Hacker News Reader

Comment by ChuckMcM | Hacker News Reader

I have some hope for this approach, but as the article points out making reliable magnets is really the key. When a fusion reactor quenches, the plasma will basically eat the reactor through "hole" in the field. I keep hoping an effort to manufacture ReBCo coils directly will be successful, it would both make them less expensive and likely more reliable. However I expect it would require something like a 5-axis 3D printer capability.

Lastly, fusion power is one of the possible 'good' future events (unlike climate change, or nuclear war) that give me hope for the future of the planet.

View on HN

Controlled fusion seems one of the few awesome future things that's unambiguously good.

Particularly if there's some way to capture carbon that's otherwise impractical due to energy requirements.

Fusion power will likely not be significantly cheaper than solar and wind for this purpose. Fusion plants are probably quite expensive and won't have an infinite lifetime. The big benefit is that you don't need batteries, but that doesn't matter if you want to use the power for carbon capture.

Why is is unambiguously good? There's great reason to doubt it could be anywhere close to economically competitive. That makes its value pretty ambiguous, in my view.

I think that an argument can be made that it will always be economically competitive. That reasoning includes the 'fuel cycle' is non-waste producing, the economics of other energy sources continues to rise thus creating a wider window for economic recovery, the liability associated with fusion will always be less than the liability associated with fission, and as the carbon externalizations of fossil fuel are priced into its production it will become in-economic as well.

Well designed fusion power should come in at or below hydro-electric power without the environmental impacts or risks associated with dams.

This is a bad argument. The fuel cycle is a small fraction of the cost of fission power. The main cost is the cost of building the plant. Fusion reactors will be much larger than and much more complex than fission reactors, per unit of thermal power output, so they should be more expensive. Fission reactors today are uncompetitive, even with liability artificially set to a small value, so that argument fails too.

The reason nuclear fission is uncompetitive is its high operating costs. The cost to build the plant is a sunk cost: if the reactor is expected to be reasonably profitable to operate, you can fund building it. If the reactor is not expected to be reasonably profitable to operate... Good luck.

That's what fusion aims to solve. The fuel is plentiful and you can easily buy it; the same can't be said of uranium. It's also much safer to run than fission, and produces vastly less dangerous waste.

Perhaps your argument is that solar and wind are sufficient to power humanity's needs, without fission or fusion. That's debatable. But compared to fission, fusion is theoretically better on some pretty critical metrics — if we knew how to build a fusion reactor, which we don't yet. If you assume solar and wind won't be sufficient, fusion seems worth research.

> The reason nuclear fission is uncompetitive is its high operating costs. The cost to build the plant is a sunk cost

This is nonsense. You can't ignore the cost of building plants, when trying to determine if a technology is competitive. Nuclear plants have to be paid for; they're not given to us free by the Nuclear Fairy. If you include capital and financing costs, you will find they contribute more to the cost of energy from the plants than do operating costs.

Economics of a nuclear reactor https://youtu.be/cbeJIwF1pVY?t=58

Yep (although he's understating the capital cost of nuclear plants.)

> The cost to build the plant is a sunk cost: if the reactor is expected to be reasonably profitable to operate, you can fund building it.

You're missing one important part: the reactor must be profitable enough so that it can recoup the build cost for its lifespan. The lifespan of fusion reactors, even disregarding failure modes, will be severely limited by the high energy neutrons which it generates, and which turn all materials brittle.

And maintenance for the brittle material will mean stopping the reactor, and sending in robots to dismantle and carry away the brittle radioactive walls, and build new walls in place. It may well be about as cheap to scrap it for parts (for those parts that haven't been turned to radioactive Swiss cheese) and build a new one.

The various links to the talks by the head of the ARC/SPARC project in this comment thread talk about the size of the fusion reactor, as does wikipedia:

> ARC is a 270 MWe tokamak reactor with a major radius of 3.3 m, a minor radius of 1.1 m, and an on-axis magnetic field of 9.2 T.[2]

> The design point has a fusion energy gain factor Qp ≈ 13.6

So the reactor is about 25-30 feet in diameter, plus the steam plant / FLibE processing.

