Lockheed's Skunk Works Building Bigger Fusion Reactor | Hacker News Reader

Lockheed's Skunk Works Building Bigger Fusion Reactor

thedrive.com

175 pointsSEJeff122 comments

> Containing the reaction, the same one that occurs in our sun and other stars...

Not really. Our sun is fueled mainly by the proton-proton chain reaction, which produces no neutrons. This reaction is very very slow and happens at comparatively low temperatures. The large power output of the sun is because the sun is huge.

Useful fusion reactors need to use much faster reactions and require temperatures several times higher than the sun’s core.

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

To add to that, the sun's average power density barely rivals a compost heap.

That’s awfully simplistic. Having spent most of today day wrestling with permutations of the various radiometric flux and power formulas. It’s pretty clear that depending on what you count as the “edge” you can get some very disingenuous results when it comes to anything related to the sun’s density. What’s the math behind the compost heap estimate there?

(Genuinely curious since I’ve spent the day doing math adjacent to this one)

Physics exchange has a calculation. Bunch of articles about it accessible from google.

That was my reaction too. Indeed, among the many possible ways to think about it, the sun can warm the compost heap, but not vice versa.

On the other hand, the sun is a cosmic compost heap. ;-)

Remember the square-cube law; for two things of different sizes, the difference in surface area is proportional to the square and the difference in volume is proportional to the cube of the difference in sizes.

So, compared to any Earth-bound compost heap, each bit of the sun's surface (wherever you draw the boundary) has more heat-generating interior backing it up.

According to the inarguable science of taking the first google result for my questions…

The sun releases 3.846×10²⁶ watts. The sun has a volume of 1.4×10²⁷ cubic meters.

Therefore the sun releases less than one watt per cubic meter.

Scanning a 2010 article: Energy from Waste: Reuse of Compost Heat as a Source of Renewable Energy it looks like composting facilities make about 1 to 10 megajoules/kilogram over a fortnight. At 10⁶ seconds per fortnight gets us 1 to 10 watts per kilogram of compost. At a minimum of 100kg/heap (anything less is a pile, a mess, or a spill) it is a trivial exercise to show that compost is orders of magnitude more energetic than the sun.

Now if Hacker News would just allow stick figure cartoons maybe we could lure Randall Munroe into answering these questions for us.

Using average power is kind of misleading:

The core inside 0.20 of the solar radius contains 34% of the Sun's mass, but only 3.4% of the Sun's volume. Inside the 0.24 solar radius is the core which generates 99% of the fusion power of the Sun.

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

Lockheed enlarged their original concept, reducing the power density (counting the volume of the reactor, not just the plasma) by a factor of 100. It's now about 0.5 MW/m^3, about the same as MIT's ARC concept.

In contrast, the power density of a PWR primary reactor vessel is 20 MW/m^3 (and of the core alone, 100 MW/m^3). How exactly is fusion supposed to be cheaper than this, when the reactors are both so much larger and so much more complex? Saving on fuel is irrelevant, as fuel is only a small part of the cost of running a fission reactor.

What’s meaningful about power density as a metric for reactors?

Power density determines size (for a given power output), which is strongly correlated with cost. Simply put, larger machines are more expensive than smaller ones. And cost is the showstopper for fusion, not the comparatively shallower issues of plasma physics.

Cost also rapidly translates iteration time too. The bigger you get, the more expensive it is and the more you need larger groups or collaborations, which means more politics. If you can shrink the reactor enough you can go from requiring governments to work together to just needing a single company.

I'd assume it's a decent proxy for cost. It would also be an important metric for mobile reactor usage

You're not just saving on fuel, you're saving on waste

And safety. Fusion has the enormous advantage that you can just switch it off, without the core continuing to generate heat and causing a meltdown if the cooling system fails.

0.5 MW capacity costs around $1M with solar, so if 0.5 MW of fusion power takes up a cubic meter that doesn't seem like a deal breaker.

The argument involves the comparison with fission. The low power density means it's going to be more expensive than fission, and fission already fails economically.

The difference between fission and solar is that solar lacks many systems (like, turbines and generators) that fission and fusion would need.

