Comment by mrterry | Hacker News Reader

Comment by mrterry | Hacker News Reader

Parent Full thread

I've worked in ICF for a nearly a decade. I think NIF research is a necessary step to a ICF based power source. If NIF demonstrates ignition, it can be used to validate our physics models of ICF (turbulence, fusion product transport, thermal conduction, etc). These physics questions are shared by all ICF approaches. However, a fusion power plant will use entirely different laser technology [1], and most likely a different target design [2]. After showing that the physics works, there is still a lot of engineering work needed before building a reactor.

[1] An ICF power plant will be pulsed at 1-10 Hz. NIF is a flash-lamp pumped glass laser, which takes ~12 hours to cool between shots. A power plant would likely be diode-pumped sold state laser since these can meet the required repitition-rate.

[2] The indirect drive target that LLNL is pursuing on NIF is not very efficient. You spend a lot of energy heating the hohlraum. Directly driven targets (blast the capsule directly rather than heating a gold can to make x-rays) should be much more efficient. There are also several ideas for ways to ignite a target more efficiently (shock ignition, fast ignition), but these need additional laser hardware.

@sam Though I work in ICF, I was sad to see the innovative confinement concepts (magnetic confinement) cut a couple years ago. I think it is short sighted. Fusion need to work and get smaller and we should keep our options open. Hope you managed to get a thesis out before the walls fell.

Good to see a commenter with an ICF background (most of what I know of ICF was only from water cooler chit-chat).

I didn't manage to get a thesis out. I left the PhD program before the funding was cut, but the writing was on the wall.

Yes, it is short sighted to cut funding of small scale plasma confinement concepts. They are high risk / high reward projects which are not expensive ~$1M. And even if they don't pan out as viable confinement schemes, they make for great training platforms for graduate students.

I thought the goal was to have a self sustaining reaction that would only require laser pulses to hold it in place. Is a continuously pulsed laser necessary to provide thermal energy to drive the reaction?

Also I've read (sorry no citation) that the indirect drive was necessary to ensure even dispersal of the heat generated by the initiating laser blast.

Any clarification would be much appreciated, this isn't my field.

ICF is intrinsically pulsed (magnetic confinement like ITER is completely different). You use a laser to spherically compress a 1 mm diameter spherical capsule. The capsule implodes, stagnates, and blows (releasing energy in the explosion). Then you do do it again 0.1 seconds later.

NIF uses an "indirect drive" design. Instead of directly illuminating the spherical target with a bunch of lasers, you blast the inner surface of a gold cylinder with the shell at the center of the cylinder. The cylinder gets hot, emits x-rays, which are absorbed by the capsule. The x-ray drive tends to be "smoother" than direct illumination.

The big problem with ICF is hydrodynamic stability. It is like trying to squeeze a water balloon with your fingers. If you don't squeeze it perfectly symmetrically, it will squirt through your fingers and pop rather than getting compressed by a factor of 20.

>The big problem with ICF is hydrodynamic stability. It is like trying to squeeze a water balloon with your fingers. If you don't squeeze it perfectly symmetrically, it will squirt through your fingers and pop rather than getting compressed by a factor of 20.

Well said.

I agree with your general philosophy towards ICF.

You get the same problem with laser ignition, funnily enough.

My layman's understanding of the subject is that you want (and get) anything but a self-sustaining reaction with fusion, and that's actually one of its safety features compared to fission (where a chain reaction is allowed to happen in a controlled way -- if something goes wrong the reaction keeps going, out of control).

You do want to use the energy from one reaction to power the next, but not directly. No one's actually trying to create a star on the surface of the earth.

As with you, though, not my field so someone who knows better can feel free to correct me.

"one of its safety features compared to fission"

Not exactly. Off the top of my head I can't think of a reactor thats unstable in operation. Hands off (like that TV series life after humans where the people all disappear) almost all stabilize and eventually shutdown on their own. Maybe those crazy graphite reactors are not inherently stable.

There's two linked issues.

Fission reactors get about 10% of their heat from decay products (the "waste") decaying away. That means there is no instant off switch. The "off" position is still 10% out for hours/days/weeks (well it decays away eventually...). As the Japanese found out the hard way, 10% of a huge amount of power is enough residual heat to cause an awful disaster.

The other is surface area / volume ratio. Couple gigawatts thermal like the Japanese and there's not enough surface area to cool it without giant pumps when shutdown. Couple MW like a nuclear sub and the surface area ratio is better, theoretically you could probably walk away from one without anything awful happening. The people who know aren't talking. Fusion you're talking about something a millimeter across not multiple meters. Walk away or have a computer crash or whatever and it seems impossible for something that small to cause much damage.

There are some practical physics reasons why making a fusion reactor the size of the sun is really easy and the smaller you go the harder it gets, but there are practical engineering reasons why making one bigger than a millimeter would be a huge PITA. On the other hand if you could make a fission reactor the size of a millimeter that would certainly solve a lot of painful thermal engineering problems, but the physics just doesn't work (long story...)

There's also the small matter of the hohlraum costing $100,000 a shot.

Reply on news.ycombinator.com