Comment by jmyeet | Hacker News Reader

Comment by jmyeet | Hacker News Reader

Parent Full thread

I fully support more research into commercial fusion power eneration. I really hope it becomes economical. That being said, I'm not convinced this will ever happen for these reasons:

1. Energy loss from neutron loss. So-called "aneutronic fusion" seeks to avoid this problem but requires rare fuels (eg He-3), which kind of defeats the point. Also, it's not neutron free. It's just fewer neutrons;

2. Container destruction from neutrons (ie "neutron embrittlement"); and

3. A superheated plasma (in either a tokamak or stellerator) is fundamentally a turbulent fluid. This is inherently unstable. Any imperfection in the containment could result in significant container damage.

Fusion works for stars because fusion is relatively slow (on a per-unit mass basis, which is why stars can live billions or evne trillions of years) and they have gravity for neutron containment.

More on the difference to tokamaks [1].

[1]: https://www.sciencedirect.com/science/article/pii/S2468080X1...

>This is inherently unstable. Any imperfection in the containment could result in significant container damage.

This is a misunderstanding that conflates multiple concepts: instabilities from toroidal plasma current (unnecessary in stellarators) and edge turbulence (lower confinement time). W7-X has done 30 second shots and are working their way to 30 minute shots. It is their control that has terminated their shots, not the plasma.

You're trying to contain something at ~100 million degrees when the result of that process is destroying the container. It's not an exaggeration to say the engineering challenges are... significant.

Indeed it is and the details matter. That's why I wanted to set the record straight.

The temperature is high but the amount of heat isn't unusual. The atoms are moving fast but there aren't many of them. For a 500MW plant you're containing about the same amount of energy whether it's fusion or coal.

Yep. It's hard. That's why stellarators are so amazing: they can do it.

Exactly!

Update:

> W7-X has done 30 second shots

They're up to 8 minutes now.

https://www.ipp.mpg.de/5322229/01_23?c=5322195

In a D-T reactor you're not losing energy from neutrons, you're using the neutrons to heat a coolant and turn a turbine. (And also to breed more tritium, but that still generates heat.)

He3 is rare but deuterium is decidedly not. Fusion deuterium is about as easy as fusing He3, and the waste product of D-D is half He3, and half tritium which decays to He3 with a 12-year half-life.

Fusion company Helion is building a hybrid D-D/D-He3, and says the combination would release only 5% of its energy as neutron radiation, compared to 80% from D-T. The neutrons would also have lower energy, below the activation energy of common reactor materials.

>gravity for neutron containment.

Is it gravity, or is it gigameters of nuclei in every direction?

you think that high pressure material would just sit there in space, absent gravity?

How much damage can a few grams of gas do to a massive reactor vessel?

Gas? Not much. 14 grams of D-D fusion byproduct plasma? This much damage: https://www.youtube.com/shorts/wqKn_3iJOP4 (I chose 14 grams to make it the same).

While it's ridiculously difficult for fusion to go off like that, even spreading that same energy over 2.488 hours (for a nice 1 GW-thermal reactor) leaves you with the problem that it is in the form of high energy radiation that's more than strong enough to knock atoms out of their latices.

Hm, the fusion plasma would rapidly cool and stop fusing as soon as containment fails. I don't see how a few grams of very hot gas can do more that a little scratch to a reactor wall that is designed to be bombarded by radiation and carry off megawatts of power to a steam turbine.

It's not just gas, it's VERY angry gas.

One part of it is that the plasma sometimes carries a massive current - like, megaamps - that gets dumped into the wall when the plasma contacts it, melting it. This scenario isn't theoretical - it's a big concern when designing and operating tokamaks.

The speed of the particles themselves is also relevant. They're moving rather fast, so they impart a lot of energy when they hit the wall. It doesn't take very many to melt the wall, actually. If some part of the wall sticks out a bit, particles on what are called "banana orbits" will often melt that part off, which isn't good if it's an important part of the wall (e.g. it had been protecting some sensors).

I suspect there are multiple PhDs each focussed entirely on one aspect of the problem of damage from the fusion plasma. The example I gave, of atoms being knocked from their lattices? I only know about that because of meeting someone in an infamous Cambridge geek pub who was working on it.

Reply on news.ycombinator.com