More fundamentally, solar is distributed (even on a large solar farm), so there's a lack of coupling between elements that makes everything easy to assemble and very fault tolerant. In fission and fusion plants, the intricate dependencies mean everything has to be highly reliable. Safety (for fission) and the extreme difficulty of repair of activated structures (for fusion) also drive this very expensive reliability.

The argument involves the comparison with fission.

With the current LWR/HWR fission reactors, where core is solid.

> switch it off, without the core continuing to generate heat and causing a meltdown

A number of fission designs have this property, too. Most operational fission reactors are from 1960s to 1980s, and indeed are capable of a meltdown.

And if one were allowed to design a fission core with a power density as low as a fusion core, it could also be made highly resistant to meltdown, because it would have so much more thermal inertia.

One might ask why fission reactors are not designed that way.

Technically true, but not in the way a casual reader may think. The storage and ultimate disposal of spent nuclear fuel (i.e. "nuclear waste") is a small fraction of the overall cost of the nuclear power enterprise. The Nuclear Waste Fund has been accumulating capital from rate-payers for years and has about $50B in it in the USA, from ~100 operating plants that have sold ~$1T of electricity in their lifetimes.

It's the short-term handling of fission products (that eventually cool off and become nuclear waste) that makes fission expensive. Since fusion products are much less radioactive than fission products, fusion has a major advantage over fission in terms of worrying about containing radioactive material. To contain it in fission reactors, we have lots of redundant cooling systems, control systems, instruments. We have a powerful regulator checking everything, and requesting major changes. We have teams of people planning maintenance work to minimize dose to the workers. We have big 50km radius emergency planning zones with sirens and drills and all that.

Nuclear fusion is the dream energy source because it has all the positive elements of fission (24/7, zero-carbon, can build anywhere, tiny fuel requirements b/c nuclear) without the major downside of radioactive fission products and transuranics.

I got into the nuclear industry specifically to help solve climate change with fusion, but I got distracted by advanced Gen-IV fission reactors for now, assuming they'd be shorter-term developments. Fusion is still the dream though, for sure.

The major downside of fission isn't fission products or transuranics. Would that those were! The major downside of fission is that it's too expensive. Fusion would take this, THE showstopper for fission, and make it worse.

It's an example of not focusing your problem solving on the actual problem.

Cost is the proximal issue, but why? My claim is that if the fission products and TRU were not radioactive it would not be expensive. This point is pretty easy to argue, as most cost ratcheting has been related to protecting against radiation hazards.

Fusion products are far less radioactive but the engineering to get to the proper physical conditions is very hard and expensive. Fission is the opposite. Trivial to get fission chain reactions given the material (done in 1942 by Enrico Fermi with natural uranium), expensive to deal with radiation.

Fresh fission products will always be radioactive, that problem isn't going away easily. Engineering challenges of fusion could possibly be solved by good engineering and fancy materials. That's a lot easier to postulate.

Thus, fusion remains a potentially-attainable dream.

Your claim is that because fusion doesn't have fission products and transuranics, it will be cheaper?

This doesn't make any sense to me.

For one thing, those isotopes are packaged up nicely in fuel elements. They don't get spread around the reactor. And disposal of spent fuel is not what makes fission expensive or uncompetitive.

Perhaps you are talking about decay heat from the isotopes leading to safety engineering costs. This has nothing to do with transuranics. Cooling does demand high reliability, which adds to expense, but it's not the only thing that leads to a need for high reliability. Another problem is that anything in the reactor proper is very difficult to repair if it breaks. This will push designers to try to make the reactor as reliable as possible, and designing complex systems that don't experience breakdowns is really expensive. This problem will be worse for fusion, not better, since the fusion reactor itself (which will be far too radioactive for hands on maintenance, even after prolonged shutdown) will be much more complex than a fission reactor, and under higher stress.

Oops, I clarified my point above while you were responding.

See last two paragraphs above. Fission is expensive because of radiation hazards. Fusion is expensive because of engineering/physics. The latter is easier to postulate solving.

Now you are claiming that not only fission products make fission expensive, they do so in a way that is immune to engineering solution (unlike, you are saying, the problems of fusion.) I cannot understand how this makes any sense.

