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.
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).
>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.
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.
* 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.
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.
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.
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.