5

Reuters recently published the following piece:

Nuclear energy too slow, too expensive to save climate: report

Inside, it says:

The cost of generating solar power ranges from $36 to $44 per megawatt hour (MWh), the WNISR said, while onshore wind power comes in at $29–$56 per MWh. Nuclear energy costs between $112 and $189.

It surprises me that the nuclear figures are so high.

Are those figures accurate?

  • Just an other random article about this: world-nuclear.org/information-library/economic-aspects/… The article seems published by a pro-nuclear organization. If you look at the table in the section "NEI 2013 Financial Modelling" you see they claim nuclear is around that cost (they divide in two categories ranging between 85 and 121.90). However they claim wind would be 112.9$/MWh which is more than twice what Reuters claim... – Bakuriu Sep 25 at 21:07
  • 2
    Do they include the price for safely storing the produced waste for many millennia? – Hagen von Eitzen Sep 25 at 21:22
  • @HagenvonEitzen: I don't know. My wild guess is "no" or "partially". – einpoklum - reinstate Monica Sep 25 at 21:48
  • Paging @EnergyNumbers who has a history of writing excellent answers to these sort of questions. – Oddthinking Sep 25 at 22:55
  • @AndrewGrimm I think it is a dupe. But might be worth adding some of the content here into the other question. – matt_black Sep 26 at 11:28
7

The figures seem consistent with most sources:

  • A Lazard 2018 study puts nuclear firmly in that range.
  • A UK government study puts new nuclear construction at 95 GBP per megawatt hour, which sits somewhere in the middle of the range stated by Reuters, depending on the exchange rate of the day.
  • The OpenEI configurable chart trends lower, but only outliers lie below the bottom of Reuters' stated range.

There appears to be a fair amount of agreement on the figures in question, across a number of countries and governing bodies.


Speculation - accounting for fuel disposition after use and decommissioning the plant at end of lifetime adds substantially to the low marginal cost-per-MWh during operation. On top of this, it's hard to think of any utility project with higher up front capital costs than a new utility-scale nuclear power plant, and that factors into the lifetime generation costs as well.

This isn't to say that nuclear might not still outperform the fossil fuel plants (particularly the dirty ones like coal) if their externalities (cleanup, pollution) are taken into account, but it's easier to do cost estimation for nuclear plants because their storage and decommissioning costs are well established at this point.

  • 3
    You also have to factor in that nuclear is available all the time - renewables aren't so for some of the time they are not competing and so a more complex consideration is needed (unleast until much cheaper large batteries are available) – user151019 Sep 26 at 8:08
  • 1
    @Mark - That depends on the renewable. Concentrated solar thermal is available 24/7, and offshore wind is very dependable. – jdunlop Sep 26 at 17:47
  • 1
    @jdunlop - hydro seems pretty steady, as well. The big challenge is the aged infrastructure. It needs to be updated for many reasons, but that is a prerequisite for a truly functional renewable generation model. Then whether the wind is blowing at location X becomes less important than the fact that wind will pretty much always be blowing at some combination of X, Y, and/or Z (or land-based wind, for example). – PoloHoleSet Sep 26 at 21:47
  • @jdunlop: IIANM, coal-fired plants can be dramatically less dirty with gassification technology. – einpoklum - reinstate Monica Sep 26 at 21:55
  • @einpoklum - Potentially true, but then you have the cost of gassification, you're still dealing with carbon emissions and nitrogen oxides (smog), and you have to dispose of the slag/fly ash/sulphur dioxiode somewhere as well. Plus, of course, the damage done to the environment by coal mining. – jdunlop Sep 27 at 0:28
-4

It is not really possible to come to a conclusion. For example:

http://www.windfarmbop.com/category/roads/

  1. A large share of wind costs is in the roads. This is often provided as public infrastructure and not LCOE.

https://www.google.com/amp/s/cleantechnica.com/2011/11/19/solar-profit-margins-compared-to-other-industries/amp/

https://www.atkearney.it/energy/article?/a/solar-power-and-india-s-energy-future

https://www.world-nuclear.org/information-library/economic-aspects/economics-of-nuclear-power.aspx

  1. Solar costs are an apples to oranges comparison. Approximately half of solar LCOE is just materials. Therefore the room for labor automation is fairly minimal. On the other hand, 12% of nuclear plant construction costs are materials, and this falls further if operating labor costs are included. So the automation potential may be lower for solar.

  2. Raw materials costs fluctuate wildly year to year. You could achieve pretty much any conclusion by selecting the appropriate year.

  3. Lastly, all estimates of any cost are from interest groups or consulting companies and are therefore useless.

So published costs are simply not relevant. At best, you can ask for specific estimates of individual sites. Saying "all X is better than all Y" in this context is an extreme statement, and quite frankly absurd.

  • 3
    Are you really arguing that the cost of roads should be incorporated into the LCOE for wind? But somehow not for other sources of power? – Fred Stark Sep 26 at 5:50
  • Your fourth point is particularly, egregiously, bad. Governmental bodies charged with impartiality are not "interest groups". Utilities' statements of cost are not "useless". To make a blanket statement like that smacks of paranoia. – jdunlop Sep 26 at 18:14
  • For example, dunlops government source is actually a private company (nera consulting) – Isa Wal Sep 26 at 18:30
  • @jdunlop if you can show me a utilities estimate that applies for power plants in general and not a single one off project I will change my answer. – Isa Wal Sep 27 at 2:40

Not the answer you're looking for? Browse other questions tagged .