Fusion reactors will be regulated by the same regulators as fission reactors. The NRC has already said that fusion reactors fall under their purview. Tritium release alone would be sufficient to ensure they will be regulated -- the tritium burned by a 1GW(e) fusion reactor in a year would be enough (if all released) to raise 2 months of the flow of the entire Mississippi river above the legal limits for tritium contamination in drinking water. Fusion reactors will be regulated to contain even small amounts of tritium, and this will not be cheap.

I explained already why the redundancy and expensive reliability engineering of fission plants will also be necessary for fusion. It's not for the reason of safety, but because any breakdown in the hands-off part of a fusion plant will be catastrophic for the economics of the plant.

Fusion is starting way behind fission in the struggle to be sufficiently reliable, because fusion reactors are so much larger, under much higher stress, and are much more complex.

MSR/LFTR are fission, produce a fraction of the waste, and would require a fraction of the investment to make them market ready.

Don't get me wrong, happy to see fusion advancing. But we need lots of power, today, market-ready, outside the lab. No, not the US or Europe. China, India, Africa. They want to grow, now, and we don't want them to use coal/oil, nor outdated PWR/LWR/HWR.

If fusion is 20y and $1T away, and we can have hundreds of MSR/LFTR 1GWe plants by investing $1B over 10y years, desalinating water, making aviation fuel out of sea water, and recycling fission waste from older reactors, I'd stick with fission for now.

The biggest hurdle for fusion power is, ultimately, how to deal with the neutrons. These particles are electrically neutral and their trajectories are unaffected by electromagnetic fields. The approach so far is to put a «blanket» around the fusion core to slow them down, but this fails to mention that the blanket will be slowly (or not so slowly) damaged by neutron induced transmutation. Since a lot of the energy released in fusion power comes in the form of kinetic energy of neutrons, it’s also essential to convert that energy into eg steam, for power generation.

Is putting a big pile of concrete around the reactor not a suitable solution? Concrete is often used as neutron radiation shielding because it's full of hydrogen (in the water chemically bound up in it) and because of its desirable structural properties.

They already do this with ITER[1] for the "bioshield" to protect humans from radiation, the facility is semi underground as well.

However by the sounds of it, a reasonable fraction of the output power is in the form of neutrons. If you had no lithium blanket and just concrete, the energy of the neutrons would go into heating the concrete, where it would either be wasted, or require the concrete to be actively cooled which would be difficult as concrete is not a great conductor.

[1] https://f4e.europa.eu/mediacorner/newsview.aspx?content=941

The main concern with the neutrons is damage to the reactor itself — you don’t want to have to replace the main chamber walls with any frequency. Fusion reactors generally do not release enough neutrons to hurt people. So shielding on the outside doesn’t do much except provide additional fail safes.

The neutron flux inside a fusion reactor is many orders of magnitude higher than the maximum allowable for human exposure. The flux is low on the outside precisely because of the many absorption lengths of shielding between the reactor and people.

The blanket also has to have a high cross section for elastic collision with neutrons to efficiently capture their energy as heat. ITER is planned to have liquid lithium as this blanket because it has a high cross section and the odd neutron capture will result in helium and much desired tritium that can be extracted for use as the main fuel in the reactor. Unfortunatey, elementary lithium is a very temperamental material.

I heard a story a while ago about a facility working with a lot of molten sodium. The whole building had no running water at all, no bathrooms even, because water and sodium mixing is a problem.

Lithium doesn't seem a whole lot safer.. but I don't think a fusion power plant could be water free. As far as I'm aware there is no more efficient method of using heat to spin a generator than a steam turbine. A lithium/water heat exchanger seems pretty frightening. Do they have a better plan for that?

From somewhat outdated memory: lithium is thermodynamically less stable in the presence of water, but the reaction is slower than it is for sodium. (And so on: potassium and cesium react even more violently with water.)

But this is largely moot, as lithium in salt form (Li+) does not react with water. I would be more concerned about the beryllium salt dissolving in water and escaping. Beryllium is nasty.

If memory serves, lithium is more reactive than sodium. ITER will have some kind of heat exchanging system to cool the lithium. But I don't know what that secondary system uses.

Haven't looked at the proposal, and not commenting either way on the risk of using water in this context, but the water/reactive-metal issues have been widely investigated in other liquid metal reactor designs and generally intermediate cooling loops with some benign medium are used.

