Ahh, Wendelstein is that stellerator reactor. The stellerator is really cool, and an alternative to a tokamak reactor. Tokamak is the doughnut shaped reactor, and it has a problem where the plasma near the outer circumference has less magnetic confinement. The stellerator is similar, but confines the plasma to a ribbon and folds it over on itself in a mobius-like arrangement.
I used to be really interested in this, but forgot it existed over the years. Glad to see it works!
Even just reading "mobius" and I cannot not hear the line from Warf "There is this theory of the mobious. It twists in the fabric of space where time becomes a loop. Time becomes a loop. Time becomes a loop." Of course, I'm referencing the Orbital track
I didn't either, since I spelled it Warf! I never watched Star Trek (probably get my HN privileges revoked for that), but I did know that much. And then the Reading Rainbow guy's "whatever happened will happen again". I know the character's name is Geordie or whatever, but he didn't have the right accent for him to be a Geordie, so he will forever be Reading Rainbow guy instead.
I think you're not considering the first album with an actual track called Mobius. You're thinking of the opening cut on Orbital II (the brown album). That one is the one where the "where time becomes a loop" becomes fun. The Mobius track is the one where "whatever happens, will happen again" is also included. Sorry, but I will have to deduct Orbital fanboi points, and you will lose your next turn
It's two copies of "...where time becomes a loop." One on the left, one on the right. One of them is slightly shorter than the other, so that as they repeat again and again, they slide in and out of phase. The track ends after they come back into phase.
If you listen to this with headphones, or speakers with decent separation, paying attention to this feels interesting. It's similar to the way listening to "binaural beats" can do interesting things to your brain.
Also if you are in the habit of putting an entire album on repeat and this is one of your favorite albums, then you've probably heard this a zillion times. If you have your music player set in "randomize by album" mode, then, well, it's the first track on this album, so every time it comes up you'll hear most of it unless you instantly decide you are not in the mood for Orbital 2, and even if you're not in that mood it may be pleasant to let everything come back into phase before going to another album.
----
The next track on the album starts up entirely in the left ear, with a tinny, distant little loop, and the words "Even a stopped clock tells the right time twice a day". Once it brings in a deep bass, this bass is also doing some weird cross-ear phasing things.
And then the third track also opens with "Even a stopped clock..."; a theme has been established at this point. Time is a loop, and a stopped clock is right twice a day. The opening of "where time becomes a loop" is also a bit of a joke; Orbital's musical craft is very much about making a bunch of short loops that work together, and bringing them in and out over each other for four to seven minutes. Occasionally as much as thirty minutes, the extended version of "The Box" is glorious. This is something that utterly dominates most electronic dance music now, but Orbital was one of the first notable acts to really go hard on this, and this is their second album; they are saying "yes it's just more loops, we think they're good loops, enjoy!".
By the time you get to the last track, you've probably forgotten about Worf's repeated mantra. Especially if it's your first time listening to the whole thing and Halcyon + On + On just blew all the cobwebs out of your head. But Orbital returns to the idea, with two different loops that are very close in sound and length, played on both channels: "Input Translation"/"Output Rotation". They begin in phase with each other, drift out, and come back together. And the album is over.
Or, if you have the CD player on repeat (remember, this album is from a time when people bought CDs and probably stuck them into a one-disc player, maybe a 3 or 5-disc player if they were lucky, and the whole album is built with an awareness of this), you're back where you began, inputs translated and outputs rotated, and ready to be reminded of the Theory of the Moebius.
Time has become a loop. Come out of the trance Orbital has put you in. Do you want to experience this loop again? Does it feel rude to jump to another album before Worf's come back into phase again? You may as well let him get you back in sync with the moment the album began before going back into normal time.
>remember, this album is from a time when people bought CDs and probably stuck them into a one-disc player
You just describe the first time I every danced with Lucy. I went back and forth with this disc and The Orb's A Huge Ever Growing Pulsating Brain That Rules from the Centre of the Ultraworld.
BEST HN THREAD EVER (and WOW is that an essential explanation of the first two Orbital albums)
is this where we talk about
- the 39:59 mix of The Orb’s “Blue Room” (which, like several other Orb singles, is better than the album version thanks to Jah Wobble providing a proper bassline)
- the academically-verified lack of repetitiveness in Autechre’s “Flutter”
- and Orbital’s “Criminal Justice Bill?” on the “Are We Here” CD single, which is four minutes of silence
Funny thing about Blue Room, but that was the first time I had ever heard of Haile Selassie, then I was told that sample is a comedy bit from a prank call or whatever.
The Orb Live '93 CD was one of those that if a CD could wear out like a cassette, that would have been one (two technically) that would have from my collection.
I find the geometry of things like this fascinating. We typically think in such simple shapes. I feel like my brain can do triangle, rectangle and maybe hexagons and that's about it. I remember when I finally understood radians enough to really understand circles and waveforms—I felt so enlightened. Like I actually remember the moment when it clicked. For years I was just "doing the work" without actually understanding what I was doing, but once I was able to understand it… it's like something changed in my brain.
I want to be able to think in mobius, but my brain is currently like, "No thanks."
Doing CAD design is really interesting. A lot of stuff is just 2.5D, extrusions of 2D sketches sitting on other 2D sketches.
Then you accidentally make something truly 3D by intersecting things and realize you have no idea what you're looking at, couldn't imagine it if you closed your eyes, couldn't replicate it if you had a picture of the result and didn't know the 2D inputs that made it... and then you realize there are probably people out there who can see that entire design in their head.
To me it's like unicycling on a tightrope or skateboarding or realistic oil painting or playing piano well. I have no real concept or reference point for what that experience must be like.
Interesting. With a very strong reference for the experience of skateboarding, reading your sentence made me think about how hard it is to explain! I suppose when it’s going well, it feels like body/mind flow, when it’s going poorly it feels like physics :)
That kind of thing is super amazing! I've secretly suspected for a while that people who are able to skateboard or do similar things must live in 4 or 5 dimensions.
I can't figure out any kind of way someone could do that with anything resembling what I understand "thought" to be like, so I assume you must be able to process entire sequences and their alternate possibilities simultaneously?
Coming from 3d sculpting to CAD was...eye opening. It really makes you realize how much of design constrained by the CAD tools, which are mostly constrained by manufacturability.
If true 3d printing* ever gets cheap, it'll be interesting to see how much form will be able follow function, rather than manufacturing cost.
* true 3d, as in overhang are allowed. Something like a cheap FDM is more 2.5d, since overhangs aren't allowed.
Well, there's SLS which is tolerant of overhangs and other exotic features. Formlabs has a machine in the 5 digit price range, and I wouldn't be surprised to see competitors come out with a 4-digit one in the near future.
CAD is hard! I think about this a lot as I'm doing a makerspace at an elementary school and adults think it needs a 3D printer, but what are kids going to print on it?
I think for me, the prerequisites for mastering CAD were 1) the practice I got visualizing 3D shapes so I could translate them into unambiguous mechanical drawings on paper (I swear, I'm not that old but my college was behind and we were the last class to actually do mechanical drawings with a pencil), and 2) having a procedural thought process from coding so that I could sequence the CAD operations to get where I want.
For me CAD is pretty much the opposite, I rely heavily on the apps, I spent about a week trying to learn to draw before deciding I didn't really want to spend months or years on it.
I'm not a mechanical engineer, I mostly only do incidental CAD and hobby level work, so it's not really essential to have the deep understanding of space that real MEs need.
I often don't know what sequence of operations I'll need until I actually open the app. Generally it's more of an "Oh I need a mounting hole, lets look around on the screen and see where one could go" thing, a lot of the thinking is in the app rather than in the mind.
Of course you can't make nicely parametric things without a lot more thought so I will often wind up having to redo things that aren't one offs...
Grasshopper for rhino is awesome too. It's actually really relevant for this topic. It's the only syste I know that integrates so well into a precise cad design system and also allows you to do complex stuff easily. I love it. Drawing a stellarator precisely is probably not that hard on it... Will see if I can throw something out quickly today.
I can't really think in (visualize) 3d shapes, so I depend a lot on 3D geometry programs when I design things like for my microscope. A fair number of people I've talked to can visualize complex shapes in their head, rotate them around, do interference checking, etc.
What about those tests where they say to fold a piece of paper X number of times, and then punch holes in specific places. You then have to pick the image of what the paper would look like unfolded. Do these types of visualizations give you the same issue? I have known several people that just could not visualize these tests, and I'm curious if these are related.
