about when and how EVs will be charged. Most non-engineer types do not think about this stuff and will do it whenever it’s most convenient, not when it’s most optimal.

California already offers you a “deal” on power if you’ll setup your A/C such that they can remotely shut it off during high load (for a few hours). It seems entirely reasonable to me to make it so that E/V chargers have a similar setup so that in cases where the grid load is getting high they can be remotely shut off or deferred. So instead of doing rolling blackouts you can first do E/V shed.

In fact why not make it so you can set your E/V to “charge to X% by dawn, using non peak power if possible” - for even during normal usage?

This doesn’t even go in to the potential for making EV charging stations more power wall-like and allowing bidirectional energy transfer.

This capability exists in some chargers and I’m not aware of any power companies who have bothered. They just don’t care about small residential loads, and I’m fairly sure that even as EVs increase, they’re not going to bother with too many small loads. The AC load shed things might happen, but I don’t see it as being that likely.

And, of course, getting daytime charging solves a lot in most areas - morning solar production is a good fit for EVs.

Here’s an example of the CA thing: - and I suspect something similar for EVs will be offered once they become a significant part of grid load.

I think utilities don’t care because market penetration is barely above zero. Above some threshold they would suddenly get very interested in shedding EV loads.

As I said above, Boobus Americanus is going to charge at whatever time is most convenient, likely right after getting home from work. Most people will be blissfully unaware that there is even a difference, and most of the remainder won’t be bothered enough to change habits. You need a mechanism to incentivize daytime charging.