Edit to expand; having now looked it seems some of the intermediate loop designs have the SAME material (e.g. sodium). This at least means the loop at risk of mixing with water is isolated from the core (where an explosion is difficult to manage). I thought I'd seen other designs, just can't find them now.

MIT's design uses molten FLiBe salt as the blanket. It's chemically stable, and the beryllium multiplies the neutrons for breeding with the lithium.

The inner wall of the reactor is 3-D printed, and the reactor uses jointed superconducting tape that allows it to be opened up annually to replace the inner wall.

MIT's design also uses 40% of the world's annual mine production of beryllium, for a single reactor. Using the entire estimated Be resource in these reactors would supply maybe 1% of the world's primary energy demand.

That was once said about lithium and batteries. Typically once the market demand is there we suddenly find a bunch of viable reserves to tap. This is tragically happening with petroleum where we're tapping incredibly difficult reserves via deep sea drilling and fracking.

Asteroids!

I don't think asteroids would actually be a good source of Beryllium. IIRC they have approximately the same overall elemental distribution as the earth does, just more evenly distributed - which is great for finding dense materials like gold or rare-earth metals that have sunk towards the earth's core, but not so great for one of the lightest metals in existence.

Gold and the like have sunk into the core not because they are heavy, but because they chemically partitioned into a phase (liquid iron/nickel) that was heavy.

Uranium, which is also heavy, partitions into light silicate minerals, and is enriched in the Earth's continental crust by 3 orders of magnitude above the average of the planet. Without this concentration fission energy would likely never have been considered as an energy source.

It is my understanding that beryllium is concentrated in the Earth's crust by about a factor of 100 over carbonaceous chondrites. Beryllium is actually very rare on a cosmic scale, since it is one of the "X-process" elements made by cosmic ray spallation, not in stars.

> The biggest hurdle for fusion power is, ultimately, how to deal with the neutrons. These particles are electrically neutral and their trajectories are unaffected by electromagnetic fields.

Layman here, but surely they must interact or bond with something? Perhaps to produce something else that could be used.

Well, yes, they interact with most things, by colliding with their atomic nuclei, thereby turning their atoms into radioactive isotopes, and generally destroying any object or living matter in their path.

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

I'm a total ignoramus when it comes to particle physics, but what you just described sounds like a particle-based process. Is there an equivalent description if you think of neutrons as waves?

Not really. At these ranges and energies a particle is the best description.

Can we use neutron absorbers? Here's what wikipedia [1] says about them:

"The most important neutron absorber is B [boron] as B4C [boron carbide] in control rods, or boric acid as a coolant water additive in PWRs. Other important neutron absorbers that are used in nuclear reactors are xenon, cadmium, hafnium, gadolinium, cobalt, samarium, titanium, dysprosium, erbium, europium, molybdenum and ytterbium; all of which usually consist of mixtures of various isotopes—some of which are excellent neutron-absorbers."

By the way, boron carbide [2] has an extraordinarily high melting point (2763 C), maybe this could be handy if one wants to convert the thermal energy of neutrons into electricity.

[1] https://en.wikipedia.org/wiki/Neutron_capture#Neutron_absorb...

[2] https://en.wikipedia.org/wiki/Boron_carbide

The neutrons in a DT reactor have to be used to breed more tritium, by absorption in 6Li. One cannot afford to lose many; indeed, some neutron multiplication by (n,2n) reactions on various structural materials would be needed to make the cycle close.

I see. So you are saying it's better to use a neutron reflector [1] rather than a neutron absorbant. Either way, it doesn't appear to be an insurmountable problem.

[1] https://en.wikipedia.org/wiki/Neutron_reflector

Fusion is hard. Neutrons are a concern, yes, but none of this matters without ignition, without break-even (choose your definition). Until break-even is achieved, it’s all premature optimization to me.

Compact or not, clever neutron handling or not, integrated tritium-breeding or not, break-even is the minimum bar to cross, and no one has crossed it outside of fission-triggered bombs. That must be the focus. Once that is achieved, there will be plenty of resources to solve the myriad of other problems, but until break-even is achieved (by ANY means other than fission trigger), solving those other problems just doesn’t matter.