Yea they are kinda confusing, especially when you get into the 3D ones like klien bottles and roman surfaces. I also recently learned that if you make a mobius shaped transmission line (like a ladder line) and you send a pulse down it, that pulse will continue looping until it dissipates (or forever if it is a superconductor).
Are real life superconductors ideal, i.e. they truly would store a charge forever (at least until the material disintegrated)? Or is there some sort of loss, albeit much less than typical resistance?
So long as you do not exceed a certain current, they have zero resistance. There are other forms of parasitics though that can effect it.
Those crazy electromagnets they use on these stellarator are simple superconducting loops that they 'charge' by inducing a current. That current is maintained so long as the superconductor stays below a certain temp. There is even something called a SMES (superconducting magnetic energy storage) that stores power this way, as I understand it they have a 0% self discharge rate.
It is my understanding that electrons on a curved trajectory will lose energy to electromagnetic radiation. Is that not the with electrons in a superconductor?
If you have a single free electron, and move it on a curved trajectory, you’ll see electromagnetic radiation.
If you have a hypothetical continuum of charges moving in a circle, and you use Maxwell’s equations, you’ll find that given a constant charge density, no EM radiation will be emitted.
If you have a superconductor, you should really be using quantum mechanics to understand it. If you can imagine that an electron can “orbit” a nucleus without emitting EM radiation, then you can imagine that current can flow in a loop without emitting EM radiation. The behavior cannot be explained by thinking of the behavior of a single electron, but must be explained by considering the behavior of many electrons in a quantum mechanical system.
Also note that the actual speed of electrons (the “drift velocity”) moving around a loop of wire is extremely low, so if you treat electrons as point charges and ignore quantum mechanics, and you calculate the amount of EM radiation that should be emitted by a typical loop of wire, you will get an extremely low amount EM radiation emitted, which would be very difficult to measure.
1. Something “in a superconducting state” has zero resistance, but the transition to superconducting states is not sudden, and there are various things which disrupt that state like magnetic fields.
2. Alternating currents will dissipate even with zero resistance, because the circuit will emit EM waves.
>I want to be able to think in mobius, but my brain is currently like, "No thanks."
I spent a lot time staring at Escher images as a teen, so I think my brain says "yes please". I have no idea what to do with any of it, so it's not like it does me any good.
I'm not that guy, but I can speak to what you're asking. I've followed Wendelstein 7-X for almost a decade.
Nuclear fusion occurs at extremely-high temperatures. As you heat your fusion fuel to sufficiently-high temperatures to allow fusion, the matter transitions into a plasma, which is great: plasmas react to electromagnetic fields. As such, a major challenge with achieving viable nuclear fusion is making a vessel capable of holding the fusion reaction. Because we can't create on-demand gravity wells, the next best option for confinement is using electromagnetic fields to hold the plasma in the air.
So, you now have an "electromagnetic bottle" capable of suspending a fusion reaction above the reactor's walls. Now, you have another issue: how do you ensure the fuel will sufficiently mix to sustain a fusion reaction? One approach is to move the plasma in a loop. The topologically-simplest method to accomplish this loop is the torus. Such a plasma-confinement device is called a tokamak. A tokamak uses two magnetic fields, torodial and polodial, to accomplish its task. The torodial field is driven through the plasma to push it forward, while the polodial field pulls the plasma in toward the center. Proper balance of these fields will allow the plasma to circuit the vessel following a helical path, achieving confinement.
However, driving two separate magnetic fields is energy-intensive, and a successful fusion reactor will want to minimize its own power consumption to maximize the amount available for external usage. Enter the stellarator. The stellarator also drives the plasma around in a circle, it but uses a single magnetic field. How? It "tricks" the plasma into "thinking" there's only one magnetic field by using computer-optimized magnets with highly-complex geometries. This provides stellarators with a major engineering advantage over tokamaks and is a primary reason Wendelstein 7-X would have chosen it.
With the confinement vessel topology largely identified, the next main step is to figure out how to build a vessel able to contain a sustained fusion reaction. For context, fusion experiments traditionally only operate on timescales of milliseconds to maybe a second. The reason? Fusion occurs at millions of degrees, and keeping the reaction vessel cool, ensuring a continuous supply of fuel, and dealing with reaction "exhaust" (e.g., alpha particles) and stray high-energy neutrons from the common deuterium-tritium reaction (which irradiate your reactor walls because neutrons don't react with electomagnetic fields) is a major, major engineering challenge. Any operational, net-positive fusion reactor must be able to operate for days, weeks, and months on end.
What Wendelstein 7-X has been attempting to do for years is demonstrate that building such a vessel is even possible. Their overall goal is to sustain a fusion reaction for about 30 minutes. Such a timescale will show a proof-of-concept system which enables sustained fusion reactions to occur.
Currently, the preferred fuel is deuterium-tritium because the fuel is generally available and has an attainable fusion temperature. The stray neutron issue can be mitigated by lining reactor walls with lithium to breed tritium fuel. Even better is to use the helium3-helium3 reaction, which completely annihilate to produce pure energy as the output (welcome to e=mc^2, enjoy your stay). The main holdups are: (1) the reaction occurs at much higher temperatures than deuterium-tritium, and (2) he(lium)3 is quite scarce on Earth. Once Wendelstein 7-X shows how to engineer a proper confinement vessel at a "lower" temperature, you can then work on the higher temperature levels required for he3-he3. Also, he3 is plentiful on the surface of the moon, so mining the surface of the moon will be performed to obtain the required fuel, which is the fundamental premise of the movie "Moon".
Someone asked for information on electromagnetic plasma containment folding. I recommend reading up on magnetohydrodynamics (MHD). It's the mathematical and physical foundation of your interest.
> Even better is to use the helium3-helium3 reaction, which completely annihilate to produce pure energy as the output (welcome to e=mc^2, enjoy your stay).
It doesn't completely annihilate to produce pure energy. It produces helium-4 and two protons. Or you can react helium-3 + deuterium to produce helium-4 and one proton. The point is that helium-4 and protons are easier to shield against than neutrons, don't turn your reactor radioactive, and at least in theory their energy can be extracted directly (eg through induction) instead of through heat.
Edited to add: except helium-3 + deuterium still produces neutrons, because sometimes the deuterium will react with itself to produce helium-3 and a neutron.
You're absolutely right, thank you for the correction. he3-he3 doesn't produce neutrons, which is the major advantage, in addition to the massive energy output.
One guy asked why the mobius aspect is needed and I couldn't answer. I know a lot of stellarators aren't odd-period like Wendelstein, and the old designs didn't do folding at all. Do you know what improvements the mobius design has over something like TJ-II?
The helical path creates a twist in the plasma which cancels out the drift forces. This is what I meant by "tricking" the plasma. User mjfl gives an even more technical explanation:
> By twisting the plasma into a shape where the curl of B (proportional to J) is parallel to B, i.e. a helix, the cross product is 0, and thus there are no net magnetohydrodynamic forces on the plasma.
Hope all that's a good answer for you.
> Mobius aspect
You might avoid using the word "Mobius" and instead use "helical." A Mobius strip is important because it has two faces which form a single surface. The surface aspect isn't relevant in this context, so a term which refers to the shape would likely dispel confusion in a reader.
As far as I'm aware, each section of a stellarator is periodic in its own right, which means the end and start points of each section are the same. Though I'm not certain, the choice of four versus five is more likely an engineering factor rather than one of physics, whereas the distinction between a tokamak and stellarator is of physics and not just engineering.
If a 'particle' (I don't know a better word) finds itself near one of the top divertors, at the same point in the next orbit it will find itself near the bottom divertor. That is a product of the "mobius-like" shape, so although it isn't really a 'ribbon' and isn't really a mobius, it helps explain the concept concisely. I just don't know WHY that shape helps lol. Maybe it doesn't and it was just a practical design change like you said.
Did a little research to try and understand this better.
The most precise term to describe the "twisted ribbon" flux tube in W7-X is "toroidal helix". The toroidal quality comes from the general torus shape of the stellarator, and the helix quality comes from the twisting of the magnetic field by magnets. (The torus shape is required only topologically; look up the knotatron to see what I mean.)
The "ribbon" we're talking about is properly called the flux tube. The flux tube is the volume created by the flux surface, which is where the magnetic field lines lie. A given volume of plasma contained within a flux tube should remain inside it, causing magnetic confinement of the plasma.
