Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Marginal power

This site may earn commission on affiliate links.
I often see the following argument: Marginal electricity is coal, EVs represent a new demand, therefore EVs run on coal.

This is wrong.

As an electricity consumer in Norway, I can choose what power I want to buy. That destroys the whole marginality thing in one crushing blow. For the sake of argument, let's pretend that I don't have a choice.

If I lived in a more sunny country, I might have installed solar PV on my roof. Let's pretend I don't have that choice either.

Marginal power is the power that is least expensive to generate that is not already in use. If demand rises some small increment, marginal power will meet that demand. If it falls, marginal power will be shut down.

It's not quite that simple, however.

Marginality also assumes that no one is able to plan ahead and that everyone immediately finds out what the best course of action is. The world does not work that way.

There will also be different types of marginal power for different scenarios. There are four classes of power plants: 1) Base load, running 24/7 *EDIT deleted "full power" which I don't really know whether is true or not * 2) load-following, which follow the load curve throughout the day, 3) peak power which is only started up when load is approaching the average high and 4) reserve power, plants that are so inefficient that they have been closed down completely and are only started up when base power demand is higher than usual, like during a cold spell in the winter.

If load is very high and peaking plants are already running, then more demand will be met by injecting more gas into the turbines, or by building new gas turbine plants if the load is getting dangerously high. If we're talking about a general increase in night-time demand, the peaking plants are unaffected. This will be met by running base load plants harder or building new base load plants.

If, all of a sudden, huge numbers of new EVs were to be charged during the night, then reserve power would have to be fired up. They were shut down because of high fuel costs, so if the new demand can be expected to last, then it makes sense to build modern plants so they can be shut back down again.

The introduction of EVs will be a slow and predictable process. Reserve power will not be used to meet such a slow increase in base power requirements, this is something that can be planned for. I don't think the marginality concept is very useful in this situation.

Also, if marginal power is coal, then why are new gas turbine plants being built all the time?

Finally, with a smart grid and lower charge cycle cost, EVs can help smooth out the demand curve.
 
Last edited:
You need to clarify the time-scale over which your reasoning is to apply, and whether it varies between the individual points. There is a big difference in day-to-day or week-to-week demand-following strategy (adjusting output of existing plants), as compared to demand planning 3-5 years from now or 15-20 years from now (building/expanding/retiring plants).

1) Base load, running full power 24/7, ...
...
This will be met by running base load plants harder...
As presented, these two statements are logically incompatible.

Finally, starting a logical path of reasoning with this declarative yet unproven (at that point) statement
"This is clearly wrong."
will throw off lots of people that aren't already convinced of your position. Lay out the facts and your deductive points, and allow the message to come out at the end.

Now as there are several others on here with actual utility power experience, I'll pipe down and let them review the finer points :smile:
 
As the grid exists now, in most places, EV's are marginal demand, and if charged at night they will most likely be charged from coal power, and if during the day probably natural gas. You may be able to choose where you send your money but the power is coming from whatever is available and cost effective at that time. You might be able to pay only for wind power but you won't actually be getting any of it. The extra load your car puts on the grid can't be met by turning on a windmill.
 
You may be able to choose where you send your money but the power is coming from whatever is available and cost effective at that time.

While true, choosing where you will send your money puts that amount of "type x" power into the grid at some nebulous point in the future, as opposed to not having it go into the grid at all. So even if it isn't running your car, it's running something that would otherwise be powered by coal. A kWh used is a kWh used whether is is running a car or watching TV.
 
You need to clarify the time-scale over which your reasoning is to apply, and whether it varies between the individual points. There is a big difference in day-to-day or week-to-week demand-following strategy (adjusting output of existing plants), as compared to demand planning 3-5 years from now or 15-20 years from now (building/expanding/retiring plants).


As presented, these two statements are logically incompatible.

You are absolutely right. I should have thought this one through a couple more times before clicking the send button.

I didn't separate the short-term and long-term effects.

If demand continues to rise over time, then new power plants must be built. If I as an individual plug in my EV one night, then some power plant must consume more fuel or water for a few hours. These effects have very different costs and time scales, and this is also part of the problem with the "coal is marginal, so EVs run on coal" argument.