The focus on ignition is strange. If the engineering and economics look hopeless, then why focus on ignition? The only reason I can think of is that the people working on it want to keep their jobs and defer the likely practical showstoppers until after the ultimately irrelevant plasma physics problems are solved.

Well it's a step in the direction. If you can get that to work then maybe you can then figure out the other stuff afterwards.

In the same sense that climbing a tree is a step in the direction of getting to the moon.

In the same sense that flying to orbit is a step to getting to the Moon. We did Mercury before Apollo.

Climbing a tree is useful for going to the moon. Having access to heights from which to drop things is a prerequisite step for the first experiments that lead to an understanding of gravity and acceleration ;).

Global level strategy suspicion, guess: The US has to try because China, Russia, even North Korea, are trying and it just MIGHT work, or something from it might work and be the basis of something serious eventually.

Broad observation: Physical science research was a sleepy area until The Bomb. Then supposedly Ike with J. Conant, et al., said, IRCC, "Never again will US academics operate independent of the US military." Then the US Congress via the NSF and DoD went to the leading US research universities and "made them an offer they couldn't refuse" -- accept the grant money or cease to be a leading research university. And, oh, BTW, you can take 60% or so as overhead for the English department, the string quartet concerts, the little theater production, the glossy alumni magazine, the new front gate to the campus with the big, round fountain just inside, the new art gallery, a new, on campus, 6000 square foot Georgian house for the president with, of course, a black limo. Or some such!!!!

Then the bio-medical sciences mentioned to Congress that a big fraction of Members of Congress are old and need medical care for heart disease, cancer, other diseases of aging, etc. so should have, say, the NIH to do for bio-medical what the NSF does for the STEM fields. Congress went along.

If ignition hasn’t been achieved, why bother with clever engineering or economics? Without ignition/breakeven, the best engineering and economics in the world are irrelevant.

Ultimately, physics showstoppers matter more than any others.

If you ignore all other parameters and achieve that one, you run the risk of walking down an engineering path that cannot deal with those downsides. Seems 30 years out still...

I just find that these other considerations are either an excuse or a distraction from actually meeting this minimum objective goal. Breakeven or get out.

Every journey starts with a first step, but so does every dead end.

If the engineering and economic issues are showstoppers, then focusing on breakeven is simply wasting money. It's almost a canonical example of shortsightedness.

It's actually much more true the other way around: if physics is somehow fundamentally preventing you from fusion ignition, then focusing on engineering or economic issues is a complete waste in money and a dead end. Because while economics and engineering change over time, physics[1] does not. It's like spending money trying to solve the economics challenges of a perpetual motion machine[2].

[1](Although our understanding of physics does.)

[2](...which fusion is not)

A showstopper is a showstopper. Yes, physics can rule things out. But so can economics. There is no reason to give the first priority over the second, if the second is sufficiently bad. And I will claim those practical considerations render fusion very dubious. And yes, they change over time: they have gotten WORSE for fusion. It used to be that fusion was competing with fission to be that post-fossil energy source. That was, perhaps, not too unreasonable, especially if the fission had to be with breeding. But now, it competes with dramatically cheaper renewables.

Yes there is because again, physics does not change. Physics showstopper rules out all others.

To the end user, it doesn't matter if physics or economics caused a technology to be useless. It's still useless. And it doesn't matter to the financier either. In either case, any investment he made is lost.

Your attitude is a reflection of a culture that considers physics to somehow be a deeper subject than engineering, facing more fundamental and important problems. The details are "just engineering", somehow not important or relevant.

The problem is that we need it to just barely break even and then a little bit more. Having a tremendously energetic fusion device is both useless and a solved problem: we already know how to light one off, but it only lasts a millisecond and it destroys everything for miles around.

For what it is worth, fission also suffers this problem. A promptly critical fission reaction is possible, indeed quite easy to achieve, and very undesirable. All fission technology is dedicated to keeping the reaction just barely critical.

http://julien.hillairet.free.fr/wiki/doku.php?id=list_of_fus... is the most complete list I've seen, of private fusion efforts

Maybe also interesting in this context:

https://en.wikipedia.org/wiki/Wendelstein_7-X

If you understand German I can only recommend you give this interview a listen: https://alternativlos.org/36/

Edit: I should mention that it is about Wendelstein 7-X.