The optimality of the confinement of the flux tube is expressed with the term "omnigeneity". Conceptually, a flux tube has onmigeneity if ideally all of the non-colliding plasma inside the tube stays in the tube. W7-X's flux tube appears to be approaching omnigenity. (Another experiment which approaches omnigenity is HSX. Interestingly, HSX has one set of primary magnets, whereas W7-X has two. That's likely because HSX achieves omnigenity via quasisymmetry, whereas W7-X uses various stellarator optimization techniques.)
With these points, we can call the W7-X "ribbon" a near-omnigenous toroidal helix flux tube, which sounds way cooler. So, all that said, why is a helical property desired? From what I've read, the twist in the flux surface reduces plasma drift inside the flux tube.
I think it makes sense to analogize this stuff as a circular semi-permeable pipe filled with a high-pressure "magic fluid" flowing around-and-around inside. By semi-permeable, it means fluid will leak from the pipe if the internal pressure is too high (remember that this is magic fluid). Trying to understand the helical twist along this analogy, I think the effect is evening of internal pressure across the pipe surface to reduce fluid turbulence and permeation while maximizing laminar flow. At least, that's my best analogous interpretation of "why" the twist helps.
The divertors are useful for long-term reactor operation but have no direct relevance to the magnetic field geometry. I'm guessing there's two divertors for engineering reasons (performance, redundancy, etc.) and not for reasons of basic physics.
Yes I was referring to odd. But I've been reading some papers and I think even 4-fold had the mobius effect. I'll comment here again tomorrow when I have learned a bit more on the topic.
>Also, he3 is plentiful on the surface of the moon, so mining the surface of the moon will be performed to obtain the required fuel, which is the fundamental premise of the movie "Moon".
This is repeated a lot, but the practicalities are... questionable. Here's an article I consider to be the definitive criticism of the concept:
Seconded. This is fascinating stuff and reminds me of some crazy rant some guy was telling me about anti-gravity and how electromagnetic “ribbons” could propel you. Obviously the guy was nuts, right? How would one go about learning more about electromagnetic plasma containment folding?
Haha, no I am just some random guy who reads too many Wikipedia articles. "Electromagnetic plasma containment folding" does sound like something a crazy person at a bus station would rant about.
My explanation was definitely over simplified, but I'm not knowledgable enough to go into detail on the topic. I can't even point you towards something to read on the topic since everything I read about it is like 15 years old at this point.
“…the supporting structure can only withstand the forces if the interfaces between the ten individual segments of the central rings, which weighs several tonnes, are built with a level of precision of less than 100 millionths of a metre…” - and they found a small family business in the north of Italy capable of doing this!
thus, 100 millionths of a meter = 0.1mm, or ~4 thou in American units. Easily achievable by hobbyists, let alone by serious, professional equipment.
Sure, that is a pretty exacting specification for what I suppose is a big machine, but I'm pretty sure very normal things like say, car engines get made to far tighter tolerances.
millionths of a meter are known as micron so most people would call this '100 micron' (or '100 micrometers') which is indeed close to 4 thou, as you calculated, and is the level of accuracy of my ~$500 3d printer.
1 thou was achievable in routine shops in the 1940s and a tenth of a thou (2.54 micron) is a common accuracy to target these days. Obviously it depends on the context and the size of the object, at some point you move away from cutting to using grinding and lapping to achieve your results, which is ultra-timeconsuming.
What usually matters more than absolute tolerances is relative tolerance, aka ppm. 100 micron / 4 thou tolerance can be achieved with hand tools and a bit of patience on the benchtop scale, say a 4" part. That's about 1000 ppm, or 0.1%. If I gave you a meter stick, you could probably eyeball marking something +/- 1mm.
Getting the same finish on a 120"/3m coil is 33 ppm. 100 ppm / 0.01% for any operation or process tends to be where things start to get really challenging. Deflection goes up by the length cubed, so increasing the size of all the tooling relative to the tolerance gets really challenging really fast.
> Dr. Ning Li of Huntsville, AL passed peacefully away on July 27, 2021. She was 79 years old. One of the world's leading scientists in super-conductivity anti-gravity. Dr. Li had constructed first 12" HTSD of the world in late 90s.
I just read a bunch of stuff about her from her son. About how he took care of her in her advanced years and her alzheimer's disease. Sad but also peculiar about her DoD work and how she "never talked about it". I wonder what it was? trying to get an alien craft working again? developing anti-gravity weapon? a ship? a hoverboard? please say it was a hoverboard.
Her claim that "You can take a bowling ball and place it and it will stay." is fascinating. I would love to see footage/video of this. Small electro marbles and globes are one thing, a bowling ball or other large non-magnetic object!? man oh man!
Just curious as to why a Mobius strip type arrangement is better than a toroid? Is it anything to do with the turbulence in the plasma flow being easier to control?
Draw a torus and then draw rectangular "bands" across it, they will represent the containment magnets.
Due to pure geometry, the area closer to the center will have a smaller distance between bars. This means that the magnetic field will be stronger near the center.
This in turn means that particles will separate (depending on charge) and drift to the sides. It seriously interferes with the containment.
You can fix that by changing the torus into something resembling "8", so that particles move to one side when they fly through the upper part, but then they'll move back as they fly through the lower part.
Of course, you can't just do that in 2D because the part in the middle of "8" will have no magnetic field. You need something without self-intersections. You can try to move one side up and another down. But that doesn't quite work either because you will get another set of preferred directions.
So instead you go with the gentle twisting, resulting in the Möbius-looking shape.
These are great questions for someone more knowledgable, but as I understand it, If you follow a single point on the surface all the way around the loop, it will spend as much time in high confinement as it does in low confinement.
That explains why folding is important, as for the mobius, I oversimplified a bit. The Wendelstein has 5 folds, making it a mobius, but I think I read about one in Spain that had only 4 folds. That would mean the mobius isn't imperitive, but I'm sure there is a good reason for it.
Really a stellerator doesn't need 'folding' at all, they can be as simple as a twisted torroid. I didn't want to go into excruciating detail though, the more in detail I go the more likely I am to say something that is wrong lol.
Edit: I looked it up, the one in spain is called "TJ-II"
No, twisted pair wires are really cool but different. When you push current down one wire you pull current down the other. The signal is passed through the differential of those 2 wires. If the wire is hit with EM interference, that change will be seen as a 'common mode voltage', that is, both wires will be 'pushed' or 'pulled' the same amount, and you won't see a differential.
That effect works both ways too, where a single wire with a digital signal will spew out radio waves, 2 wires with opposing signal cancel each other out and emit no em waves.
The effect with the stellarator is more like stirring a pot.
It primarily has to do with the physical construction of the magnets, in a toroid the inside of the toroid effectively has more windings per meter of circumference than the outside causing uneven containment.
With mobius strip you regularly flip between inside and outside, so the plasma particles get more even force applied.
I think it's about ensuring the plasma heat/energy distribution is more uniform so you get fewer outlier particles with high enough energy to escape confinement and damage the interior of the reactor. Or something like that.
What's interesting is that stellarator actually is not just an alternative, but a wholly parallel branch of evolution - it's not like one was invented strictly after another, and the authors of both designs never knew about the other's work before they completed theirs.
What's even more interesting is that the fusor - the simplest possible design for a thermonuclear reactor, so simple that anyone skilled in electrical engineering and having access to proper civilan equipment can build one with ease - seems to be invented _after_ both stellarator and tokamak.
That said, I never particularly liked stellarator design. The very _complexity_ of it somehow feels subtly wrong, like doubling down in the wrong direction.
However, this is one of the cases where I would absolutely love to be proven wrong. We are far past due big breakthroughs in the field.
Compared to a tokamak, the stellarator bring engineering efficiency while matching performance. Though ideas for the tokamak and the stellarator may have emerged together, the main reason tokamaks were built first is because they COULD be built. Without computer-assisted magnet design, stellarators simply couldn't be properly built; the magnetic geometries are just too complex.
In the long run, it's not known stellarators will be the eventual winner in the long race for a viable fusion reactor. The attributes in a winner will be net-positive operational efficiency and superior energy harvesting abilities. Perhaps multiple approaches will be viable.
I understand the theory. I just hope I will live long enough to see a winner in this race.
