The Wendelstein 7-X concept proves its efficiency(ipp.mpg.de)
ipp.mpg.de
The Wendelstein 7-X concept proves its efficiency
https://www.ipp.mpg.de/5125328/05_21
40 comments
Stellarator theory has been booming for the past 20-40 years, but very few machines have been made to push it forward. I'd love to see a mid-scale (sub 1 BnUSD) HTS machine be built as a platform to develop coil winding techniques and test turbulent transport models.
If I remember correctly, you replied to one of my comments on fusion power in the past and that you are actually working in research yourself.
Most fusion projects seem to take years between major steps. Do you think a rapid-iteration approach (like what SpaceX is doing to develop Starship) could work for fusion research?
Most fusion projects seem to take years between major steps. Do you think a rapid-iteration approach (like what SpaceX is doing to develop Starship) could work for fusion research?
I'm stepping out to appease the two-body problem.
My impression is that a Manhattan Project for fusion would be successful. We already have all the technology needed to make a successful reactor. The reason things like HTS coil winding are moving so slowly to scale up is because there is no market for it. How could anyone invest in something with no return?
If a lot of money did suddenly become available then innovative and radical platforms could be seriously considered. Things like stellarator platforms designed with adjustable geometry coils to explore many different optimizations without needing to build a new machine every time would be on the table.
Issues related to first-wall and divertors require large machines to develop. They are not unsolvable problems, but those are the ones that need a serious amount of money to test and develop. Theory is way ahead of experiment here. I think it would be easier to find the money for these big machines if the performance metrics of smaller machines continued to increase. The problem is there isn't enough money to keep these new small-to-medium machines coming. Look at the history of FES spending in the DoE budget.
SPARC might be the community's best chance at convincing the public that fusion is worth investing in.
My impression is that a Manhattan Project for fusion would be successful. We already have all the technology needed to make a successful reactor. The reason things like HTS coil winding are moving so slowly to scale up is because there is no market for it. How could anyone invest in something with no return?
If a lot of money did suddenly become available then innovative and radical platforms could be seriously considered. Things like stellarator platforms designed with adjustable geometry coils to explore many different optimizations without needing to build a new machine every time would be on the table.
Issues related to first-wall and divertors require large machines to develop. They are not unsolvable problems, but those are the ones that need a serious amount of money to test and develop. Theory is way ahead of experiment here. I think it would be easier to find the money for these big machines if the performance metrics of smaller machines continued to increase. The problem is there isn't enough money to keep these new small-to-medium machines coming. Look at the history of FES spending in the DoE budget.
SPARC might be the community's best chance at convincing the public that fusion is worth investing in.
>The reason things like HTS coil winding are moving so slowly to scale up is because there is no market for it. How could anyone invest in something with no return?
This seems like an odd question. You seriously don't think people invest in businesses with no profits, or even revenue? In 2021? One of the financial trends recently has been investing in blank check companies that don't even have a business let alone revenue or profits. Also the biggest investing celebrity recently, Cathie Wood, is known for buying "innovative" companies based on their far future prospects, and has reaped spectacular returns so far.
There's a joke about two economists, where one of them sees a twenty dollar bill on the ground and says so, and the other says, "can't be, someone would've picked it up already".
Practical fusion power is like a multitrillion dollar bill on the ground and it needs some really creative explanation why it wouldn't be picked up.
This seems like an odd question. You seriously don't think people invest in businesses with no profits, or even revenue? In 2021? One of the financial trends recently has been investing in blank check companies that don't even have a business let alone revenue or profits. Also the biggest investing celebrity recently, Cathie Wood, is known for buying "innovative" companies based on their far future prospects, and has reaped spectacular returns so far.
There's a joke about two economists, where one of them sees a twenty dollar bill on the ground and says so, and the other says, "can't be, someone would've picked it up already".
Practical fusion power is like a multitrillion dollar bill on the ground and it needs some really creative explanation why it wouldn't be picked up.
I think you've just identified a lucrative research field. Measuring progress towards practical fusion power plants in terms of engagement.
There is clearly capital available for things other than social media.
Setting aside all the biotech IPOs:
"SpaceX has raised a total of $6.6B in funding over 46 rounds. Their latest funding was raised on Jul 30, 2021 from a Secondary Market round."
https://www.crunchbase.com/organization/space-exploration-te...