That is very informative, thanks!

For those who don't understand German, this is an interview in English, also from someone at Max Planck about Wendelstein 7-X:

https://omegataupodcast.net/312-the-wendelstein-7-x-fusion-e...

Thanks, knew this one already :)

And MIT's SPARC reactor plans:

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

https://www.psfc.mit.edu/sparc

https://www.youtube.com/watch?v=KkpqA8yG9T4

And General Fusion still seem to be going: https://en.wikipedia.org/wiki/General_Fusion and claiming 2023 as a date for a working demo: https://www.nextbigfuture.com/2018/05/general-fusion-will-co...

GF recently redid their concept, because the previous one wouldn't work (magnetic fields were far too high, which would cause liquid metal to vaporize off the plasma facing surfaces, hopelessly contaminating the plasma). The new concept uses slower compression, and has a solid metal conductor down the middle of the chamber. This conductor would experience enormous neutron loading (it cannot be shielded from that by a thick layer of liquid metal) as well as enormous forces from magnetic fields up to 100T.

I give this scheme very little chance of being workable.

Can you share a link to the design changes mentioned?

https://generalfusion.com/wp-content/uploads/2018/07/ICPP_20...

https://arpa-e.energy.gov/sites/default/files/11_LABERGE.pdf

Thanks!

and those :

https://lppfusion.com/

who make lots of reports and thus, are interesting to follow. However, I've no idea how serious they are...

They've gotten papers published in serious journals, and present at fusion conferences. That doesn't mean they're not a long shot but whether or not it works out, they're doing real science.

"Despite slower than expected progress"

I mean on the one hand this is the story of fusion. On the other they should have preempted that by being extra pessimistic, so was extra pessimism built in? Did they guess they were going to be perennially 5 years away when in fact they're 10 years away. Or did they think the project would be done in 18 months, and is still 10 years away? Any bets on what the predicted timeframe will be in 10 years time?

Five years:

https://news.ycombinator.com/item?id=5323504

The original article is gone, but the click-bait title of that HN post was "Lockheed's Skunk Works: fusion power in four years?"

It did end in a question mark, I guess.

>"How do you scale it up to generate power for a city or an entire town? That’s all ahead of us,"

If they think they can get a working fusion reactor capable of powering 80,000 homes into a shipping container, I'd suggest not building them any bigger. Being able to scale in container sized units over time via the existing transport links sounds just peachy. Is quite a big 'if' there.

Why was this downvoted? I had the same idea. If you can power 80000 homes with a container sized reactor, why not start deployment? Easy to transport, you could power small neighbourhoods. No more wide spread black outs.

Not sure how much maintenance will factor in, but a reactor that size sounds already practical.

It's still a nuclear reactor... You don't just put them into containers and ship them around, or turn them on or off by pressing a switch. They require specialized staff to operate, maintenance, handling of materials, safety and protection precautions and so on. Either you built them big or they are not worth it to operate.

OK but why can't we have just one of them running somewhere as a demonstration?

Kind of like computers back in the day?

Exactly. The reasons we have giant centralized power production is that it's cheaper to fuel and maintain one big electrical power station than it is to do so with many small ones. If a new technology doesn't need frequent refueling or maintenance and doesn't become dramatically less efficient when scaled down, it's probably better used in a distributed fashion with the power grid just helping to load level.

One of the other reasons nuclear power plants in particular have tended to be large is NIMBY. The amount of opposition and regulatory burden you get to a 5GW plant is about the same as you get to a 0.5GW plant, but then you get ten times as much power from it.

There is now a new theory to handle this, which is to build smaller reactors and put them on barges or railcars. Then you can mass produce them in one place with one regulatory approval and the time it takes to go from there to supplying power to a new city is the 24 hours it takes to tow in the barge. Which deprives NIMBYs of one of their most annoying obstruction methods -- raising construction costs through intentional red tape and changing requirements after construction has already begun. That can't happen if you can go from local approval of operations to selling power in the course of an afternoon. And then you can't lose your investment due to an unfavorable or changing regulatory environment because in the worst case you tow the barge to some other place with lower hostility.

Hopefully that will work better.

