Comment by TaylorAlexander | Hacker News Reader

Comment by TaylorAlexander | Hacker News Reader

Parent Full thread

TaylorAlexander

It is still speculative but this week I did a deep dive in to the SPARC fusion reactor from Commonwealth Fusion Systems [1] and it seems extremely promising. They recently proved out their high field magnet design and got 1.8 billion dollars in additional funding, and they claim they can demonstrate net energy gain (a world's first!) by 2025. This would pave their way for a larger production plant design ARC which they currently claim they can have running by early 2030's, though of course these things can be delayed.

But the major points are that the tokamak is a well understood design. Their system uses the same physics basis as ITER, but their more powerful high temperature superconducting magnets mean they can produce a machine with the same Q gain as ITER in 1/10th the total volume. That smaller size means that they can build the machine much faster.

A few youtube videos to absorb all of this:

[2] Is their original pitch lecture from 2016, and makes a compelling argument that tokamak is well understood and simply using new high temperature superconductors allows them to design better magnets to make a small reactor work. It’s the oldest but it’s my favorite talk on the system.

[3] Is from last year so is more up to date. It is a recorded video meeting so it is kind of boring, but the subject matter is very interesting. I actually got this funny feeling like "it is the future and we are discussing fusion power plants over boring zoom meetings". Like it feels more real because it is almost boring. But a good talk.

[4] Is a good 15 minute high level overview for a general audience. If you don't want long lectures, listen to this.

[5] Is a short 4 minute video with highlights from their event last september where they demonstrated a 20 Tesla field in a large bore magnet for the first time ever.

----------------------

[1] https://cfs.energy/technology

[2] https://www.youtube.com/watch?v=KkpqA8yG9T4

[3] https://www.youtube.com/watch?v=uKwWGUT8rCw

[4] https://www.youtube.com/watch?v=-KEwkWjADEA

[5] https://www.youtube.com/watch?v=rAv6p3grFVM

I'm a big fan of fusion research, but I strongly believe that fusion energy can NOT help with the climate crisis-- it's too late for it.

Because IMO ANY path towards commercial fusion (that is not a total moonshot) needs to go through:

1) small demonstrator (JET, Wendelstein 7-x)

2) scaled up research reactor (ITER-like)

3) demo plant capable of electrical energy generation (ITER successor)

4) Mass construction of a somewhat standardized design (kinda like EPR or APR-1400 construction)

IMO, considering somewhat comparable steps in nuclear fission reactor design it is SUPER unrealistic to expect steps 2, 3 and 4 to take less than 10 years each (and even that is EXTREMELY optimistic already).

Throwing more ressources at the problem might help, but not that much; IMO only step 4 is really amenable to parallelization...

You would be right if not for one thing. ITER and its planned successor DEMO take forever to build because they are huge. The whole point of SPARC and its planned successor ARC is that they use new high temperature superconducting magnets with a significantly higher field strength than ITER. This means they can build a machine with the same Q gain as ITER in 1/10th the size. Because of this they can go through steps 2 and 3 much faster. And they have already done step 1 in an existing reactor at MIT. They claim they can do step 2 in 3 or 4 years because the machine is so small.

They recently proved their magnets work in a demonstration last September or thereabouts. They have the strongest large bore magnets ever created, and this directly leads to a much smaller machine. That is why I say they are promising. Really, check out the links!

As optimistic as I am about recent developments, there is no practical path for fusion to be a path forward for replacing our energy consumption. Fusion will likely supplement the technologies that we instill to replace carbon-based energy, but will not itself be part of the replacement of carbon-based energy.

Well it's hard to say right now. It is hard to believe we could have working commercial fusion plants in 2035, but I do think it is possible. If the USA continues to drag our feet on switching to carbon neutral energy, that may unfortunately line up with our timeline. Especially if these fusion power plants are actually simpler to build than fission plants. Unfortunately it is very hard to plan for something that is unproven. But if we spend the next ten years going hard on renewables, we will then have an idea if CFS is going to succeed.

We could build out fission power plants, but if the Vogtle power plant in georgia is an indication of how fission build outs will go, it will be extremely expensive and take a very long time.

  > Especially if these fusion power plants are actually simpler to build than fission plants.

That really is the only hope. Even if the plant itself costs more money to build, the reduced regulation and ability to build almost everything on-site may greatly simplify the process.

I don't believe the hype yet. Fusion numbers do not represent the practical power output [1]. 2025 seems a bit too optimistic.

[1] https://www.youtube.com/watch?v=LJ4W1g-6JiY

Actually she never discusses the SPARC reactor. (I would love to see her discuss it.)

2025 is the date that they will complete their demonstration system and show net energy gain. But it will not be a fully functional power plant. Their slightly larger successor system will do that.