Searching crunchbase for companies in the energy industry that have raised > $1B returns 24 results, too.
Setting aside all the biotech IPOs:
"SpaceX has raised a total of $6.6B in funding over 46 rounds. Their latest funding was raised on Jul 30, 2021 from a Secondary Market round."
https://www.crunchbase.com/organization/space-exploration-te...
Searching crunchbase for companies in the energy industry that have raised > $1B returns 24 results, too.
Investing in hardware companies has not been as fashionable though.
And fusion is hardest of the wares.
And fusion is hardest of the wares.
As I see it the problem is also in approach to big new projects. Usually you have long careful preparation to build something about what you have no clear idea how to do it, since it was not build before. But investors, in this case usually government, wants to see some classical project, with schedule / waterfall but in reality it should be more, as what mentioned SpaceX is doing. Which as I understand is fast iteration on unclear solution with clear goal. But I can understand that investors prefer to have ordinary projects as it is more classic approach. The issue is the way it works every time is the same. Project starts everything is fine. Some small delays on milestones but OK, then someone finds out that state of project is actually much worse than expected. Director is fired and the cycle starts again. On the end there is somehow working somehow done project.
It's not just investors, it's government too. Remember they are stewards of money taken from the people; if they were to give it to some crazy person who turned out to be a sham, it would be very bad (and that happens too).
What do you think about ITER?
It's an exciting project. Breakeven isn't some magical barrier that we somehow don't know how to break, so it's headline milestones are not what excite me. It will be a platform to really look at divertors. We'll see just how the first wall at the tiny strike points in a large machine age. We'll also learn a lot about burning confined plasmas.
After working with a 50 kW gyrotron and 1 m major radius machine, the raw industrial scale of the project is awe-inspiring [0]. ITER has 20x 1 MW gyrotrons and a 6 m major radius. Every system in the machine is fascinating.
All that said, it is a slow and expensive project. The issues are mostly political. The design is very old at this point. All of this is fine, but in the context of potentially scaling up HTS supply chains to make a reactor that is much smaller, it becomes clear that the ITER path is not the one the first generation reactors will take. It's an important stepping stone in terms of plasma science and public opinion. It's going to be a lot of fun to follow its successes.
0. https://youtu.be/cvsj0nqHwic
After working with a 50 kW gyrotron and 1 m major radius machine, the raw industrial scale of the project is awe-inspiring [0]. ITER has 20x 1 MW gyrotrons and a 6 m major radius. Every system in the machine is fascinating.
All that said, it is a slow and expensive project. The issues are mostly political. The design is very old at this point. All of this is fine, but in the context of potentially scaling up HTS supply chains to make a reactor that is much smaller, it becomes clear that the ITER path is not the one the first generation reactors will take. It's an important stepping stone in terms of plasma science and public opinion. It's going to be a lot of fun to follow its successes.
0. https://youtu.be/cvsj0nqHwic
How much money is needed?
For a successful fusion power crash program?
This 45-year old analysis [0] says that 15 Bn USD 1978 (63 Bn USD 2021) stands between us and a first gen reactor. Afaict that is still accurate as we have been on the Logic I path since then. How quickly you want to spend that 63 Bn USD dictates what logic (II-V) you're on.
0. https://fire.pppl.gov/us_fusion_plan_1976.pdf
This 45-year old analysis [0] says that 15 Bn USD 1978 (63 Bn USD 2021) stands between us and a first gen reactor. Afaict that is still accurate as we have been on the Logic I path since then. How quickly you want to spend that 63 Bn USD dictates what logic (II-V) you're on.
0. https://fire.pppl.gov/us_fusion_plan_1976.pdf
I was confused but looked things up so correct me if I'm wrong:
* W-7X uses 3T magnets (traditional type-II superconductors), so it has to be big (to give more room for plasma instability to not hit the walls?)
* ARC can be smaller because it uses 21T HTS magnets, so the plasma can be more tightly confined.
* LTS magnets have an abrupt phase transition. Imperfections have resistance, which causes heat, which can quench the magnet if current gets too high. (?)
* HTS magnets for - reasons? - don't quench like this. The magnet just gets hot and slowly degrades.
* HTS magnets are really brittle and finicky to work with, so winding them without causing imperfections is hard.