To be honest, I've been interested in the ___domain for quite a time and I still want to build a fusor or a polywell at some point just to see it glow. Probably won't happen though.
> The very _complexity_ of it somehow feels subtly wrong, like doubling down in the wrong direction.
This made me think of modern jet fighters being designed to be aerodynamically unstable, making them all but impossible for human pilots to operate without flight computers. Apparently the maneuverability benefits make the added complexity more than worth it.
> The very _complexity_ of it somehow feels subtly wrong, like doubling down in the wrong direction
I read an article a while ago which sold me on the stellarator that said something like: "The tokamak has magnets in a configuration that gives simple engineering but hard physics, whereas the stellarator has hard engineering to make a complex magnet but that results in simple physics".
The engineering is a much more understood beast. It was still fairly novel as they had to have a computer do the design of the magnets, but that is now a solved problem. But then if that allows us to simplify the (very difficult and novel) physics it feels like the "obviously" correct decision.
The other thing that makes me a stellarator fan is that the JET/ITER work is later and more expensive that predicted at every stage. The W7-X provided a plan for the runs they wanted to do and upgrades to the reactor and they have basically run entirely to schedule.
The big plus of stellarator design is inherent absence of plasma instabilities affecting tokamaks. Notice that future upgrade of Wendelstein may allow to hold plasma for a hour compared with minutes at best with tokamaks. Many physicists for that reason believes stellarator is the only way to archive practical fusion.
It not only works. It beats tokamak reactors by orders of magnitude to the point that it and ITER are the only fusion reactors that even matter, so saying that you forgot about one of the most important fusion reactors in history is highly condescending.
The Stellarator is theoretically a superior design over the Tokamak, designed to neutralize the JxB force, where J is the current through the plasma and B is the magnetic field guiding the plasma around the device. By twisting the plasma into a shape where the curl of B (proportional to J) is parallel to B, i.e. a helix, the cross product is 0, and thus there are no net magnetohydrodynamic forces on the plasma.
'Theoretically' is the right word for sure. iirc, the predecessor of the Wendelstein led to the bankruptcy of the engineering firms building the parts, because tolerances were so tight and they failed multiple times to land within the constraints.
The first Alternativlos Podcast of two conducted with the leader of the Wendelstein X project relates how certain magnetic coils were purpose made by a kind of emeritus engineer in Swiss. Which seems corollary to the common wisdom about EMI, it's magic, so the manufacturer must be a wizard.
On the other hand it is reminiscent of a Georgian I met who used to be occupied with winding regular tire sized coils by hand, for over land transmission lines. This is chirurgical precision, literally hand-craft.
We do a lot of thinking with our hands. It stands to reason, metaphorically speaking, that Wendelstein is an experiment to gain hands on experience. Therein lies the difference to megalomanic projects that exceed initial estimates, eg. BER airport, which are a running gag by now.
Insolvency means the investment returned no profits so investors on those projects stopped paying. It likely doesn't mean that the cheques bounced on liabilities. And it obviously doesn't mean that investment in this space had to stop.
true. but on the other hand, the 'theoretical' is being turned into practice as evidenced by this 8 minute containment. the best a tokamak can do is half a second.
They could go longer. 7-8 mins is an arbitrary cut off out of fear that something might break (quench), knowing that enough data has been gathered for the time being.
W7-X is a new stellarator design. It’s magnet arrangement was optimized using a lot of compute time and is designed to overcome the faults of previous more naive designs.
The last time it was in the news I think naysayers listed the main caveat with stellerators as something along the lines of very low plasma density compared to tokamaks, which makes them unable to get anywhere close to the energy break even point.
The main problem with stellerators is their murderous complexity. You need to manufacture several thousand different parts with complicated 3D geometry, micron-level precision, and from unobtanium-class materials.
All while not being able to properly simulate the outcome on a computer.
Stellarators are superior to tokamaks, so an energy-positive stellerator will be about 2 times smaller than a tokamak. But we're still talking about a building-sized vacuum chamber.
That's why for ITER it makes sense to go with a simpler design to de-risk the main objective: building a burning plasma laboratory.
Doesn't really track. Stellarators can operate above the Greenwald density limit. They just have shorter confinement times for a given field strength and major radius.
Found it: "Energy turnover is defined as the amount of heat multiplied by the duration of the discharge[1]." By "amount of heat" I assume they mean "heating power delivered to the plasma" b/c the the only way to multiply by time and get Joules is to start with power.
> The energy turnover results from the coupled heating power multiplied by the duration of the discharge
The numbers:
> The energy turnover of 1.3 gigajoule was achieved with an average heating power of 2.7 megawatts, whereby the discharge lasted 480 seconds
Also:
> Within a few years, the plan is to increase the energy turnover at Wendelstein 7-X to 18 gigajoules, with the plasma then being kept stable for half an hour
Possibly the poster is from a country like America where "turnover" is not a preferred term to refer to gross receipts of a business. (Americans typically use "revenue" instead.)
The 1.3 GJ is how much electricity they put in. There is no fusion reaction or energy production, it's just an experimental device to study plasma containment.
There has never been a net-positive-energy magnetic confinement fusion experiment. Inertial confinement fusion has had 2 events that were "more energy out of the fuel than delivered to the fuel." But is still about a factor of 100 away from what is needed for "more electricity in than out"
This is perhaps an obvious question to some, but I'll ask it anyway: How is the power generated here converted into usable electricity?
I know for conventional fission reactors the heat of fission is basically used to run a steam turbine. Given the extreme heat of the plasma, and that it must be magnetically suspended so that it doesn't even touch the sides of the containment, how is that heat transferred to some other medium to generate electricity?
Most of the answers are missing the actual (proposed) mechanism.
The energy of the reaction is mostly carried away as high-energy neutrons. So, the way to get energy back is to "capture" those neutrons. Since neutrons are not electrically charged, you can't use them to directly create electricity, so all you're left with is using them for heat.
Unfortunately, since they are electrically neutral, they're also relatively hard to catch. You need a dense material where they will have a good chance to hit some nucleus. The proposed designs are typically some kind of liquid metal blanket being circulated around the reactor and onto a place where it can boil water to produce steam to spin a turbine. Lithium is the metal most proposed for this, since it also has the advantage that it can produce tritium when bombarded with neutrons (tritium being the super rare half of the fuel that goes into the reaction).
It's always fascinating to me that, no matter how many interesting new ways to release lots of energy we develop, we are still stuck with the same method for converting it to electricity: release the energy as heat, use heat to make steam, use steam to drive generator.
The reason just that this is a simple process for which a steam turbine can achieves 90% of the thermodynamic optimum. To my knowledge, the only reason people consider alternatives is to reduce capital costs. You're still capped by thermodynamics though.
That's what I was wondering: if heat -> steam -> turbine is close enough to the theoretical upper limit and you get diminishing returns by other means.
Still, it seems very indirect. Like generating solar power by using a parabolic mirror to heat water instead of photovoltaic panels... but of course I just found an example of doing that too: https://en.wikipedia.org/wiki/Parabolic_trough
Which doesn't make sense for power generation since there will always be some percentage of neutrons produced by any type of fusion reaction that can only be useful for generating steam.
To entirely skip the steam cycle portion is to intentionally make a much less efficient design.
For space-constrained, high-value, applications where economics don't matter that much, such as a submarine, that would make sense, but otherwise...
Helion's fuel mix produces just 6% of its energy as neutron radiation, and if you harvest it you'll lose a third of that. As long as you have enough net energy, collecting that 4% might not make financial sense.
With fuel costs insignificant, your cost per kWh is mainly capital cost. Let's say it's all capital just to keep it simple. I don't know how much the input energy will be but if your choice is between, say, generating net energy of 50MW without a turbine or 54MW with a turbine, then you would skip the turbine if it adds more than 8% to the capital cost. I suspect Helion has done this calculation in detail.
It does seem to hinge on the cost of fuel, I have some doubts about whether they can secure a fuel supply so cheap as to skip out on that extra 4 MW, even after factoring in the cost of a small steam turbine installation.
Deuterium costs several thousand dollars/kilogram. But even though it is just one part in several thousand of the hydrogen in water, there's enough deuterium in your morning shower to provide all your energy needs for a year.[1] Cost of deuterium is definitely insignificant.
Helion's other fuel is helium-3 which they'll make themselves by fusing deuterium. So the helium-3 cost will directly depend on the capital cost of the reactor producing it.
