2 ms·

I was not reasoning by analogy, I presented an example that contradicts the claim that machines meant for widespread adoption must resemble their experimental f

by jjk166 5y ago

I was not reasoning by analogy, I presented an example that contradicts the claim that machines meant for widespread adoption must resemble their experimental form. The laws of physics are unchanged, computers improved because of clever workarounds that allowed us to accomplish the same tasks in easier ways. Practical fusion reactors need to produce power via fusion, but beyond that the how is irrelevant. Swap out different lasers, different magnets, different power supplies, hell even the fuels might be different, it's frankly ignorant to claim that these changes are the same physics but "use a shorter wavelength in your fab" isn't. A modern computer in terms of number of logic gates is astronomically more complex than ENIAC, however what matters for practicality is not complexity but cost. The cost of complexity decreases as you learn what you're doing. Chicago Pile 1 cost $17 million in today's dollars to produce about half a watt of thermal power, or $34,000,000/Wth. A modern nuclear plant costs about $2/Wth. ITER will cost $45 Billion to generate 500 MW. If fusion sees 0.001% of the economic improvement that fission experienced, that's $0.52/Wth.

pfdietz 5y ago
The conclusion that DT fusion reactors will be expensive follows from basic engineering considerations on heat transfer, not from specifics of any design. The volumetric power density will be lousy, so cost will be high. > A modern nuclear plant costs about $2/Wth This may have been the projection before the AP1000 and the EPR flamed out. In practice, the cost can escalate well above that. Their complexity made them difficult to build. And yet, these designs have reactor power densities orders of magnitude better than ITER, and are far simpler (and more fault tolerant) than a fusion reactor would be.
- jjk166 5y ago
  The cost is not a function of volumetric power density. I believe we have had this conversation before. The most expensive nuclear plant projected to be constructed is the Hinkley Point C reactor. With all its overruns, it's projected that the cost will be £23 Billion, which works out to $3.37/Wth. $2/Wth is a $6 billion 1000 MWe plant, quite typical. Note that this is for reactors in the west, in China, where non-technical issues like NIMBYism aren't a concern, the cost is closer to $1/Wth.
  - pfdietz 5y ago
    > The cost is not a function of volumetric power density. I believe we have had this conversation before. Yes, and you failed to make the case that larger things aren't more expensive than smaller things (all else being equal, but note that fission reactors are made of steel, not the complex sophisticated equipment of fusion reactors). We can continue the argument here, if you can come up with any argument for your position that makes any sense.

by jjk166 5y ago

pfdietz 5y ago

The conclusion that DT fusion reactors will be expensive follows from basic engineering considerations on heat transfer, not from specifics of any design. The volumetric power density will be lousy, so cost will be high. > A modern nuclear plant costs about $2/Wth This may have been the projection before the AP1000 and the EPR flamed out. In practice, the cost can escalate well above that. Their complexity made them difficult to build. And yet, these designs have reactor power densities orders of magnitude better than ITER, and are far simpler (and more fault tolerant) than a fusion reactor would be.

jjk166 5y ago
The cost is not a function of volumetric power density. I believe we have had this conversation before. The most expensive nuclear plant projected to be constructed is the Hinkley Point C reactor. With all its overruns, it's projected that the cost will be £23 Billion, which works out to $3.37/Wth. $2/Wth is a $6 billion 1000 MWe plant, quite typical. Note that this is for reactors in the west, in China, where non-technical issues like NIMBYism aren't a concern, the cost is closer to $1/Wth.
- pfdietz 5y ago
  > The cost is not a function of volumetric power density. I believe we have had this conversation before. Yes, and you failed to make the case that larger things aren't more expensive than smaller things (all else being equal, but note that fission reactors are made of steel, not the complex sophisticated equipment of fusion reactors). We can continue the argument here, if you can come up with any argument for your position that makes any sense.