Right up until the barge sinks or crashes and the spilled waste renders an area uninhabitable for 40 years.

This is the real reason why fission reactors are either large, or military. The organisational capabilities that are required to manage them safely require scale and funding.

Not necessarily, a cargo container sized unit would have a limited quantity of fissile material making it more readily dealt with by proper authorities. Given newer self-shutdown models that don’t go critical or meltdown tons of highly reactive material on failure removes issues similar to Fukushima where melted nuclear material erodes containment and embeds into the ground and/or seeps into ground water. During transport, if the system is off, you could end up with readily contained debris given the containment system is designed to withstand external forces (think black boxes, but with nuclear material). Nasty still but much more controllable. I’d argue it’s on par with some of the nastier chemicals already routinely transported.

Another benefit of smaller scalable units would be the ability to readily transport aging units back to a central factory that routinely handles old units before major structural degradation occurs. One issue at Fukushima (and a majority of other nuclear reactors) is that the size and scale of each one is enormous requiring a lot of one-off resources to clean up, and so the incentives are to run them far past designed lifetimes and safety recommendations. Nuclear fission is very detrimental to materials, even in modern failsafe designs. But having an organization which routinely “recycles” units long before failure could allow both economic efficiencies, better safety due to constant practice, and dealing with smaller units which could be largely handled with industrial robotic systems.

The issue I’d worry about with such units would be security. Smaller units could mean potentially less security which would make for easier targets for motivated terrorists or rebels.

I believe there used to be a number of small reactors back in the day. But some poor design issues saw them canned.

https://spectrum.ieee.org/tech-history/heroic-failures/the-f...

McMurdo ice camp in Antarctica briefly was run via contained nuclear reactor.

I also wouldn't call it a Nimby issue. If a disaffected terrorist decided to blow up a small nuclear unit the issues would be considerable. Currently nuclear power plants are guarded like wartime POW camps.

The army did also try to build portable nuclear power generators

https://youtu.be/HPWDMHH4rY4

If they can make it cheaper than a similarly sized fossil fuel plant, they‘ve got a market. The military would be interested, too. That new russian nuclear power plant on a boat can’t have been cheap, either.

Fission power plants are not now competitive even if the nuclear part were free. The turbines, generators, cooling towers, etc. make them too expensive. So unless fusion can do away with those (that is, use direct conversion of plasma energy into electrical energy) it's unlikely to be competitive either.

Fusion reactors, unlike fission reactors, are best in the smallest workable units. That's because they are limited by the power/area at the first wall, so (by the square-cube law) their volumetric power density gets worse as one makes them bigger.

How far can this premise be pushed? And will I one day be able to get it in a surface mountable package?

Unfortunately, there is a lower limit set by the need to absorb neutrons. The blanket has to be roughly 1 meter thick.

One might be able to evade that if one could have a large number of tiny, or at least thin, reactors sharing a common blanket, much as the fuel rods of fission reactor share a common moderator. However, it's still likely to be inferior to a fission reactor in power density, as these tiny reactors would irradiate each other.

Beyond that, there's the need to confine the plasma to these very small reactors. Even in complete absence of plasma instability and turbulence, ions will scatter off each other and diffuse out and be lost. This classic diffusion takes time proportional to the square of the minimum reactor dimension / ion gyroradius.

There is not much substance in this article. The most interesting part seems to be the slide with the milestones. They seem to be three or four prototypes away from ignition. How many years is that? How many years from the ignition prototype to something that actually produces more energy than it consumes?

Well, based on the last 50 years of fusion research, the time from 'now' to 'net positive energy production' is quite consistently between 20 and 50 years.

I think everybody knows that joke by now, but also most people know that fusion research has never received the funding necessary to make fast progress. Given the lack of enthusiasm for funding it, it is in fact surprising that ITER is not cancelled yet.

That's kind of a tautology - because there's no substantial progress. The real problem is that every time the fusion community has tried something some unexpected effects and behaviours have emerged derailing the technical approaches and sending the community back to the drawing board.

Yep, if governments had committed Manhattan project scale funding to Fusion say 30 years ago, it would almost certainly have gone into a technical approach that is now thoroughly obsolete. It simply wouldn’t have worked, because we just didn’t understand the problem domain well enough. IMHO we still don’t.