(This may be the same reactor, both generating electricity and breeding He3. Or they may use dedicated He3 breeders, and minimize the D-D reactions in the generators.)
Nitpick: You’re always going to get some parasitic D+D reactions because it’s reaction cross section is appreciably higher than D + 3He below 50 keV, and non-negligible even after the crossover point.
You could always add the liquid metal blanket if you want to eke out the extra 10% (or if you want to generate tritium). But it's not worth the complication in an early prototype.
Helion is planning to use a different fusion reaction, one where the bulk of the energy will be coming out as charged particles, not neutrons.
However, D+T fusion is the only type of fusion that we have been able to sustain for any significant amount of time with reasonable energy inputs. What Helion is planning to do is completely unexplored and requires some major scientific advances.
Well, we do have some other strategies. Hydro and wind just turn the turbine mechanically, they don't heat up water to make steam. And photovoltaics create electricity directly using an effect that won Einstein the Nobel prize and began the age of quantum mechanics, so that's about as advanced as they come.
Just to amplify on your point about lithium: the tritium production function is critical. Every fusion neutron needs to produce more than one tritium atom on average, so that the reactor is sustainable (there are inevitable losses & tritium also decays radioactively) or even making excess tritium (to bootstrap other reactors). This is challenging b/c even in the best case each neutron can produce maybe 2 tritium atoms, so there's not much margin. The lithium needs to comprise most of the material surrounding the plasma, limiting the fraction that's available for other functions (structural supports, heat shielding, cooling, plasma control & heating systems, sensors, etc).
All the neutrons that don't get caught by the blanket will usually be caught by the support structure of the reactor, damaging it and making it radioactive. This puts a cap on the maximum lifetime of this type of fusion reactor until significant parts of the structure need to be dismantled and replaced, and stored as highly radioactive waste. As far as I know, current estimates suggest something like a decade - one of the biggest problems with the economics of fusion power plants.
Stars rely on gravitation to confine the plasma instead of magnetic fields, and they can reach the energy densities needed for other fusion reactions, typically D+D, which produces less neutrons. Also, there are hundreds of thousands of kilometers between the center of the sun and its outermost layers - neutrons produced in the core will have plenty of time to be absorbed by something else on their way out.
They currently run a bad-ass heatsink (which is one of the main challenges of this project, i.e., how to cool it), but eventually you will use that heat to convert it into electricity, yes.
For the German-speaking crowd here, the Alternativlos podcast guys were there twice and had lengtly conversations with the researchers there. Like, between nerds. Really cool, if you understand the language.
Omega Tau also visited and talked with some of the researchers and has a great podcast episode on it (and truth be told, all of their episodes are great).
Same way most electricity is made, you use the energy created to heat up water into high pressure steam, high pressure steam turns a turbine(s) which turn gensets that produce 3 phase AC current.
This one in particular isn't setup to do that, and as far as I know, none are yet. It's a pretty simple engineering problem, and, until we can maintain fusion for months at a time, it's not really something that needs to be built.
There is, however, one fusion concept that shows some promise that doesn't require all that that helion energy is developing (helionenergy.com) they're yet to create net-power, but, their idea has some promise, and avoids the common problems with other forms of fusion power. I don't really see it as the be-all to end-all in the space, but from what I can tell they very well might be the stopgap that is needed between large scale stellerators and fission.
Given that the plasma is several million degrees, it will radiate a lot of energy and heat up the walls even if it does not directly touch them. Just cooling the walls can heat up the cooling fluid enough to later produce steam with. AFAIK the Wendelstein machine is not configured for electricity production though, so the cooling is just cooling atm.
> Given that the plasma is several million degrees, it will radiate a lot of energy
That doesn't entirely follow. 2 particles whizzing past each other at relativistic speeds have extreme temperatures but don't offer much energy. Mass is in this equation.
What if you have 10^20 particles? Each charged particle emits photons with energy/frequency proportional to their speed (Bremsstrahlung). This is mostly from electrons because they are much lighter and so are much hotter/faster. Plasmas are quasineutral though so you'll have those electrons present. There is a long line of research trying to get away from that constraint with little luck so far (but it should continue to be worked on!).
Jumpjng back up the stack: photon radiation is mostly considered a loss since it transfers energy out of confinement and does not impart it on other fuel. You nominally extract your heat via neutrons: same as fission reactors. Some designs (Helion) aim for reactions with charged byproducts. The reaction produces a current that can be coupled by a surrounding coil, much like a transformer but powered by current induced by plasma rather than another copper wire.
I'm not sure if it's the case for this specific reactor, but the common answer to this question is that you need cooling in the surrounding walls and the coolant that runs through the walls transfers the heat out where it can be used to do useful work.
The reverse actually. Reactors need to be shut down when rivers get too warm because they are not able to properly cool the reactor. Nuclear reactors don't produce anywhere near enough energy to heat a large river by a measurable amount.
here there is no power generated as it's not working with deuterium-tritium. most of the heating will heat the plasma and a fraction of this will reach the cooling system. To make a comparison ITER is expected to have 50 MW heating for 400 seconds approx. = 20 GJoule. Using a DT mix will result though in 500 MW Fusion Power
I love these guys, they are just knocking down the engineering challenges in their plan to completely characterize and control a fusion stream. Sometimes they feel like the Tortoise in the race to a working fusion power plant but they are answering questions (managing wall temps and hold fusion in streams[1]) that the Tokamak folks have yet to solve. My bias though is I'm way more on the "D" side of the R&D spectrum and following ITER often feels like pure "R."
Speaking of fusion does anyone know what is going on with SPARC at Commonwealth Fusion Systems? I have been very excited about their system but they are understandably in a deep development and construction cycle after a $2B investment, so all their news page has for the last year are updated business deals and awards. I would love to hear how reactor construction is going.
My best friend works for them in diagnostic sub-systems development. The product is still a long way off delivery with many systems being actively designed and refined. Basically it’s busy but will still be quite a while (3-5+ years at least).
I expect they'd happily give access to anyone who wanted to cover it. Say, if someone wanted to do quarterly update videos, and had the appropriate skills, they'd only have to find the right person to ask, on the academic side. Hint.
Thanks. I do have the appropriate skills but I live in California and I’m busy with my own world changing projects that need videos produced about them! I just want to read blog posts about what’s happening.
i’ve been following this project for 10 years. it’s been successful. but how do projects like these move faster? the wendelstein 7x is never going to generate usable electricity. it’s supposed to be the pre cursor to the producing reactor
Stellarators in particular suffer from very long development cycles. It takes years and years of research to develop the algorithms used to optimize the coil geometries, and then the production of the coils and assembly of the vacuum vessel within the coils is much more challenging than for a tokamak. The coils are hard to produce because they have highly irregular shapes, and tight tolerances. Assembling the vacuum vessel is hard because the coils cover much more of the "toroidal-ish" surface area than in a tokamak.
The is a lot of interesting work going on in stellarator design optimization now, but it will likely be many years before that research is realized in another actual reactor.
For a few billion USD you could build a real power plant of this type. Sounds expensive, but consider how much money nuclear fission did cost initially, and how much money we burn on other stuff, then it's not unthinkable to have somebody rich chip in and make it happen. (Germany just gave $10bn subsidies for a domestic Intel factory.)
That's how research works. The first fission experiments couldn't power half a small country either.
When you listen to the guys from this original article then you'd know that for 10-20bn USD you could likely build a real power plant with this tech within 5 years. It's obviously not without risk, which is why nobody is doing it. But the technical feasibility is there.
They also point out that once the first-of-a-kind installation exists, subsequent models will be way cheaper and way better since you'd have learned a lot and streamlined the process.
But we choose to use public money for fossil subsidies instead, cause jobs, or something.
The US has spent (inflation adjusted) $34 billion on fusion as of the end of 2021. Assuming no real change it would be $35 billion as of the end of 2023.
Fusion has been "a decade away" since before the turn of the millennium.
To say we should just throw $10-20 billion at a power plant and hope something comes out is not a good idea.
We should wait until one of the many multi billion dollar research plants is able to get even 10% of a reasonable to target energy output before even thinking about that.
Otherwise we would likely just sign the death of fusion in the public eye. Could you imagine the backlash if a $40 billion dollar project couldn't even produce power after two decades? (Going off how public works costs and timelines have been going is the reason for higher values)
There should be more funding in this area, but at some point you've got to build it, and that takes a ton of time. Regulations/bureaucracy could be better but at the end of the day you're not going to cut off a ton of time safely.