Fusion aside, wouldn't 21T magnets be valuable in a lot of areas? Like couldn't you do really high res MRIs with such a magnet? I don't see why there's not more commercial interest in such a general technology.
* W-7X uses 3T magnets (traditional type-II superconductors), so it has to be big (to give more room for plasma instability to not hit the walls?)
* ARC can be smaller because it uses 21T HTS magnets, so the plasma can be more tightly confined.
* LTS magnets have an abrupt phase transition. Imperfections have resistance, which causes heat, which can quench the magnet if current gets too high. (?)
* HTS magnets for - reasons? - don't quench like this. The magnet just gets hot and slowly degrades.
* HTS magnets are really brittle and finicky to work with, so winding them without causing imperfections is hard.
Fusion aside, wouldn't 21T magnets be valuable in a lot of areas? Like couldn't you do really high res MRIs with such a magnet? I don't see why there's not more commercial interest in such a general technology.
* W7-X has a large major radius because the scaling laws work out favorably for large machines but there have been very few large machines built to continue pushing out the domain of the empirical scaling laws. Afaik there is not a strong relationship between minor radius and risk of disruption.
One of the selling points of stellarators is that they can be operated without risk of disruption, but W7-X has done campaigns that push its operation to disrupt. A machine with dimensions similar to W7-X could be made with higher field magnets.
* Yes, higher field means larger machines are not needed to hit the same performance metrics, pushing down the minimum size (and cost) of a reactor. REBCO (HTS material of choice) is quite an expensive material to procure and work, so even modest sized research devices are more expensive than Uncle Sam is willing to gift right now. Private investment into SPARC is changing my opinion about what the future of fusion research will look like.
* The details of working HTS into confinement coils is an active area of discussion. HTS coils will be operated at liquid helium temperatures but won't stop superconducting until liquid nitrogen temperatures. The critical current limit decreases as temperature increases, so the risk of quench is still there when operated at near-boundary conditions. One thing that is discussed is whether to insulate the HTS tape or not. Conventionally, coils have insulation between the windings. In the case of superconductors it isn't strictly necessary. Superconductors are embedded in large copper jackets. When the core is superconducting the copper acts as a virtual open. When the core quenches the coil turns into a one turn copper ring and can dump all of the stored energy through the leads quickly without boiling cryogenics because the inductance is much lower.
High field magnets are useful, just not to many areas. MRI is the singular commercial application and they are more interested in lower cost (ie being able to run a superconducting magnet without liquid helium) than they are a higher field. NMR is the other big (but actually quite small) application for high field magnets. There are commercially available NMRs today with fields of 50 T.
One of the selling points of stellarators is that they can be operated without risk of disruption, but W7-X has done campaigns that push its operation to disrupt. A machine with dimensions similar to W7-X could be made with higher field magnets.
* Yes, higher field means larger machines are not needed to hit the same performance metrics, pushing down the minimum size (and cost) of a reactor. REBCO (HTS material of choice) is quite an expensive material to procure and work, so even modest sized research devices are more expensive than Uncle Sam is willing to gift right now. Private investment into SPARC is changing my opinion about what the future of fusion research will look like.
* The details of working HTS into confinement coils is an active area of discussion. HTS coils will be operated at liquid helium temperatures but won't stop superconducting until liquid nitrogen temperatures. The critical current limit decreases as temperature increases, so the risk of quench is still there when operated at near-boundary conditions. One thing that is discussed is whether to insulate the HTS tape or not. Conventionally, coils have insulation between the windings. In the case of superconductors it isn't strictly necessary. Superconductors are embedded in large copper jackets. When the core is superconducting the copper acts as a virtual open. When the core quenches the coil turns into a one turn copper ring and can dump all of the stored energy through the leads quickly without boiling cryogenics because the inductance is much lower.
High field magnets are useful, just not to many areas. MRI is the singular commercial application and they are more interested in lower cost (ie being able to run a superconducting magnet without liquid helium) than they are a higher field. NMR is the other big (but actually quite small) application for high field magnets. There are commercially available NMRs today with fields of 50 T.
That seems to be the approach that MIT is taking with the SPARC reactor.
They've been working on their ARC reactor design, and decided they wanted a smaller, easy-to-make prototyping platform that they can run tests with. If I understand correctly, SPARC isn't designed for long-term use, or practical energy extraction. If they destroy one, they'll just make another.
