3 ms·

tl;dw: Sabine Hossenfelder explains that public is being misleaded by fusion industry touting "efficiency" as ratio of energy that goes only directly out of/int

by rini17 5y ago

tl;dw: Sabine Hossenfelder explains that public is being misleaded by fusion industry touting "efficiency" as ratio of energy that goes only directly out of/into plasma. This is not total consumption vs. usable energy output of whole plant as you might expect. That one might well be 100 times less.

willis936 5y ago
>That one might well be 100 times less. It might be, but it isn't. Turn to page 170 of your copy of The Future of Fusion Energy. It goes through typical efficiencies of each subsystem then concludes: >"Multiplying these conversion factors (i.e. eta h&cd = 0.4 from the heating/current drive systems, eta steam = 0.4 from the steam cycle, and f blanket = 1.3 from the blanket), we see that Qeng is roughly a factor of 5 smaller than Q of the plasma." What's more is that due to self-heating, increases in Q becomes exponential. The distance (in terms of machine performance) from Q=1 to Q=infinity is much smaller than the distance from Q=0.001 to Q=1. So why are scientists lying to the public? The answer appears that they are not, but they simply are not being asked to explain the entire history and projected designs of fusion reactors. You need more than a TED talk or an audience who at least knows anything about physical sciences in order to even start that.
- rini17 5y ago
  If anyone gets that kind of money, is merely "they were not asked to explain" good excuse for you? Anyway, what kind of error bars do these numbers have? We still aren't even at Qplasma = 1. As Sabine says it went from 0.67 to 0.7 in 24 years. For me that's clearly too soon to talk about "typical setup".
  - willis936 5y ago
    That was a misrepresentation by her. NIF is not MCF, it's an ICF weapons program. NIF also gets healthy funding while MCF does not. Look at PPPL's FIRE 1976 report [0]. We've been on logic 0. Given that, the progress we've made with the discovery of H-mode, integration of superconductors, MHD optimization and control, and now empirical turbulence modeling is wonderful. These advances are reflected in simultion and in real machines as triple product increases. True nuclear machines are expensive. We could have made a Q = 50 machine in 1990. We didn't. Want to fund it? If she wanted to actually show the progress of MCF performance she would have plotted triple product vs. time, where gains have been made faster than Moore's Law when Moore's Law still applied, for a period that began before Moore's Law and is continuing past Moore's Law. 0. https://fire.pppl.gov/us_fusion_plan_1976.pdf https://fire.pppl.gov/us_fusion_plan_1976.pdf
    - rini17 5y ago
      Not a VC myself but we're on startup funding platform here so I encourage you, go pitch it.
- simonh 5y ago
  Please watch Sabine’s video. She gives numerous examples where blatantly misleading statements are made by fusion scientists.
  - willis936 5y ago
    I did watch her video. More than that I had already read the entire transcripts of a few of the DoE Fusion Energy Sciences congressional testimonies[0], where the Bigot excerpt was pulled from (or rather Sabine pulled them from another video with an agenda that pulled them from somewhere I don't know). I didn't draw my conclusions from 3 second supercuts of testimonies or TED Talks presented by someone who's making conclusions for me. Read the testimonies. Read the questions given to them by congress. Then you're at least partially equipped to pass judgment on whether or not Bigot was being deceitful. To be fully equipped you'll need to learn the history of fusion energy research and how the DoE conducts its communications. 0. https://docs.house.gov/meetings/SY/SY20/20160420/104815/HHRG-114-SY20-20160420-SD003.pdf https://docs.house.gov/meetings/SY/SY20/20160420/104815/HHRG...
Robotbeat 5y ago
I guess so, but these plants aren't meant to produce electricity. You'd build them totally differently if they were. Achieving ignition is a different kind of goal, and it's something that OUGHT to be pursued without trying to make the whole thing be a net positive producer of electricity. In reality, the only crime here is that fusion progress is just bloody and unnecessarily slow, and ITER folk seem to not care.

by rini17 5y ago

willis936 5y ago

>That one might well be 100 times less. It might be, but it isn't. Turn to page 170 of your copy of The Future of Fusion Energy. It goes through typical efficiencies of each subsystem then concludes: >"Multiplying these conversion factors (i.e. eta h&cd = 0.4 from the heating/current drive systems, eta steam = 0.4 from the steam cycle, and f blanket = 1.3 from the blanket), we see that Qeng is roughly a factor of 5 smaller than Q of the plasma." What's more is that due to self-heating, increases in Q becomes exponential. The distance (in terms of machine performance) from Q=1 to Q=infinity is much smaller than the distance from Q=0.001 to Q=1. So why are scientists lying to the public? The answer appears that they are not, but they simply are not being asked to explain the entire history and projected designs of fusion reactors. You need more than a TED talk or an audience who at least knows anything about physical sciences in order to even start that.

rini17 5y ago
If anyone gets that kind of money, is merely "they were not asked to explain" good excuse for you? Anyway, what kind of error bars do these numbers have? We still aren't even at Qplasma = 1. As Sabine says it went from 0.67 to 0.7 in 24 years. For me that's clearly too soon to talk about "typical setup".
- willis936 5y ago
  That was a misrepresentation by her. NIF is not MCF, it's an ICF weapons program. NIF also gets healthy funding while MCF does not. Look at PPPL's FIRE 1976 report [0]. We've been on logic 0. Given that, the progress we've made with the discovery of H-mode, integration of superconductors, MHD optimization and control, and now empirical turbulence modeling is wonderful. These advances are reflected in simultion and in real machines as triple product increases. True nuclear machines are expensive. We could have made a Q = 50 machine in 1990. We didn't. Want to fund it? If she wanted to actually show the progress of MCF performance she would have plotted triple product vs. time, where gains have been made faster than Moore's Law when Moore's Law still applied, for a period that began before Moore's Law and is continuing past Moore's Law. 0. https://fire.pppl.gov/us_fusion_plan_1976.pdf https://fire.pppl.gov/us_fusion_plan_1976.pdf
  - rini17 5y ago
    Not a VC myself but we're on startup funding platform here so I encourage you, go pitch it.
simonh 5y ago
Please watch Sabine’s video. She gives numerous examples where blatantly misleading statements are made by fusion scientists.
- willis936 5y ago
  I did watch her video. More than that I had already read the entire transcripts of a few of the DoE Fusion Energy Sciences congressional testimonies[0], where the Bigot excerpt was pulled from (or rather Sabine pulled them from another video with an agenda that pulled them from somewhere I don't know). I didn't draw my conclusions from 3 second supercuts of testimonies or TED Talks presented by someone who's making conclusions for me. Read the testimonies. Read the questions given to them by congress. Then you're at least partially equipped to pass judgment on whether or not Bigot was being deceitful. To be fully equipped you'll need to learn the history of fusion energy research and how the DoE conducts its communications. 0. https://docs.house.gov/meetings/SY/SY20/20160420/104815/HHRG-114-SY20-20160420-SD003.pdf https://docs.house.gov/meetings/SY/SY20/20160420/104815/HHRG...

Robotbeat 5y ago

I guess so, but these plants aren't meant to produce electricity. You'd build them totally differently if they were. Achieving ignition is a different kind of goal, and it's something that OUGHT to be pursued without trying to make the whole thing be a net positive producer of electricity. In reality, the only crime here is that fusion progress is just bloody and unnecessarily slow, and ITER folk seem to not care.