Once you have a working model iteration gets much much faster, but we've simply been hitting walls for decades.
I'm honestly not sure if you're joking but in case you're not, the "minimum viable size" is hardly the largest issue with what you're proposing and it sounds like you're not getting what the key issues are.
You're talking about taking a technology that's so finicky we've barely gotten it to work after almost 100 years and rocketing it into space? We're no where near good enough at this to get a test that would work after the extreme violence of an escape velocity launch.
Further fusion reactors aren't like fission. "exploding" really isn't a problem . Keeping the reaction going in an efficient manner is.
IF exploding was a problem, space is probably the worst place for it? Putting it way underground would be vastly easier and a hell of a lot safer because you won't have material possible falling back to earth/hitting satellites in orbit.
That was not a real question, it was a subconscious plea for elon musk to take over and make it work.
"Shooting it into space" is a reference to how SpaceX disrupted the rocket industry through a "fail fast" mentality, aggressive goals, and sheer force of will.
Fusion reactors are pretty non-explodey. Really all they can do is spring a leak, then fill with air and extinguish the plasma. Maybe if you quench the magnets hard enough you might get something dramatic like leaking a gram of tritium.
I'm no expert on this, but a minimum viable size exists and it's much larger than what you could just launch into space. I watched a video on this years ago, I don't recall the exact relation to size but if I'm not remembering this completely wrong there's a minimum size you need for a fusion reactor to "ignite", ITER is huge for a reason.
Yeah there is, at least for a tokamak. The smaller the device, the stronger the magnets need to be, otherwise you can get instabilities that will prevent the plasma from being contained. This is why commonwealth fusion is building their own superconducting magnets, for example.
In the (German) podcast alternativlos two nerds interview the Wendelstein 7-X people and ask them specifically what stops them from moving faster and their answer was: money.
You can't jump from idea to production power plant in one step. This is research about the fundamental science involved. What they're doing is incredibly difficult and complex. They have a plasma at millions of degrees mere centimeters from superconductors at near absolute zero. The field geometry and interactions are so complex it brings even current supercomputers to their knees. The device wasn't even possible to simulate until the late 90s using the biggest machines in the world.
What they've already demonstrated is a tremendous accomplishment. But apparently if it doesn't go from idea to an option in door dash in 6 months flat that's not good enough for people here.
It's a research reactor, not a production reactor. Generating useful electricity was never the design goal. The goal was to learn how to build a reactor that could generate useful electricity.
We are researching fusion technology. It would take a reactor many times larger to get more energy out of the facility than you put in. The technology still needs to mature before a reactor that size would be financially responsible.
In the German Alternativlos podcast the Wendelstein team (Prof. Dr. Thomas Klinger, Dr. Adrian von Stechow) recently stated that it is already feasible, they estimate a cost of ~€20B and a 5 year construction time for a commercial fusion power plant if we started now.
For less than 15 billion euro you could buy enough solar to power a country the size of the Netherlands. With 5 billion to spend on batteries you might even make it through night time usage.
Or in other words: Fusion is too expensive at this point to be useful.
Then some countries stepped up the subsidies game and booom, prices fell dramatically since suddenly everybody wanted a piece of the cake. And competition drove this all down.
All you need is for somebody to start. Or we just keep telling ourselves that it's too expensive, shrug, and move on.
Also note how the goal posts changed. Until recently, everybody made fun of fusion by basically saying it's too hard, it's too far in the future. Now it's not too hard anymore, it's just too expensive. What's next? Too loud? Too big? Induces headaches with the esoterically minded?
And it was pointed out 40 years ago that DT fusion will be inherently expensive (specifically, more expensive than fission, which itself has demonstrated it cannot compete.)
Solar has the advantage of scaling down. 1000 people can give 1M US dollar which can produce approximately 1GW of power on a thousand of power plant in a year. You can scale down as low as 400W of energy production and distribute the financial cost to many people.
No. Right now there is no reasonable way to have a 100% renewable and reliable grid in Northern Europe, excepting classic hydro.
I specifically studied the German grid, and it needs about a MONTH of storage to compensate for a once-in-a-century Dunkelflaute (a period with little wind, no sun, and cold temperatures).
You're wrong about that. Back up with green hydrogen is quite plausible. Europe has enormous salt formations in which cavities can be solution mined for gas storage (this is one of the chief ways natural gas is stored). Storing hydrogen, the cost of these caverns per unit of storage capacity in these would be about $1/kWh. The total potential capacity there is in the petawatt hours, far more than would be needed.
A combined cycle power plant costs about $1/W of capacity (and for rare events, simple cycle would be even cheaper), so one could back up the entire grid with these at a small capital cost compared to powering the grid with nuclear. For Europe, these would also be useful for seasonal leveling, allowing solar to provide a larger fraction of Europe's energy demand.
Hydrogen is an example of "Power to X" (PtX), where excess power, when available, is used to make some very storable commodity. This review article talks about how important these are to reaching 100% RE.
"With every iteration in the research and with every technological breakthrough in these areas, 100% RE systems become increasingly viable. Even former critics must admit that adding e-fuels through PtX makes 100% RE possible at costs similar to fossil fuels."
> You're wrong about that. Back up with green hydrogen is quite plausible.
I have not seen any real plan to achieve this. Right now, it's basically a giant asterisk with a footnote saying: "Magic happens here".
One plan I've seen where authors went totally wild and actually tried to compute what's needed, required converting 80% of housing to district heating with molten salt storage, all kinds of energy storage, and 2x price electricity increase.
I've seen estimates that simply building out hydrogen backup will cost on the order of $300B in power line and pipeline upgrades (because hydrogen can't just be piped through natural gas pipes). And it will still require expanding the renewable fleet.
I'm not at all optimistic about that.
FWIW, I think power-to-natural-gas has the biggest chance, because it can re-use the natural gas infrastructure. But it's still going to be too expensive.
Argument-from-ignorance is not an argument. If you haven't seen "any real plan" that just reflects your disinterest in seeing such a plan.
There is nothing preventing this from being applied to Europe. All the technologies are available. It's just a matter of integrating existing capabilities, which is the surest kind of innovation.
No pipeline upgrades are needed for hydrogen for grid storage, since there's no need to move hydrogen away from the storage caverns. It can be created and consumed there. It could be useful to build pipelines, of course, but it isn't necessary. I am NOT suggesting using hydrogen to replace natural gas in distributed applications.
Power-to-natural-gas has the problem of where does the carbon come from. CO2 capture (either from the atmosphere, or from the exhaust of the CC plants) would add to cost, and then the CO2 needs to be stored also. And, the round trip efficiency will be considerably below that of hydrogen. Power-to-liquid fuels would make more sense; it doesn't cost that much more to turn CO2 + H2 into such fuels instead of to methane. Liquid fuels (normally for air or ship transportation, for example) could also serve as a rare event backstop along with hydrogen, for once-in-a-century events, as long as the CC plants can burn both.
> If you haven't seen "any real plan" that just reflects your disinterest in seeing such a plan.
No. I did a full literature search and I read most of the articles in that area.
> There is nothing preventing this from being applied to Europe. All the technologies are available. It's just a matter of integrating existing capabilities, which is the surest kind of innovation.
What is "this"?
> No pipeline upgrades are needed for hydrogen for grid storage, since there's no need to move hydrogen away from the storage caverns. It can be created and consumed there.
The thing is, most of German storage is in the northern part (Rehden, Etzel, Epe, etc) due to geology. That's not where the consumers are, so you need to build a huge amount of power lines.
To give you a perspective, a fairly typical natural gas pipeline can transfer around 1 Bcf of gas per day, which translates to about 12GW of power. This is the same as the largest ultra-high-voltage direct current (UHVDC) line in the world (in Brazil), built at the cost of around $2.5B for 2400 km.
And you'll need many, many such lines to transfer power from the points of generation and consumption to the hydrogen hubs. This is in addition to already expensive hydrogen production and gas turbines.
I don't see this ever becoming cost-competitive with plain old PWRs.
> Right now there is no reasonable way to have a 100% renewable and reliable grid [...] and it needs about a MONTH of storage to compensate for a once-in-a-century Dunkelflaute
If you accept slightly less than 100% renewables, you could use diesel or gas backup for these once-in-a-century events.