If I remember right, they're working on re-mountable magnet coils for ARC so that you can just de-solder the windings and take the whole thing apart and replace the inner lining when necessary. I don't think SPARC has that.
https://en.wikipedia.org/wiki/ARC_fusion_reactor
https://en.wikipedia.org/wiki/SPARC_(tokamak)
They've been working on their ARC reactor design, and decided they wanted a smaller, easy-to-make prototyping platform that they can run tests with. If I understand correctly, SPARC isn't designed for long-term use, or practical energy extraction. If they destroy one, they'll just make another.
If I remember right, they're working on re-mountable magnet coils for ARC so that you can just de-solder the windings and take the whole thing apart and replace the inner lining when necessary. I don't think SPARC has that.
https://en.wikipedia.org/wiki/ARC_fusion_reactor
https://en.wikipedia.org/wiki/SPARC_(tokamak)
How is energy extracted from the reaction? Once you get a bunch of energetic particles flying around, how to do you get from there to electricity?
Same way it is pretty much everywhere, including in nuclear power plants. Heat water and use the steam to run a turbine.
I have always wondered, is heat->steam->turbine->generator efficient? In my mind it just doesn't seem like it would be efficient at all. Surely there must be other ways to convert heat to electricity (Thermoelectric effect?), but why do we still use steam?
The thermal efficiency of turbines can approach 90%, which is far and away better than any other method we have. Turning energy into heat is usually not a problem.
And in a lot of cases extra heat can be put to use beyond electricity production. Steam heat and cooling is a thing for large metropolitan areas, and it can also be used for desalination water. A few Soviet nuclear plants were designed to provide both electricity and water, actually.
Peltiers and other thermoelectric generation is really really inefficient compared to basically any other energy capture method. They're about 5-15% max that we know how to make currently. Basically they only get used for things like RTGs where you need a zero maintenance very long lasting electric source; in the US that's mostly been spacecraft but in the USSR they were used for extremely remote lighthouses around the Arctic Circle where solar was impractical and resupply (and maintenance) of a diesel generator was prohibitively expensive.
Any engine is limited to the carnot cycle; the themoelectric effect still requires a cool side and heat gets conducted along the two leads to equalize the temperature. That's where the power comes from, the temperature differential across the thermocouple wires, and you consume the temperature differential when you generate electricity.
Different types of engines have different details of where that energy gets lost, but ultimately they are devices that if left to run indefinitely would equilibrate in some way and the way that they idle tells you a fair bit about where the energy flows are.
Turbines are pretty impressively efficient, but something can always be better. They often utilize the low grade waste heat in steam heating though, so a lot of times the efficiency is extremely good when you include free heating or process heat (in a plant they might use that steam to heat another piece of equipment).
Different types of engines have different details of where that energy gets lost, but ultimately they are devices that if left to run indefinitely would equilibrate in some way and the way that they idle tells you a fair bit about where the energy flows are.
Turbines are pretty impressively efficient, but something can always be better. They often utilize the low grade waste heat in steam heating though, so a lot of times the efficiency is extremely good when you include free heating or process heat (in a plant they might use that steam to heat another piece of equipment).
Current thermoelectric tech has very low efficiencies. It's an active area of research where any small improvement could be significant. It would be an ideal way to produce power more or less directly from thermal neutron bombardment, IF a method / material can be found that can make direct thermoelectric conversion efficient enough to compete with a water or molten salt heat exchanger loop.
> ...why do we still use steam?
It hits the sweet spot on fiscal, engineering, safety, and science fronts. Energy extraction is very challenging, we're likely hundreds of years away from aneutronic fusion, for example. The research area you are asking about is direct energy conversion, of which aneutronic fusion is one small branch (though within it, there are many scientific and engineering branches to explore).
It hits the sweet spot on fiscal, engineering, safety, and science fronts. Energy extraction is very challenging, we're likely hundreds of years away from aneutronic fusion, for example. The research area you are asking about is direct energy conversion, of which aneutronic fusion is one small branch (though within it, there are many scientific and engineering branches to explore).
I think I may have said this before, but a statement like "we're likely hundreds of years away from aneutronic fusion" seems to me necessarily meaningless.