The grid can't be powered only by plants running at 100% 24/7 even during the night. It is my understanding that even base load plants can be adjusted, but I don't know to what extent that is being done vs. running the base load plants at full power and using other load-following plants in addition even during demand lows.

Finally, starting a logical path of reasoning with this declarative yet unproven (at that point) statement

will throw off lots of people that aren't already convinced of your position. Lay out the facts and your deductive points, and allow the message to come out at the end.
You are right again. Thanks. I had realized that by myself and was hoping no one had answered so I could edit :)

My main points are 1) that this issue is more complex than some make it seem, there are many different classes of power plants and plugging in at different times of day has different effects, and 2) that you can't just say that EVs run on coal power based on what happens the instant you plug it in. A change in the demand curve that persists over time has a different effect.
 
Last edited:
A lot depends on your local grid characteristics. If you add gobs more night time EV charging in Ontario it will actually help Hydro One with their excess nighttime generation problem. Normally none of the carbon generators are running at night, because they are used for peak load.

Yes electrons are fungible - you get what the grid is producing, not what you "paid" for. But I don't see how it is legitimate to assign "marginal load" to an EV just because it is a "new" device. People are buying new big-screen TVs all the time and no one calls them "marginal load". As I see it, loads are just as fungible as the electrons.
 
As the grid exists now, in most places, EV's are marginal demand, and if charged at night they will most likely be charged from coal power, and if during the day probably natural gas. You may be able to choose where you send your money but the power is coming from whatever is available and cost effective at that time. You might be able to pay only for wind power but you won't actually be getting any of it. The extra load your car puts on the grid can't be met by turning on a windmill.

I'm aware that my electrons are not being moved by the utility I paid to do it. But you have to either accept that power can be traded or not.

In Norway we used to have 97% hydro. That's not true anymore, because the utilities now sell the green-ness of 60% of their power to Germany and the Netherlands. We don't have anywhere near half the transfer capacity to transmit the actual electrons there even in theory. If they can do that then I can equally insist that I'm charging with hydro because I paid extra for it.

Now anti-EV economists and statistics types are telling me that my EV is running on Danish coal, because the Danes have to power up some old mothballed coal plants during extra cold winters when there is little water in our dams (i.e. coal is the marginal power - true, but misleading). They simply cannot have it both ways - either the utilities sell green power to Germany and I buy green power for my EV, or my EV is still running on 97% hydro as it always did. And either way, it's crazy to label the power consumed by EVs as "dirty coal" just because we depend on Danish coal for 5% backup during extra cold winters. This is application of seasonal timeframe marginality to both longer and shorter timeframes, which leads to the wrong conclusion.

We also buy Danish coal and Swedish nuclear energy during the night and sell hydro back to them during the day. This does not mean that my EV is being charged with marginal Danish coal during the night. It means that we're helping the Danes use their generating capacity more efficiently by not having to ramp. They're paying us good money to do it - their power is cheap at night because no one needs it and we can take it. Our power is expensive during the daytime because they can buy it to avoid ramping. Our power is more valuable because it is more easily regulated, so it makes sense to use it to lower the peaks of the Danish supply curve and use spare Danish generating capacity to meet as much of the Norwegian base load as possible to save water for demand peaks and wind lows.

It would often be correct to say that during the night, Danish coal is marginal, so new night-time demand will be met with Danish coal power. It would be wrong to say that EVs are dirty because they charge during the night - we sell the same amount of hydro power back to them the very next day. Neither is correct when there is a lot of wind or rain. When the wind is blowing they need to get rid of their excess wind power first, and when our dams are full and the rain keeps falling, we need to offload excess water at any price higher than zero. Last summer and autumn the price of electricity was very low because the turbines had to run all out all the time because the rain would not stop. We hit a new electricity export all time high. In 2010 we had a dry autumn and extremely cold winter, and we broke the import record.

The marginality concept can be used to find out what will be used to satisfy the next demand increment within a given time frame, but I don't think you can use it to assign a CO[sup]2[/sup] emissions value to a specific class of consumer. Demand is the sum of demand of all consumers and the blame can just as well be assigned to people not fixing their leaky windows, and too many variables, too many different scenarios and different effects over different time frames make it impossible to arrive at a simple number anyway.

I would love to hear what pro-EV economists and those with utility power experience think about this.
 
Last edited:
I think I am now able to formulate why blaming EVs for marginal coal emissions seems so unfair to me.

