Hacker News Reader

Comment by Animats | Hacker News Reader

Oh, this is just the plan for 2025-203x. I thought maybe they'd actually got it running.

What if I told you that even if everything goes perfectly by 2045, ITER is not a power plant, just a proof of principle? They would have to do further engineering to figure out how to effectively convert the heat generated into electricity.

I'm not sure if they shouldn't rethink the ITER plan now magnets are much stronger and build something like the MIT design instead (https://www.theguardian.com/environment/2018/mar/09/nuclear-...).

There seems to be a bit of a sunk cost fallacy going on at ITER.

ITER is not the route to an economical fusion reactor. The supporters have even stopped promoting it as such. It is now a research reactor. We may find out some novel information on how plasma behaves in a certain regime, but this data is not likely to be useful in building a future reactor. The amount of resources that have been diverted from fundamental fusion research to be spent pouring cement and building magnets is a tragedy. We should have a thousand $3 million dollar or a hundred $30 million dollar experimental reactors, not this monstrosity.

Scale is everything for fusion (plasma). ITER is very important, as it represents the current most likely way. (Hence the name.)

That said, it's big, design by committee, slow, meticulous, etc.

It's not a fail fast market-driven experiment.

Also, most of the money is spent on the fundamentals, planning, developing operational knowhow, basic material science and plasma vessel engineering.

All in all, it could be better, but at least ITER is actually being built. MIT's ARC is still somewhere between "secured funding for a scaled down prototype" and actually will build something. Though it's great news, that they got funding (from Eni, an Italian energy company).

Scale is not everything. There are types of fusion that would not build such huge monstrosities.

We should look at a broader range of ways to do fusion and then finance the ones most likely to work.

As you said, most of the money on ITAR is hardly directly related to fusion. A few companies get lots of money to do research to hit very specific performance that are specific to ITAR. The problem is this does not really build a broad range of companies that work in fusion.

It was simply to early to engaged in such a huge project especially because the technologies are evolving so fast outside of the project and some parts of the design are already not up to date 20 years before it will really be ready.

> Scale is everything for fusion (plasma).

No. Q is everything in fusion. Scale is one way to improve Q in a Tokamak. The problem with improving Q by increasing scale is that it makes the reactors uneconomical. A reactor the size of ITER cannot ever be economically viable, even was capable of producing massive amounts of electricity.

Luckily, scaling up is not the only way to increase Q. The better way is to use more powerful magnetic fields.

Put more eloquently than I can: https://www.youtube.com/watch?v=L0KuAx1COEk&t=43m47s

Wow that's a really great talk, thanks for sharing!

Unfortunately, even with higher magnetic fields, DT fusion is unlikely to ever be practical. That's because the power density of the reactors is inherently limited by wall loading and minimum size considerations from neutron cross sections.

Look at the numbers. The power density of a PWR fission reactor core is 100 MW/m^3. If you consider the volume of the primary reactor vessel instead, it's 20 MW/m^3.

The power density of ITER is 0.05 MW/m^3 (counting gross fusion power). Even if you just include the volume of the plasma, it's 0.6 MW/m^3.

What about higher field concepts like MIT's ARC reactor? If you look at the paper on arxiv where details are given, the power density is 0.5 MW/m^3 -- 40 times worse than a PWR.

The low power density is devastating to the economic case for fusion. Magnetic fusion reactors will be complex, expensive things, with superconducting magnets, complex cooling systems, breeding blankets, heating and control systems. They will be much more expensive per unit mass or volume than a simple PWR reactor vessel. And much of that complex system will need to be periodically replaced due to neutron damage. If the power density is 40x worse they cannot possibly compete.

Yes, (SP)ARC is better than ITER. But ITER is so horribly bad that a reactor can be an order of magnitude better and still not have any real chance of competing.

Well, from a 2016 article on ARC they say cost 4 to 5 billion USD, output to the grid 200MW, which is not competitive with existing reactors but that would be the first one producing power built, the designs may improve. (https://www.computerworld.com/article/3028113/sustainable-it...)

I don't see why the number of cubic meters matters?

If you give me a machine that outputs 50kilowatts, I'll happily give up a cubic meter of my house for it...

Because size (and complexity) directly correlate with cost. So unless fusion can enable cost reduction elsewhere, if the nuclear island is inherently larger, more complex, and hence more expensive than in a fission plant, power from it will be more expensive than from a fission plant. And in that case, why would any utility want one? New and risky (in the sense of having a significant chance of not working as well as hoped) technologies like fusion will be adopted only if they are significantly less expensive than more proven alternatives.

The size and complexity also directly affect reliability. There is more to go wrong in a fusion reactor than in a fission reactor, and repairing anything there will be difficult because hands-on work will be impossible.

Agreed. Aneutronic fusion is the end goal and the real prize in the fusion game, and we're very very far from that. Yet with just waiting for it, we'll never get closer.

And of course eventually clean (as in proper reprocessing/recycling) fission energy would be great too, but that again is stalled, largely due to nuclear armament proliferation concerns.

The research I've seen planned isn't about plasma physics - it's about material properties, lithium blanket designs, and lots of other finicky engineering details that will be necessary for actual practical reactors of different designs.

While those are all important, I thought that the main research goal of Iter was to understand the stability of the plasma for sustained net-energy operation. Aren't there still some remaining open questions about operating a D-T tokomak reactor to produce energy, like, does the plasma begin to rip itself apart after 1ms, okay, how about 10 seconds, how about 500 seconds, and so-on.

I accept that I'm talking based on extremely vaguely remembered statements from many years ago...

After all, if the goal was really these ancillary things you mention, many of them could be studied in other environments.

There’s nothing wrong with that. That’s how hard problems get solved - you tinker with them until a new understanding and clever ideas emerge. Then you build new and better things based on those ideas, and continue tinkering with them. Etc etc.

Some things are so complex, like nuclear fusion and the Standard Model of physics, you have to build big complex things like tokamaks and particle colliders to be able to tinker with them. No one should be surprised or put out by that, it’s part of the process.

The problem with this argument is that in engineering, the goal is not just to get something that works, it's to get something that works better than the alternatives.

That means it's inescapable that there will be losers, regardless of how much is invested. There's no guarantee that fusion will win this competition, and good reasons to think it won't. Fusion may well go into the bin (with lighter than air passenger aircraft, superconducting digital computers, and MHD coal powerplants) of technologies that just couldn't cut it.

Possibly, but it’s way too early to give up on fusion now, especially considering its benefits - clean fuel, clean byproducts, less risk of meltdown and similar disaster vs fission.

Actually, it's long past time to give up on fusion, at least DT fusion. It's been known since at least the 1980s it would have lousy power density. DT fusion's putative advantages don't make up for that, nor for its grave operating issues (frequent replacement of major components and serious reliability/repairability problems.)

I would say that ITER has no potential to really point the way for fusion energy, it’s an international boondoggle born of Cold War/Post Cold War nuclear politics. As a vehicle for real science and engineering it’s bollocks. The real, slow, incremental research is being done elsewhere. We’re so far from a viable fusion plant rather than a test reactor that it isn’t even funny. So many basic problems (such as what you’ve outlined) remain unsolved even in a laboratory context, never mind the context of a large plant that would need to operate more or less continuously.

Reply on news.ycombinator.com