The pragmatic approach is to fund a variety of research projects enough to keep making progress, but not so much that too many resources get wasted on dead ends. A dead end is fine if we learn from it, but we should only commit the resources needed to learn the lesson, and as little more as possible.

Progress on the triple product :

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

has gone up faster than Moore's Law :

https://i.imgur.com/BN0pz.png

The problem is that the next device costs a lot more because it has to be much bigger, that is unless you have much stronger magnetic fields. The cube of magnetic field is proportional to the energy gain in a Tokomak. See this video at at 46 minutes to get the equation, watch more to understand why people are now doing this.

https://www.youtube.com/watch?v=L0KuAx1COEk

Tokomak Energy and Commonwealth Fusion Systems among others are looking at smaller reactors that use High Temperature Superconductors.

Elon announced they will reach net positive by 2023 and the first Tesla powered by a fusion reactor by 2026

Elon has announced no such thing.

Even if it takes another 50 or 70 years to get there,people should still be pushing for it.I wonder how this would change the political situation across the world when oil,gas and coal wouldn't be required so much anymore...

If we still require oil, gas and coal in 50 or 70 years we're fucked anyway.

Yes, it will mark the end of world war 3 because humanity will be able to survive by traveling to a new planet in a different solar system. Deep space exploration becomes viable because we are no longer dependent on solar panels, instead we carry our own mini sun within our spacecraft and simply sail forth!

Each prototype will be more complex and more expensive than the previous one and it will take longer to build. Current research devices have enormous infrastructure requirements (Wendelstein 7-X and ITER both span a whole campus of several buildings for power supply, plasma heating, cooling etc.). If Lockheed wants to scale up, they will eventually have to build all of that as well.

The Lockheed design is exciting precisely because it offers compact reactor designs. The article talks about power plants the size of shipping containers.

I believe that when I see it. It would be cool if that works, but what I know of the physics that are involved stands against that.

agreed, but they do get to discard the constraint of needing to make money on the energy at market rates. I'm guessing this will go on some eternally loitering aircraft like the nb36 but without the need for the 3 tons of shielding that made it unviable. or maybe that is just me wishing. probably will power haliburton contractor cities in the middle East that were a couple miles outside the existing grid so they "need" a billion dollar generator

A fusion reactor that is operating needs shielding against neutron radiation much the same as a fission reactor does. So a fusion reactor on a manned plane still has the same mass issue even if it were compact enough to fit.

The physics are discussed at the bottom of the article. There's an inherent 20x efficiency improvement in containment pressure over Tokamak designs.

The quite in the article isn't particularly helpful. Thisnis one of the cherry-picked quotes that journalists love to use that sound imagimative and colorful, but explain exactly nothing.

So the field has a self-refulating field containment in the radial direction. This was never a particularly crtitical issue. The issue is building non-leaking mirrors at the end caps where the field needs to separate from the confinement to loop back on itself. You can only shape field gradients there so that they form a potential that particles on escape trajectories have to either overcome or be reflected by. I forgot the exact details, but the issues is that electrons are harder to confine and they gradually build up an electric field that attracts the protons and thus counters the mirror gradient of the magnetic field.

I lost it, but a physicist who has worked on this stuff posted here once explaining how incredibly over exaggerated progress on fusion is and there is a real liklihood that little will ever come of it.

Link?

The amazing thing is that Lockheed is doing this with their own money. Not for PR; they seldom say much about the project.

People running this project (Tom McGuire) are aerospace engineers who are familiar with plasma in hypersonic speeds and computational fluid simulation. It seems that they don't have experienced particle or plasma physicists in the project.

Can anyone correct me on this?

So, it's a design using magnetic mirrors? I was asking what happend with this way of building a fusion reactor. The last that I know is that was used on the 60's.

Tokamaks are the direct evolutionary step up from magnetic mirrors because these had problems with with electron containment at the mirrors. A circular design removes the mirrors and therefore this problem. Unless these guys have completely novel mirror configuration, they must run intonth same issue because electron and proton trajectories are quite different at these energies.

"we have a fusion reactor, she is real. You just never met her because she goes to a different school"

I tip my hat to them. We need some boldness to move forward.

Reply on news.ycombinator.com