That drives up costs because you have to maintain many gigawatts of backup capacity sitting idle most of the year. It's better to have more flexible solutions that provide value the whole year. That's either baseload (fission, fusion) or grid-scale, seasonal storage (unsolved problem).
The backup capacity is cheap (in capital cost) compared to nuclear providing the same output. Like, an order of magnitude cheaper. Combustion turbines are remarkably compact and inexpensive for their power output (this is why they power our aircraft). It's wonderful what happens to machinery when you can reduce the need to transfer heat across fluid-solid boundaries. Rocket engines are an even more extreme example of this.
Rooppur Nuclear Power Plant cost $6 per Watt of installed capacity over the projected 50 years of lifetime. Simple natural gas turbines (not combined cycle) cost around $2 per Watt over 50 years in just capital costs. This doesn't take into account the cost of the fuel, or the magic infrastructure to produce, store, and deliver hydrogen.
I'm taking Rooppur Nuclear Power Plant as the base for comparison because it's an example of what you can do, when you have a "mass produced" design that you can just quickly build.
That's funded by a loan from the Russian government, not from private financial markets, so we can assume the rate is below market. The actual cost when real risk penalties are included (as they must be for an accurate cost) would be higher.
I also doubt anyone is going to be buying Russian nuclear power plants in Europe anytime soon. The strategic risk and associated cost (as seen with importing natural gas from Russia) would be far too high.
> That's funded by a loan from the Russian government, not from private financial markets, so we can assume the rate is below market. The actual cost when real risk penalties are included (as they must be for an accurate cost) would be higher.
Not much higher, though. Russia makes money on these contracts. South Korea has
> I also doubt anyone is going to be buying Russian nuclear power plants in Europe anytime soon. The strategic risk and associated cost (as seen with importing natural gas from Russia) would be far too high.
Of course. I'm not suggesting that Russia should be relied upon for ANYTHING at this point. It should be as isolated economically as possible.
I'm just using this as an example of what you can do with a streamlined construction pipeline for plain old PWRs. No fancy new technology, no breakthroughs, just regular old good project management.
First, backup generation is expensive. Right now Germany needs about 200GW, and this value will go _up_ when Germany switches from natural gas to heat pumps for heating, and expands the EV fleet.
That's a lot. Even cheap gas turbine power plants will cost around $100B to build.
And while the one-month Dunkelflaute is exceptional, the shorter versions lasting a couple of days happen basically every year. As a result, you probably need about 2-3 weeks a year of various levels of backup utilization every year.
Chile is an outlier, the plants are in remote locations in the Atacama desert where you have two compelling reasons to build solar plants: There is a lot of space where nobody lives and the sun is always shining. There are mountains but there are also lots of places which are flat for as far as the eye can see, an example would be the Cerro Dominador plant which probably didn't require any ground preparation.
Chile is cheaper, and in a solar-powered world energy intensive industries will move to such places. If (say) Sweden wants to try to preserve industries by building nuclear power plants, they'll find the expensive power from nukes competing against the dirt cheap power from Chilean (or Namibian, or Australian, or Saudi Arabian) solar.
> in a solar-powered world energy intensive industries will move to such places.
And to windy places. Happening already in Europe, building new industrial plant close to the huge and fast growing offshore North Sea wind power plants
1 euro per kW of capacity isn't that low, residential systems can get below that and it includes installation cost and an inverter which would both scale better for a bigger system.
From listening to the episode I'd say that it isn't much more than a gut feeling derived from their experience building the experimental reactor, coming from the head of W7X I'd give it at least some credibility. Without sufficient political will this isn't feasible at all.
Well, most land is either farmland, forest or desert. Forest is out of the question for any solar installations unless you want to cut down the trees. Desert is not easily accessible for most parts of the world and provides terrible conditions for solar cells that have sharply declining efficiency with heat and don't like dust.
Leaves farmland, if you want to do this kind of thing at any sort of required scale. (Sure you can put solar cells on barn roofs, but the premise was scale magnitudes beyond that.)
Hint: solar on farmland does not interfere with use as farmland. Look up agrivoltaics. Solar on pasture is even easier, and protects livestock from weather extremes.
Farmland is literally converting solar energy into sugars. It's natures solar plants. Sure, you can get a bit of shade for your cattle. But the hundreds of millions of acres used for crops are directly competing for sunlight with your solar cells.
You imagine that is so, but the science says otherwise.
In fact, most plants can only use sunlight for a few hours a day, and must then endure the heat for the rest of the day. A few crops -- wheat, corn -- offer slightly reduced yields when shaded, but many others -- particularly peppers -- yield better with partial shade. Even where yield is reduced, the extra year-round revenue and radically reduced water loss may even the score.
Related recommendation for the german-speaking crowd here:
The Podcast Alternativlos by Felix Von Leitner and Frank Rieger were twice in Greifswald to interview some of the people behind the Wendelstein.
In the first episode (http://alternativlos.org/36 from 2016) they mainly focused on the development and build process and the history. The second one is from this year and they talk about the achievements and the future of Fusion (http://alternativlos.org/51/)
Question for this knowledgeable group of people: Which fusion start-ups look promising? How does inertial confinement look? Any thoughts on the newly funded start-up Blue Laser Fusion? Any thoughts on one of the older players, TAE? How about Commonwealth? Helion? Others?
Also their concept is just so weird, I love it. Worst case it works in space as a nuclear fusion drive :)
My guess is that they're all going to tank without government money, unless by some miracle they have truly found some special low-cost of operation on the first try that beats the current heavily optimized solutions for power generation.
I feel like it's more like saying, we should build an airport here, because it's near a city and in a good ___location, and the land is cheap because no one else wants it for a number of reasons and this is the only thing that practically makes sense to go there.
It's not going to save a meaningful amount of money. But that doesn't mean it's a bad idea.
Sure. But even in your example the site selection won't help you much to build the airport, you still need to do that, and it's still going to be massively expensive.
I'm also in general skeptical about conversions of coal power plants into nuclear even for fission. Typical nuclear plants produce much milder steam temperatures and pressures than coal power plants, so their steam turbines are optimized for different conditions.
I think Helion is most promising for two reasons.
1. Even if the more traditional fusion power plants manage to generate the plasma itself in a device that’s not too expensive (big if) it seems that just the heat exchange mechanism itself would be extremely complicated and expensive. And since they’re thermal power plants you’re limited in where you can put them and how big they must be to be economical.
2. As CO2 emissions come down, I think there will be some focus on thermal power plants contribution to global warming. Helion will still be adding heat to the planet that wasn’t there before, but there will be less heat for a given amount of electric energy. It’s also not going to rely on dumping all that heat in a river.
I don’t know if Helion is feasible. But it feels like it’s the only technology that could be feasible.
I used to think Helion might have a chance, until I looked into the energetics graph. At the temperature they hope to operate, a deuterion would much rather fuse with another deuterion, producing a neutron, than with a helion.
And Helium3 is not available in useful quantity. No, not even on the moon.
Literally none of the projects running will ever lead to so much as a single erg of commercial power. They might spin off interesting plasma materials-processing tech.
It would be much more expensive to build a fusion power plant than a similar-capacity fission plant, but fission is already uncompetitive, and falls farther behind by the day. Start thinking hard about uses for stable contained plasma that doesn't fuse. Advantage is, it doesn't need to be especially hot. What they have already is more than hot enough for any plausible use.
I greet W7-X with a huge yawn. A reactor based on stellarators will still be very large and have very low volumetric power density. The beta is not good, so these would only work with DT, and suffer from the generic problems of all DT schemes.
Right. It is fun to keep plasma hot for a long time, but literally none of the fusion projects being worked on can ever lead to production of so much as one solitary erg of commercial power. Cost will necessarily be much more than for fission. Fission is far from competitive today, and falling farther behind daily.
There is a bare possibility of eventual usefulness for spacecraft propulsion. And, some spinoff might come out of a newfound ability to handle lots of hot plasma. Maybe for sewage treatment?
> In individual areas, temperatures of up to 600 degrees Celsius are reached (red areas). The divertor tiles can withstand temperatures of up to 1200 degrees Celsius.
They heated the plasma with a power of 2.7 MW for 480 s which in total deposited 2.7 MW × 480 s = 1296 MJ, i.e. 1.3 GJ, of energy into the plasma heating it up. That is the energy of 310 kg of TNT (4.184 MJ/kg) or burning 38 l of gasoline (34.2 MJ/l). Keep in mind that this energy was deposited into the plasma which has a mass of only about 10 mg.