When you add another watt of load, that watt is going to affect the grid in some way. Marginality is useful for reasoning about its effect and what must be powered up or built to deliver the last added watt of load.

However, the last watt added would have had a different effect if no other loads were present. Every watt of load must take its share of the blame for the emissions caused by the last watt, because every watt of load is equally responsible for the total load being what it was when the final watt was added, and therefore also for the effect of adding the last watt. Thus the only reasonable way to assign blame is energy consumed multiplied by average emissions per energy generated.

Or what?
 
Last edited:
EV's put an extra demand on the grid that did not previously exist. That power has to come from fuel that was not previously being used. More fuel will be burned to charge your EV. Depending on location that fuel is likely going to be coal at night and NG during the day.
 
EV's put an extra demand on the grid that did not previously exist. That power has to come from fuel that was not previously being used. More fuel will be burned to charge your EV. Depending on location that fuel is likely going to be coal at night and NG during the day.

Every new toy that people must have, and for that matter, every new person, puts an extra demand on the grid that did not exist before. The grid is constantly growing and changing depending on the needs and wants of those people. EVs are not any more to blame than exercise equipment or shopping malls with air conditioning or new restaurants. Even the mix keeps changing, away from coal, toward more sustainable, greener means.
 
On the long-term: if the Administration's regulation released last week is formally put in place (which it should occur, unless Congress does something amazingly drastic), there will be no new coal-fired power plants built in the U.S., beyond the few that are currently in the pipeline, unless/until we figure out cost-effective carbon capture and sequestration.

On the short-term marginal fuel: natural gas has gotten so cheap that, in some areas of the country where historically coal is on the margin overnight, natural gas-fired units are now cheaper.

My general rebuttal about "coal-fired cars" is this: Even if even kWh used in my EV came from coal, it is still causing less pollution than a typical gas-fired car. As the system evolves and power generation comes from cleaner sources, my EV automatically gets cleaner, too, while the ICE car stays just as polluting.
 
Right, but the typical counter argument I get, mostly from our friend John Petersen and his followers, is that a Prius class HEV is cleaner than either a coal or NG powered EV, and the well to wheels numbers seem to back that up when you include fracking emissions for NG. That does not of course include the emissions incurred by our foreign policy that is influenced by our oil dependency.

Regarding other appliances adding to grid emissions, that is of course true, but we are trying to claim that our grid powered EV's are cleaner than ICE's, no one is trying to claim their toaster is cleaner than another toaster. The cleanest EV charging possible at this time would seem to be in the ERCOT area when excess wind power at night can't be utilized and windmills must be braked or feathered to cut power. Unfortunately that area is probably going to have a very low EV penetration for a long time.
 
...
My general rebuttal about "coal-fired cars" is this: Even if even kWh used in my EV came from coal, it is still causing less pollution than a typical gas-fired car. As the system evolves and power generation comes from cleaner sources, my EV automatically gets cleaner, too, while the ICE car stays just as polluting.

Robert,

Isn't this just old data? When I hear the long tailpipe argument I now say firmly, "Obsolete!"

In the old days when there were many more coal plants (and they were much dirtier then), and electric cars could only go 25 miles with lead batteries that was true. But we no longer have vinyl records, dial phones or cameras that use film. It's a new world and electrics are now cleaner than gas cars even in the worst case.
 
EV's put an extra demand on the grid that did not previously exist. That power has to come from fuel that was not previously being used. More fuel will be burned to charge your EV. Depending on location that fuel is likely going to be coal at night and NG during the day.

I completely agree, but this does not address my argument that all loads present must equally share the blame for forcing the coal plant into action when I plug my EV in.

I'm not talking about the long tailpipe argument in itself here, but rather the argument that marginal power, usually coal, must be used to calculate EV emissions in the long-tailpipe argument.

This argument is being used to counter the argument that EVs improve at the same rate as the grid.

No extra coal would have to be burned to charge my EV if someone somewhere put on a sweater and turned the thermostat down from 21 to 19, or draught-proofed their windows, or switched a factory to more efficient motor drives, or unplugged their mobile phone chargers when not in use, or any other of a myriad options. They are equally responsible for generating the high load that puts coal on the margin. Therefore the energy mix emissions rate of the grid must be used to calculate EV emissions.

