Comment by YZF | Hacker News Reader

Comment by YZF | Hacker News Reader

I have some friends at General Fusion and I've been following them relatively closely from the beginning. I had an opportunity to invest some in the beginning and decided not to - mostly because the problems they're trying to solve seemed so hard.

They've pivoted a little bit on their approach since then. In the beginning it was all about focussed pressure waves. Now they're starting with plasma. Fusion is really hard (well, at least for power generation, perhaps the "level a city" kind is easier). I think it's harder than self driving cars (for example) or landing rockets on barges (as another example). There are many fundamental problems in different areas that have to be solved and not all these areas are well understood.

I'm pleasantly surprised they are still going strong and that they were able to raise money to continue their progress. Obviously if they are successful the business opportunity of all power generation in the world is a pretty big one...

Fusion is easy. Lots of fusion (but not quite city leveling kind) is hard.

As far as I can tell most of the difficulty arises in the plasma physics part instead of the actual fusion bit (the latter being fairly well understood now for quite a long time now). That's where 90% of the work has to go get fusion power and also the reason the early estimates for when fusion power would be available were so optimistic. It wasn't clear back then how difficult and unforgiving plasmas can be (and making the plasma extremely hot, applying strong magnetic fields and running high currents through it doesn't make it easier to work with - quite the opposite).

Sure. I believe there's some fusion happening around us all the time. It's a statistical event. But to get a useful amount of fusion is different.

Yes, plasma is very difficult. You need to get it hot enough and in the right shape and in the right position at the right time. But even when you do that there are still many hard problems to solve. One of them is irradiation embrittlement. Capturing all the energy efficiently and converting it efficiently into power is also difficult (steam turbines I guess). The efficiencies of the input side are also a problem (the energy intake of creating the plasma, managing it, "fusing" it). All these inefficiencies eat into your energy budget making it harder to get significantly more energy out than you put in... Anyways, I'm not a physicist, but every little problem I hear about related to this process seems to be of monumental proportions and there are many of those...

Evan in thermonuclear bombs, most of the benefit in fusion does not from the energy released by the fusion itself, but rather from the neutrons released that trigger fission surrounding U-238 (the non-bomb isotope of uranium).

How does the design process for these reactors work? How much design is done in simulation vs experimentally and how high-fidelity are the simulations?

From what I know a lot is experimental.

This is not my field (just an enthusiast) but consider that a typical fusion plasma has a density of 10^20 particles per cubic meter. Even if your plasma is 1 cubic centimeter, that's still 10^14 particles! Thus a full-on EM simulation, where every charged particle interacts with every charged particle, is not practical. A full-scale simulation is still unrealistic even if you typically cheat by breaking the problem up into cells and ignoring the Coulomb interaction beyond cell boundaries.

Fusion machine design is a combination of hard-won experimental knowledge and theoretical guesswork. I'm skeptical of the current wave of fusion startups simply because they don't have the weight of the decades of experimental work that have gone into tokamak research. But hey, they might make a breakthough, who knows.

Reply on news.ycombinator.com