If I understand things correctly, the problem with magnetic confinement (e.g. Tokomaks, Stellarators) is that once you have heated a plasma such that it is "fusing," how do you get the power out with out cooling the very plasma you've just spent a lot of energy heating up?
Helion, a fusion startup, claims to have solved this problem via capturing an induced current from colliding two hot plasmas together. I'd be curious if there is any way the Wendelstein can produce electricity.
Most fusion power systems assume they are doing that as neutrons. D-T fusion conveniently has the proportion of energy that gets lost from the plasma as KE of neutrons be pretty close to the amount of energy that a conveniently sized fusion reactor can afford to remove from the plasma.
Then you trap the neutrons with, for example, a lithium blanket, use them to breed more tritium, and produce energy with a turbine from the heating of the blanket.
It's a been a long time, but in a talk at Google, I think Bussard said that power output scales with the 5th power of the radius of the device. There's really no point making a small one.
This is an incredible achievement but there are strong reasons to suspect that stellarators are not and will never be plausible candidates for energy generation. For some more experimental or perhaps military tasks, it's viable.
Why do you think this? Also, suggesting only the military should use it suggests it is stellarators are well suited for energy weapons versus powering homes/industry, which is also curious, as energy weapons need huge impulse amounts of power, not power that is constantly available and ramps slowly by comparison.
Can someone tell me why this won't produce commercial level fusion for 30 years so I can shut down my eternally optimistic "physics kid" portion of my brain for a while?
Interesting fact: Nuclear fusion, even if we'll make it work, won't stop global warming, because the heat it creates heats up the earth enough to bring us outside the Paris agreement: https://twitter.com/rahmstorf/status/1605967891928596481
One immediately apparent flaw of this argument is the assumption of energy use growing by 10x over a century. But in developed countries, energy use per capita has been roughly stable for decades. The 2 main drivers of energy growth will weaken over time (population growth + countries becoming developed).
Also, if energy use does increase by 10x, the solution is simple, build giant refrigerators powered by fusion energy to cool the atmosphere. (joke)
Assuming I trust the math, that’s plotting exponential growth in energy usage out for 80 years, and assuming a fully nuclear grid. Neither of those is likely. Lastly the Paris accord is a pipe dream that will never happen. A target to aim for, and miss. Nothing more.
According to the article, we currently emit 2.1w/sqm in greenhouse gasses.
If we had 10x energy and it was all fusion - it would be 0.4w/sqm.
This sounds like a massive improvement.
Additionally, I'm highly skeptical we'll be using 10x the energy in 100 years - when there's likely to be significantly less people, and everything is getting more efficient.
Sounds like the same argument that China is going to continue growing 10% per year for the next hundred years, because it did for the last 30 years. No. China's workforce is going to decline massively. It will be so much harder for them to grow at the same rate, it would take a real miracle to keep growing at that rate.
this is a basic consequence of thermodynamics and true for all power generation. the only thing to be done to minimize waste heat is to to increase temperature of the hot side of the heat engine, with ie advanced fission reactors.
and fusion never had any advantage over fission anyways, other than that people aren't scared of it yet.
The threat of the earth heating up by 0.3 degrees due to energy production is irrelevant or at least absolutely worth it as a tradeoff for working fusion.
The dangers of climate change is not that the earth heats up by some small amount, the earth can easily cope with that. It is that continued greenhouse gas emissons are causing a ever increasing heatup due to trapped solar energy.
(It is also extremely strage that he argues for geothermal in his comments. Does he not realize what that is? Literally heating up the surface of the earth with energy from below.)
Didn’t downvote you, but having unlimited energy source is worth it and can allow us to remove heat from earth. Human-caused CO2 alone contributes 2.1W/sqm while all current human energy production is 0.04W/sqm . Removing extra CO2 alone would offset 50x energy production growth. Then you can do things like placing reflective satellites between earth and sun.
Because the argument involves unlimited future growth in energy use. Compared to the current energy use, fusion (assuming it could be made to work practically) would indeed solve global warming.
With gas cut off from pipeline terminated in Greifswald, how will they power now this bottomless energy pit? They still have some money, but a finite amount.
We know the equations for flight. Why didn't they just build a 787 in the 40s already?
Oh, is it because the technology didn't exist and first had to be developed, in incremental refinements? Initial airplanes didn't even fly and half the people trying them died? Oh...
The basics of human history show that these kinds of predictions are sometimes right, sometimes wrong, and being too sure about them is usually the wrong way to do about things. We'd all still live in caves with that attitude.
The history of engineering shows that most approaches to solving problems fail. That's because, as in ecology, there's typically only one approach (in ecology one species) per niche (in the market or in the ecosystem). The winner drives the losers to extinction.
For fusion, we have to ask why it's going to be an exception. The prior is that it won't be. If there's evidence it will be blocked, that's two (or more) strikes against it. Something very unusual is needed to come back from that far behind.
The continuing success of renewables, and their continuing progression down their experience curves, is bad news for fusion.
Hm, you are obviously not following what's going on in the fusion space currently. They are continuously churning forward towards a viable design.
What "renewables" (with which I suppose you mean solar, hydro and wind) have going against them is their environmental impact. Sure it's not as bad as fossil fuels, far from it, don't get me wrong. But the area and materials needed for solar, the animals getting disturbed by wind turbines (birds killed, wales confused, ...) and the ecosystems that get flodded by dams are not nothing. Especially in the light that the energy demands of 8bn+ people are continuing to grow, and those 8bn will soon be 9bn and 10bn.
The promise of fusion tech is that you get much more bang for the buck, and with "buck" I mean resource use. That's not going to happen soon though, so until they we'll be stuck with solar+wind+hydro, but those are not really sustainable solutions in the long run (i.e. 100s/1000s of years ahead).
Of course, if you see fusion just as some crazy idea and know nothing else about it, then I can see how your "prior" makes sense. But once you know the details, it's quite different, since fusion is also continuing to progress down its experience curve. Tokamaks and stellerators in particular.
Let's deal with the environmental impact argument first. We can estimate the cost of damage to the environment from what activities societies allow, and the value those activities produce. The largest use is agriculture. Societies allow elimination of natural ecosystems and their replacement by crop monocultures. What value is obtained? Typically, the value of crops delivered from a tract of land is two orders of magnitude less than the value that could be obtained by putting PV there.
So, if we stipulate your environmental argument rules out renewables, it also (to a much stronger degree) rules out agriculture. This is obviously absurd, so your argument cannot succeed.
As for recent "goings on"... I do follow them rather closely. You are likely misled by a common cognitive failing. That is: if we have a set of steps needed to reach some goal, then if one of these N steps is achieved, it's natural to think that we're 1/Nth of the way there. But this is only true if the steps are equally difficult. This cognitive blind spot is exploited in those collectable coupon games you sometimes see at grocery stores or fast food outlets. The # of winners is controlled by the number of a particular rare coupon; all the others are just noise.
For fusion, the immediate steps have been plasma confinement, stronger magnets, and so on. But none of these matter if there's a later showstopper. And for DT fusion, there is. That showstopper is the inability of DT fusion reactors to achieve adequately high volumetric power density. None of the recent DT reactors are promising in that respect, and there's good reason to think this obstacle is generic. Lawrence Lidsky (and Pfirsch and Schmitter in Germany) in the 1980s pointed this out. The implication of poor volumetric power density is that DT fusion will be more expensive than fission -- and fission itself cannot compete with renewables.
(I view current work on DT fusion as "making good progress toward a dead end.")
(If Pfirsch's name is familiar, it's because he, with Schlüter, discovered Pfirsch-Schlüter currents, which are important in stellarators.)
The only effort I see that has any chance is Helion's, which does not use DT, because they can evade this showstopper (by not producing their output as heat, allowing them to potentially save on the cost of the non-nuclear part of the plant.)
>Why even bother with these machines that can never be built economically?
If you can not build a research reactor which functions well, then "building a machine that can burn plasma and breed tritium at appreciable rates." is more than impossible.
Maintaining a controlled fusion reaction for eight freaking minutes seems like a pretty worthwhile accomplishment in and of itself. The only other place this is known to occur is in the center of a star. Doing it here on Earth is pretty mind-blowing IMHO.
I used to be really interested in this, but forgot it existed over the years. Glad to see it works!
https://en.wikipedia.org/wiki/Stellarator