In countries with a low grid emissions rate the outcome becomes very different for energy mix versus marginal power emissions. Marginal power is the power that is the least expensive per kWh that is not already operating. That means that until all base load plants have low per-kWh operating costs (nuclear, hydro), marginal power will always come from something that has higher such costs than nuclear. For the foreseeable future that means fossil fuel. That again implies that if we accept this argument as valid, then we accept that EVs will be charged purely with fossil fuels until all fossil fuels have been phased out!

Even in France and Norway, which have approximately 90% and 95% emissions free electricity respectively, imported electricity from coal is periodically on the margin, and I refuse to accept the blame for charging my EV with energy from coal when I know that 95% of the electricity in this country is hydro. I will accept my share of the blame (5% of the power I use), but nothing more.
 
Last edited:
Robert,

Isn't this just old data? When I hear the long tailpipe argument I now say firmly, "Obsolete!"

In the old days when there were many more coal plants (and they were much dirtier then), and electric cars could only go 25 miles with lead batteries that was true. But we no longer have vinyl records, dial phones or cameras that use film. It's a new world and electrics are now cleaner than gas cars even in the worst case.
How is the data obsolete if plugging in your EV uses marginal production, which is still mostly coal at night and mostly NG during the day?
 
I completely agree, but this does not address my argument that all loads present must equally share the blame for forcing the coal plant into action when I plug my EV in.

I'm not talking about the long tailpipe argument in itself here, but rather the argument that marginal power, usually coal, must be used to calculate EV emissions in the long-tailpipe argument.

This argument is being used to counter the argument that EVs improve at the same rate as the grid.

No extra coal would have to be burned to charge my EV if someone somewhere put on a sweater and turned the thermostat down from 21 to 19, or draught-proofed their windows, or switched a factory to more efficient motor drives, or unplugged their mobile phone chargers when not in use, or any other of a myriad options.
Here's the problem. Let's say they did all of that, which reduced grid emissions. When you plug in your EV you still end up raising grid emissions. You are basically asking everyone else to reduce their grid emissions so you can increase yours.
Areas with large amounts of hydro can use some hydro capacity for marginal loads, but it's hard to tell how much because they usually also export excess hydro power elsewhere. If you use the hydro for your EV it can't be used to displace fossil fuels in some other location, so again over all grid emissions increase.
 
An analogy:

You are the manager of a team of engineers. You're responsible for one aspect of a new product. The product must be ready at a certain date. Your team is on the critical line, you have to spend lots of expensive overtime hours and hire consultants and you might still not make it in time - because of earlier delays that you were not involved in. Are you and your team to blame for the cost overrun or a possibly missed deadline? Do you share the blame with the other teams, or is it all your fault?

I as CEO can say that you and your team caused the cost overrun because you worked so much overtime. That would not be entirely fair, would it?

But everyone would agree that your team's hours were the most expensive.
 
Last edited:
You have to take the grid as it is. The question is what happens at the time you plug in your EV. Your argument ultimately boils down to the equivalent of "what if there were so few people on the planet that hydro power could meet all marginal needs?" Even if everyone else in your grid network cut their consumption as much as possible so that everything were running completely from hydro power only, if plugging in your EV pushed load beyond existing hydro capacity that demand would have to be met by some other marginal source. Certainly you could claim that you removed an electric hot water heater and replaced it with a solar hot water heater, and that made up for your EV charging so your overall emissions remained the same, but the counter argument is if you had replaced your electric hot water with solar but did not drive an EV emissions would be even lower. Understand I'm playing devils advocate here and bringing up the potential arguments you might hear.
 
You have to take the grid as it is. The question is what happens at the time you plug in your EV.

No, the question is who is to blame for what happens when I plug in my EV, and consequently whether I should be charged with all carbon emissions of the marginal power source or whether they should be distributed among all consumers.

It's not my fault that all the "good" electricity has been taken, just like it isn't my hypothetical team manager's fault that his team's work was expensive. The cost must be shared among all those who contributed to the situation.

Your argument ultimately boils down to the equivalent of "what if there were so few people on the planet that hydro power could meet all marginal needs?"

No, that's completely beside the point. Please consider my analogy.

Understand I'm playing devils advocate here and bringing up the potential arguments you might hear.

I have no problems with that :) This discussion has helped me think much more clearly about this issue already.