But you’d have to listen to their lectures to really understand. The tokamak is a well understood system. They are just using more powerful magnets than ever before, and they recently proved their magnets work. They published their physics basis research paper in Nature, and these are academics from MIT who have talked very openly about their system design. It doesn’t sound like there are any big unknowns remaining except the engineering design of the system itself which is well underway.

Maybe building this tokamak will be harder for some reason. But it really sounds like they can make it happen.

I'm cautiously optimistic about SPARC, but I don't believe it's going to be more than a footnote in the overall energy storage

At a fundamental level, our lowest cost energy options are really simple - wind blows and makes stuff spin, dig a black rock and throw it into fire, inject gas into fire, let sun shine on a blue brick, chop wood and burn, make water flow and spin stuff

Now compare that to - keep superconducting magnets 250 degrees below ambient temperature, inject high precision beams of deuterium and tritium into a vacuum chamber, capture escaping neutrons in blankets of pure lithium and berylium. Eventually, capture the generated heat in a primary-secondary heat exchange loop of pressurized water, generating steam and spinning a turbine. All of which requires thousands of highly specialized scientists and engineers to design and operate

I doubt it will ever be cost efficient. And we're not even talking about the cost of fuel. Tritium is mind-boggingly expensive - we need deuterium-deuterium fusion, tritium is just an experimental stepping stone

It’s a fair point. But there are industrial processes that need consistent high levels of power. So large cities will probably want fusion for industrial customers, and that power will be made available to others.

Actually if you have fusion you start to think of power hungry things to use it for that were previously infeasible. Carbon capture is one.

For tritium, I’m a little fuzzy on this but I thought maybe the system generates the tritium from duterium inside the reactor? It sounded like it only needs duterium as an input.

If fusion becomes a godsent reality, do you think we will see solar and wind companies fighting tooth and nail against it just like we see oil and coal companies fighting those today?

There will most likely be some sophistry/FUD on the part of lobbyists.

Fusion in the long-term answer will provide benefits far beyond just eliminating power station greenhouse gas emissions, I suspect, but there are many technical and economic hurdles that have to be overcome, any of which could be pointed to when questions about allocation of state funding come up.

To me, it will basically be areas where solar and wind companies are weakest in a local market/jurisdiction where this will take off (as we've seen a lot of the newer solar MW capacity build out not in the US to date).

Also places that want to cater to industrial customers who may need access to continuous high power draw day and night. Like metal refineries etc.

Maybe. But oil companies were huge behemoths while solar I think is more smaller companies. (Though this may change in 15 years.) Plus the value proposition for fusion is so clear that it is going to be hard to argue against. We are going to need to use fusion power to suck CO2 out of the atmosphere.

This would / will still take ages / decades.

It could also be that it makes total sense to use 100% solar/wind + battery in everything not city and use fusion in city/for cities.

I would also argue that the 'no longer needed solar' will then be sold and used everywere else until all poor countries also are using solar energy.

Considering that wind and solar is already cost-competitive with nuclear (fission) power, fusion power might never be viable (opinionated fusion fans often gloss over this possibility).

Sure, there might be potential savings compared to fission because less safety concerns, but the reactor complexity is an order of magnitude higher, and its far from clear to me that fusion power will even be able to compete with current (admittedly often state-subsidized) fission on price, MUCH less with the price of renewables now (or, worse, in 20 years)...

Why having highly centralized power production when you can have it decentralized?

Are people unable to learn from Ukraine? Fission/fusion reactors are just one missile away from nation-wide blackout and panic.

With the new magnet engineering (basically the electromagnets are using some kind of metamaterial for the wiring that's needed to wrap around the toroid that's way better than copper, thus confinement is much easier with the stronger B field in way smaller space [the first research paper that talks about this being theoretically possible was back in 1987, but the engineering capabilities were not there to do such]), you are not limited by volume compared to strength of the B field, so you could have portable decentralized fusion reactors (they talk about this in the video).

He did say there are other practical limits to how small it can be though. You have to stuff other stuff inside the coils and the plasma has to be physically separated from the walls of the chamber by some distance since it is so hot it would melt the walls, and it just gets hard to squeeze everything in there. But I wonder for example if SPARC could be made in to a system that could operate continuously, even though ARC makes more sense from a power generation perspective.

> physically separated from the walls of the chamber by some distance since it is so hot it would melt the walls

This distance could probably be shrunk if some of the gains from matsi for meta-materials translated into fluids engineering, where some kind of "meta" fluid has certain differential properties at a given thermal load (im imagining that the side/region with the thermal load, the particles in the fluid would have some type of glassy dynamics behavior to spontaneously create a radiant barrier back toward the source, would need to talk more with my friend about this since his thesis covered quasi-1d models of glassy dynamics and spontaneous partial arrangements under various conditions on initial perturbation and at steady state).

Reply on news.ycombinator.com