Have you seen the size of a nuclear cooling tower?

That's not the reactor itself. I'm talking about the cost of the reactor. The reactor part itself of a fission NPP is quite cheap. Replacing it with a much larger fusion reactor means your containment building (and one will be needed) is larger and more expensive, and you still need the same things (turbines, generators, cooling towers, switchyard) outside the nuclear island.

Next generation fission does not need cooling towers. As soon as you go to a high temperature design, they can run perfectly fine in the desert.

Not familiar with that design. Unless you dump heat, any generator will melt down. How does the 'next generation' dump heat? A cooling tower does it very quickly, and can be scaled arbitrarily large.

Wind and solar have zero fuel cost, but they cost money to build. A fusion plant is always going to be a rather high-tech, expensive thing.

Wind and solar are non-continuous and impossible to use in space, underground, underwater. Fusion potentially could be used everywhere as a relatively high density energy source.

Solar is impossible to use in space? Please explain this claim.

The energy density of fusion is inferior to fission, when you take the mass of the reactor into account (as you must, if you are worrying about the mass of your system.)

Yes, anything beyond Mars is impractical for anything but very small missions.

Is the way from here to Mars not "space"?

What if fusion was combined with renewables? Meaning fusion served similar role as batteries. Storage costs are really high so fusion might be competitive in that space. What I am thinking of is using output from fusion in molten salt storage.

Generators are coupled to this thermal storage. When wind and sun is good, there is a continuous build up of heat stored in molten salt by fusion reactor. When it is low, net heat gets drained from molten salt storage to power steam generators.

At this level of planning, it would likely be easier to just lay the superconducting undersea cables and start building massive solar on opposite sides of the planet.

The reality of energy today is we basically have the technology to cleanly power the human race for the rest of it's existence - politics and tribalism is why we can't.

> What if fusion was combined with renewables? Meaning fusion served similar role as batteries.

Fusion, like fission, is a high fixed cost, low variable cost power source. The cost of energy from such a source escalates rapidly as capacity factor goes down. They are very poorly suited as backup sources for intermittent use. Hydrogen (made by electrolysis from renewables and stored underground, then burned in $400/kW turbine power plants) would be much cheaper as a backup source.

You are right, in short term, fusion has not much value. But in (very) long term, unlike fission, which no matter how incredible it is, is still only 'planetary-level tech', fusion is the first 'stellar-level tech' we've unlocked so far. If we ever plan to go somewhere else than Solar system, fusion is the key.

Beamed power will likely be much more practical for interstellar missions, or even large activities in the outer solar system. In particular, beamed power minimizes entropy generation at the target (and can even actively cool the target), which is vital for high thrust at high Isp.

Molten salt reactors and thorium fission is pretty exciting too.

Making reliable magnets is A key, but it's by no means the only one. Fusion faces many grave obstacles even if the magnets were totally reliable and cost nothing.

What are some of the others ? I’ve heard even if everything worked the generated neutrons would eventually destroy the reactor.

That's one. Another showstopper is that all the energy has to be radiated through the wall of the reactor. By limits on this areal power density and the square cube law, the volumetric power density of fusion reactors will suck. Compare ITER (0.05 MW/m^3) or ARC (0.5 MW/m^3) vs. a commercial PWR fission reactor primary reactor vessel (20 MW/m^3). Stronger magnets don't save a reactor from this.

The large size and complexity of a fusion reactor also means their reliability is a huge problem. There are many parts and joins there, and the machine will be so radioactive hands on access will be impossible. A single leak of coolant into the vacuum chamber renders a fusion reactor inoperable (while a fission reactor can keep operating even with multiple fuel rod leaks.)

I’ve heard the sun has a very low power density it’s just that it’s massive. Maybe it’s just easier to create a star and live in orbit around it

The key point is the Sun is already paid for. :)

Are you certain? ;) Sounds like the plot of a sci-fi in the vein of Hitchhiker's guide where galactic aliens show up with an overdue bill for the sun.

