5 ms·

Last week Elon explained that the problem of fusion is that it will be more expensive than solar and wind.

DennisP 4y ago
The real question is whether it's more expensive than solar/wind plus the batteries required for dispatchable power. The answer depends on the reactor design, which Elon doesn't know. There are a lot of possibilities, and if the first ones are too expensive, later ones might not be.
- gitfan86 4y ago
  True, but that is a moving target. https://ourworldindata.org/battery-price-decline https://ourworldindata.org/battery-price-decline
- ncmncm 4y ago
  Cost for energy storage is plummeting even faster than for solar or wind ever did.
- pfdietz 4y ago
  Almost certainly it will be more expensive than renewables + storage (note the importance of hydrogen there.) The cost of renewables + storage to provide "synthetic baseload" is likely less than fission. And DT fusion is likely more expensive than fission. https://model.energy/ https://model.energy/
  - DennisP 4y ago
    Yes, we discussed that pretty extensively the other week. It starts out with unrealistic assumptions for fission, including a 25-year lifespan (compared to the actual average age of US reactors of forty years) and twice the discount rate for fission as for everything else. Fix those things and nuclear does better. It also uses US fission prices, which are higher than in most of the world. On top of that, the US has unusually good geography for both wind and solar. Nuclear is much more competitive in many other countries. One ironic example is Germany, where the model suggests a 100% nuclear grid even at US nuclear costs, as long as the lifespan and discount rate are corrected.
    - pfdietz 4y ago
      And as I told you last week, the numbers are actually optimistic compared to US (or actual EPR) numbers. In particular, it assumed the reactor cost 6000 euro per kW, which is considerably less than what EPR ended up costing.
      - DennisP 4y ago
        And I linked you a paper surveying global nuclear costs, showing much lower costs in many areas and no overall trend of increasing costs. Here's the paper again: https://www.sciencedirect.com/science/article/pii/S0301421516300106 https://www.sciencedirect.com/science/article/pii/S030142151... Skip down to Figure 12 for a quick summary.
        pfdietz 4y ago
        Most of those points that were low were either France or Korea. France with EPR has lost whatever mojo they claim to have had (those earlier cost figures are not verifiable) and Korea turned out to achieve low costs by corruption and omitting safety features.
        DennisP 4y ago
        The lowest, sure, but those are around $2000/kW. The bulk of the points on the graph are still below $4000/kW. That includes Japan, India, and Canada, along with a lot of early US reactors. Also, do you have a source for early French costs being unverifiable? I don't see where that's mentioned in the paper, so it doesn't seem right to just dismiss the data unless there's another source giving us good reason to do that. As for the EPR, seems to me the conclusion there is just to not build any more EPRs.
    - hansthehorse 4y ago
      In America the NRC issues a 40 year license to a reactor which then, after a review, can be extended. 40 years is the lifespan of the initial license not the plant.
ncmncm 4y ago
Like the stopped clock, even Elon is right once in a long while.

The real question is whether it's more expensive than solar/wind plus the batteries required for dispatchable power. The answer depends on the reactor design, which Elon doesn't know. There are a lot of possibilities, and if the first ones are too expensive, later ones might not be.

gitfan86 4y ago
True, but that is a moving target. https://ourworldindata.org/battery-price-decline https://ourworldindata.org/battery-price-decline
ncmncm 4y ago
Cost for energy storage is plummeting even faster than for solar or wind ever did.
pfdietz 4y ago
Almost certainly it will be more expensive than renewables + storage (note the importance of hydrogen there.) The cost of renewables + storage to provide "synthetic baseload" is likely less than fission. And DT fusion is likely more expensive than fission. https://model.energy/ https://model.energy/
- DennisP 4y ago
  Yes, we discussed that pretty extensively the other week. It starts out with unrealistic assumptions for fission, including a 25-year lifespan (compared to the actual average age of US reactors of forty years) and twice the discount rate for fission as for everything else. Fix those things and nuclear does better. It also uses US fission prices, which are higher than in most of the world. On top of that, the US has unusually good geography for both wind and solar. Nuclear is much more competitive in many other countries. One ironic example is Germany, where the model suggests a 100% nuclear grid even at US nuclear costs, as long as the lifespan and discount rate are corrected.
  - pfdietz 4y ago
    And as I told you last week, the numbers are actually optimistic compared to US (or actual EPR) numbers. In particular, it assumed the reactor cost 6000 euro per kW, which is considerably less than what EPR ended up costing.
    - DennisP 4y ago
      And I linked you a paper surveying global nuclear costs, showing much lower costs in many areas and no overall trend of increasing costs. Here's the paper again: https://www.sciencedirect.com/science/article/pii/S0301421516300106 https://www.sciencedirect.com/science/article/pii/S030142151... Skip down to Figure 12 for a quick summary.
      - pfdietz 4y ago
        Most of those points that were low were either France or Korea. France with EPR has lost whatever mojo they claim to have had (those earlier cost figures are not verifiable) and Korea turned out to achieve low costs by corruption and omitting safety features.
        DennisP 4y ago
        The lowest, sure, but those are around $2000/kW. The bulk of the points on the graph are still below $4000/kW. That includes Japan, India, and Canada, along with a lot of early US reactors. Also, do you have a source for early French costs being unverifiable? I don't see where that's mentioned in the paper, so it doesn't seem right to just dismiss the data unless there's another source giving us good reason to do that. As for the EPR, seems to me the conclusion there is just to not build any more EPRs.
  - hansthehorse 4y ago
    In America the NRC issues a 40 year license to a reactor which then, after a review, can be extended. 40 years is the lifespan of the initial license not the plant.

ncmncm 4y ago

Like the stopped clock, even Elon is right once in a long while.