Turning power plants off and on?

In this post I attempt to address head on an issue we've been dancing around for a while. In previous posts (such as Solar in Winter and Solar in Summer) I've shown how adding Solar generation above 10 to 13% necessitates that portions ("blocks") of coal or natural gas power plants be turned completely off to accommodate the solar (this is known as "cycling"). I've argued that utilities are loathe to actually do this -- but I've only spelled it out in terms of what the utilities are economically motivated to do. The real question (at least in my mind) is how much worse are carbon emissions when a power plant has to be turned off then back on again (cycled)? Some have argued that utilities are just -- for lack of a better word -- "whiners" that should just suck it up and cycle their plants to make room for carbon free renewable electricity. On the other hand others argue that that cycling results in so much additional carbon emissions that the cycling completely negates the carbon saved by the renewable electricity!

First a little review. Recall that a coal or natural gas power plant is divided into multiple "blocks" -- each of which is essentially independent of the other. One can turn off and back on each block individually. This is kind of like some cars that have an 8 cylinder engine but at times are able to turn off 4 of the cylinders -- imagine that a 4 block power plant is a 4 cylinder that can run on 1,2,3 or 4 cylinders at any given time. A typical power plant has 4 blocks. Turning off an individual power plant block is a lot more complex than turning off an engine cylinder however. After the block turns off, it is still using fuel, so during this time coal or natural gas is being burned but no electricity is being made (ie it is all waste). Obviously if we're trying to minimize carbon emissions this is bad. The amount of wasted fuel when a block is cycled off then back on can be divided into three portions. The first portion is the amount of fuel burned during cool down -- remember that because of the extremely high temperatures in the metal pipes and such the temp needs to be lowered slowly so that nothing shrinks and cracks. The second portion is the fuel used during warm up -- basically the same idea as with cool down. The third portion is a baseline amount of fuel burned to make electricity to keep various parts of the plant operating (such as pumps and fans). This portion is referred to as "auxiliary power." Note that cycling a plant is different from turning it completely off. It is possible to turn a power plant completely off, but this is typically only done for maintenance purposes -- it takes many hours or even days or more to turn the plant back on. Anyway let's look at those 3 portions of waste during cycling:  

In the figure above you can see the 3 portions labelled 1,2 and 3.

After looking at various reports it appears that during typical use when you total Portions 1,2 and 3, a Coal plant block wastes between 7 and 15%, whereas in a Natural Gas plant wastes between 1 and 10%. A study I found that reported that a Coal plants averaged 7.5% and Natural Gas 3.5%, so I'm going to go with those numbers. Just portion 3 (the auxiliary power) for Coal comes in around 4% and Natural Gas is about 1.5%. As an aside, this means that if you have a Coal plant just sitting there idling in case the sun doesn't shine or the wind doesn't blow, a Coal plant will use 4% of max and a Natural Gas plant 1.5% of max.

This issue is very complex and I'm not going to pretend to have any exact answer. People spend careers modeling stuff like this and modeling it on supercomputers. Why is it so complex? Because there are so many variables involved and each interacts with the others. You have to account for changing demand, the specific generation mix (coal, natural gas, solar, wind), a crazy web of transmission wires, weather, and on and on... But that said, I think we can at least get a "back of the envelope" idea. So let's imagine a couple of scenarios where fossil fuel plants are forced to cycle. We'll start with a Coal plant:

Back in the post "From CA to WV" in the figure "All Coal in Winter" we saw that an all Coal powered grid could accommodate up to 11% Solar without cycling. Let's throw that one up again:

So we've got two figures. In the lower graph there are 42 green dots, whereas in the graph above we have 56 green dots. Therefore by cycling 2 out of 4 Coal blocks this grid was able to accommodate an extra 14 units of Solar generation. Now the total electricity the Coal plant would have made without any Solar would have been 370 dots (ie there are 370 boxes under the double humped black line). The cycling caused a waste of 7.5% in 2 of the 4 Coal blocks. 370 dots x 7.5% = 28, and that divided by 2 (because only 2 of the 4 blocks were cycled) = 14. So the waste was just as much as the Solar! Yikes! Basically what this tells us is having Solar generation above 11% in a nearly all Coal grid (like West Virginia) does not lower carbon emissions. Again, I'm not going to pretend that this is an exact answer, but even if I'm off by 50%, that would cap us at 16.5% Solar.