Or find an existing unsettled one and move in!

Tricky bit will be getting to that nearby star...

If only we had an engine with high enough Isp to accelerate to a decent velocity! Maybe something based on nuclear fusion...

That's not that hard actually. We already know how to do fusion, we just have a hard time confining it into a reactor. In space you don't have that problem, bombs work fine.

It's true, but at least partially deceptive because the fusion only happens in the core, which is <2% by volume

The rest is needed to keep the pressure and the magnetic field.

Well .. we do already live in the orbit of a star. (intended joke?)

And we have lots of desert land, that could be covered with solarthermal plants.

https://en.m.wikipedia.org/wiki/Desertec

Also not easy, but sounds more predictable than a possible fusion future.

A good comparison would be, what's the capital cost of a hypothetical fusion plant, compared to equivalent renewables (e.g. battery-backed solar/wind)

https://arxiv.org/abs/1409.3540, Table 11 and Section 6.6.

Yikes. For a 250MWe reactor:

> *Assuming the higher cost estimate for the REBCO tape, the materials costs for ARC total $428M and the total fabricated component cost estimates total $5.56B.

Given it's a novel nuclear project - it's probably safe to double/triple that cost estimate. Still wildly uncompetitive. That's actually really disappointing.

Their cost estimation methodology in that paper seems absurdly simplistic, though.

The fabricated cost estimates were extrapolated from other novel nuclear fusion projects, so doubling it would be double-counting.

> Compare ITER (0.05 MW/m^3) or ARC (0.5 MW/m^3) vs. a commercial PWR fission reactor primary reactor vessel (20 MW/m^3). Stronger magnets don't save a reactor from this.

Stronger magnets absolutely would save the reactor from this, because for a fusion reactor of any given size the power density goes to at least the fourth power of the magnetic field. If we could have ITER, but with 40T in the on-axis toroidal field, it would have a power density of ~150 MW/m^3. This is of course ludicrous because the thermal output of 1.5TW would not be possible to contain, not even considering how the coils are supposed to be kept together. But this clearly show the path that also allows beating the square-cube scaling: You scale magnetic field up, and reactor size down, until you have reasonable power outputs from reasonably priced reactors.

There is potential future in fusion, but it specifically requires better superconducting magnets. ARC at least is research in the right direction of high-field superconductors, even if the REBCO magnets they are using are not quite up to the task of making a reactor that would be economically viable.

> Stronger magnets absolutely would save the reactor from this, because for a fusion reactor of any given size the power density goes to at least the fourth power of the magnetic field.

... right up until you reach the limit on what your first wall can handle. At that point, making the magnetic field stronger gains you little or nothing, because your reactor does not survive operating at the fusion power density it would enable. The claim I am making is that this power/area limit causes DT fusion to be inferior in volumetric power density to fission. The factor of inferiority is roughly (minor radius of fusion reactor)/(radius of fission reactor fuel rod), and the numerator there at least about a meter due to the need to stop neutrons. The factor is independent of magnetic field.

What strong magnetic fields might do is allow you to go to advanced fuels. The hope there is avoiding a thermal power cycle entirely, saving on the non-nuclear side of the power plant. Tokamaks would be hopeless for this, though; their beta is too low.

I'm also thinking of the highly diffusive radioactive tritium gas which is a PITA to contain and work with.

https://en.wikipedia.org/wiki/Tritium

They're hoping to use a lithium "blanket" that would absorb and use the emitted neutrons to breed tritium (that's needed for D+T fusion anyway).

That doesn't save the confinement walls from becoming more and more brittle, it just limits the damage to those areas, and it helps recover some of the tritium it leaks.

It's not that they hope to breed tritium; tritium breeding is absolutely required for DT fusion to make any sense at all. Making the tritium in fission reactors for a 1 GW(e) DT fusion reactor would cost $20B/year.

Reply on news.ycombinator.com