We're only a bit over a hundred years from developing quantum mechanics and relativity, how can we possibly say anything about hundreds of years from now?
If something is well enough understood to accurately predict the timeline, we could do it much sooner.
If we have no idea how to do something, then "hundreds of years" means nothing, except maybe "not proven impossible yet".
We're only a bit over a hundred years from developing quantum mechanics and relativity, how can we possibly say anything about hundreds of years from now?
If something is well enough understood to accurately predict the timeline, we could do it much sooner.
If we have no idea how to do something, then "hundreds of years" means nothing, except maybe "not proven impossible yet".
I’ve long wondered this too. Intuitively, it just seems like the losses from all those energy transfers would be unacceptably great.
I think it basically heats up the walls of the chamber, from which you can use the heat to make electricity in the usual way.
I don’t know if stellarators have a different plan, but in a production tokamak, it’s a combination of:
- deuterium-tritium fusion, which produces lots of high-energy neutrons
- the neutrons help to keep the plasma hot, but being neutral, they escape the magnetic bottle; there is a lithium blanket around the reactor to catch them
- the neutrons that hit the lithium produce tritium for the reactor, and steam for the turbines
- deuterium-tritium fusion, which produces lots of high-energy neutrons
- the neutrons help to keep the plasma hot, but being neutral, they escape the magnetic bottle; there is a lithium blanket around the reactor to catch them
- the neutrons that hit the lithium produce tritium for the reactor, and steam for the turbines
Wasn't there a fusion project where the end product was a charged particle so energy could be extracted from currents generated by changes in the magnetic field as they slowed down?
Here's one such promising project: https://lppfusion.com/technology/fusion-energy-generator/
"Dense Plasma Focus device with aneutronic hydrogen-boron (pB11) fuel"
"Dense Plasma Focus device with aneutronic hydrogen-boron (pB11) fuel"
As I've said before I love this project. The engineering is top notch and they are just knocking down the unknowns one after another.
It is for me a wonderful example of how a large engineering project to build a new thing should be approached. Lay out all of the questions that are currently unanswered for which the answer will affect the next step. Then start building with the goal of answering the questions in dependency order so that the next question/build can incorporate your new understanding given the answer to the previous question. Iterate until you've answered all the engineering questions and you're sitting there looking at a fully functional device that does this new thing.
It is for me a wonderful example of how a large engineering project to build a new thing should be approached. Lay out all of the questions that are currently unanswered for which the answer will affect the next step. Then start building with the goal of answering the questions in dependency order so that the next question/build can incorporate your new understanding given the answer to the previous question. Iterate until you've answered all the engineering questions and you're sitting there looking at a fully functional device that does this new thing.
Projects like this make me respect institutional knowledge so much. I can't even imagine having to start from scratch with something like this, and these projects have decades of it under their belt. It's such an important part of our knowledge as humans.
Gene Kranz book could be up your alley. Not too technical though but I am going through it now and it’s really exciting to go through NASA’s incremental steps to the moon
Good day for fusion tech it seems. Two stories on hn for different projects
For those of you who speak German. There was a great episode of Alternativlos where they talk about details of the Wendelstein 7-X Projekt with some of the guys involved with it.
https://alternativlos.org/36/
https://alternativlos.org/36/
I wanted to post the exact same thing.
One of the most interesting parts of that episode was how the described their relationship with the hardware manufacturers. The deep level of co-working on a very hard task and jointly developing the tech needed was really inspiring.
One of the most interesting parts of that episode was how the described their relationship with the hardware manufacturers. The deep level of co-working on a very hard task and jointly developing the tech needed was really inspiring.
There is another german podcast episode, but in english, covering that topic:
[1] http://omegataupodcast.net/312-the-wendelstein-7-x-fusion-ex...
Also, something older from the middle of 2014, again in german:
[2] https://resonator-podcast.de/2014/res032-der-wendelstein-7-x...
[1] http://omegataupodcast.net/312-the-wendelstein-7-x-fusion-ex...
Also, something older from the middle of 2014, again in german:
[2] https://resonator-podcast.de/2014/res032-der-wendelstein-7-x...
Lots of news about fusion those days. Is it just a coincidence, a fad, the "funding renewal negociations incoming, better get some PR" phase, or can we have the slighest spark of a small quantum of this weird and recently undocumented thing called "hope" ?