Next is Nat Gas:

In the post "CA to WV" I showed that an "All Nat Gas in Winter" figure that could accommodate up to 15% Solar generation without cycling. Let's throw that back up:

 In the bottom figure with the 15% Solar there are 56 green dots, and in the top graph with 20% Solar there are 74 dots, a difference of 18 dots of Solar generation. We'll do the same math as above. Without Solar there would have been 370 dots. The cycling caused a waste of 3.5%.  370 x 3.5% = 13 dots. Divide 13 by 2 (as only 2 out of 4 blocks were cycled) and you get 7.5 dots wasted. We got 18 extra dots of carbon free electricity when we increased the Solar from 15 to 20%, but 7.5 dots went to waste because of cycling, so our "real life" gain was 10.5 dots. Another way of saying this is that of the 18 dots of extra solar electricity, 58% went to waste. Again this is just a rough estimate, but even I think we can see that no matter how you slice it the waste is pretty significant. This means that in an all Nat Gas powered grid like California, possibly half of Solar generation above 15% would go to waste. Ugh. 

Finally is our typical 50% Coal / 25% Nat Gas / 15% Nuclear (similar to Ohio):

The figure above shows a Coal plant cycling 3 out of 4 of its blocks. Compare this to:

13% Solar Winter with Curtailment.jpg

The bottom figure shows the max that this grid can accommodate without cycling. In the top graph the cycling allows the grid to accommodate 46 dots of Solar generation, compared to 36 in the second graph, a difference of 10 dots. The total dots from Coal if there had been no Solar at all would have been 240 dots. 240 x 7.5% = 18 dots. 18 x 3/4 (only 3 out of 4 blocks were cycled) = 13.5. Therefore in this scenario we gained 10 dots of carbon free Solar generation but wasted 13.5%! Again this is just an estimate, but I think you can begin to see that the waste from cycling is quite significant.

Basically want I wanted to show was that the issue of utilities whining about cycling their fossil fuel power plants is more than just about money. It also results in significant fossil fuel burned that all goes to waste. As you can see above if we are thinking in terms of minimizing carbon emissions it seems like Coal plants should not be cycled at all. Natural Gas power plants are surely better, but the cycling still results in a lot of waste when Solar generation goes above approximately 18%. But really, whether it is 18% or I'm off and it's more like 25%... we need to go way higher if we are going to go carbon free. So what if you could simply store all that extra electricity?

Bonus section for true Nerdom: as you might imagine I am not the only person to make an attempt at this. At one end of the spectrum there is this study: Katzenstein, W.; Apt, J. (2009). "Air Emissions Due to Wind and Solar Power." Environmental Science and Technology (43); pp.253-258. All I could find was the abstract, but basically this paper argues that cycling leads to so much waste that the carbon emissions saved by Solar and Wind are completely negated. I requested a copy of the paper but never heard anything, so I can't say if the paper is any good or not. So that's one end. On there other hand there is a large and very terrific study done by the National Renewable Energy Laboratory (NREL) called the Western Wind and Solar Integration Study. This is a huge supercomputer type study that looks at integrating various percentages of Solar and Wind into the grid of the Western US. In this study they were able to demonstrate in their model that if the Western USA had 16.5% Solar and 16.5% Wind the carbon emissions saved by the Solar and Wind would dwarf the waste caused by cycling. They achieved a similar conclusion with a mix of 25% Solar and 8% Wind. To achieve this result in their model however, they had to assume that 1) all the new Solar and Wind plants had been placed in optimal locations for transmission, 2) that load forecasts were perfect (ie that grid operators would have perfect knowledge of how much electricity would be needed the next day) and 3) they added an additional 10% of high voltage transmission lines to carry the Solar and Wind to where it needed to be. The cost of these additional 10% of transmission lines they estimated at around $10 billion dollars.

Basically what I took home from the Western Wind and Solar Integration study is that there is a world in which up to 25% Solar can be integrated into the grid without wasting it on cycling... but it's not the world we live in. Maybe (hopefully!) it will be one day. But it's not the reality today.

Interestingly the highest Solar they used was 25%... why is that? That is the subject of the next post.