> A traditional tokamak has a toroidal confinement area that gives it an overall shape similar to a donut, complete with a large hole in the middle. The spherical tokamak reduces the size of the hole as much as possible, resulting in a plasma shape that is almost spherical, often compared with a cored apple.
Is it possible I’ve only seen pictures of a spherical tokamak? All the renditions or pictures I’ve seen had a core pillar that was 1/8 or 1/10 the height of the chamber. Which has me wondering if there’s another tokamak out there with a much larger inner diameter.
The stellerator for instance is more traditionally toroidal.
That is because it was discovered that the D-shape tokamak was more stable than toroidal ones. All of the major ones since the 80s have used the D-shape.
The D-shaped "advanced tokamaks" and spherical tokamaks are similar shape but the spherical has smaller center hole. That is the most efficient but it is harder to build. The center hole is the most confined space in giant ITER.
I like that Grounding looks like a first class passenger in there but definitely it's one big electric code violation in a box.
Interesting to compare with the Nvidia latest keynote [1] where they opened up a 3D rendering of some ASML internals and how slick those systems are put together given the fantastic level of complexity involved also.
The ASML stuff is production equipment where downtime costs millions in chip production.
This is a dev system where a few second run after a few tries is excellent and they constantly tweak stuff.
I'm not sure what ASML's dev setup looks like.
if you want to understand fusion and what the heck tokamak is - I can highly recommend Lex Friedman conversation with Dennis Whyte [1]. It’s a great source of info not only about the fusion energy - Dennis describes so much more there! Amazing guy.
I feel like commercial fusion is such a lost cause at this point. For electrical generation, it’s going to have high fixed costs which will likely make it uneconomical for the same reasons fission struggles. The cost of fuel already very low for traditional fission and renewables.
I don’t see what advantages it would have over wind, solar, fission, etc.
That’s because commercial fusion doesn’t exist yet. All new technology isn’t commercially viable when only viewed from the perspective of early prototypes.
This is as equally true for consumer hardware as it is for massive scientific endeavours.
Many (most?) novel technologies never become commercially viable. I think what @charlieao is saying is that they suspect fusion will be one of the failures. That's a different claim from the one you're countering.
Of course predicting the future of new tech is always a kind of bet. I don't know enough about the details to make a judgement either way. Fairly obviously it just isn't going to be ready in time for the thing it's most often touted for - ie. the decarbonisation exigency. That's just the usual research funding opportunism.
But assuming we do make it over the next few decades (very unlikely in my view), cheap limitless energy is inevitably going to be transformational.
Usually those novel technologies that don’t become viable is because it’s either a technology we don’t want (like a sentient toaster) or there is another technology that can solve the same problem but cheaper.
In the case of fusion, it’s something we do want (fission and burning fuels aren’t sustainable indefinitely) and as our energy needs increase year on year, the cost for energy will also increase if we don’t find a way to supply that demand. This makes a mass producer of clean energy more viable.
Renewable energy is great but it’s not exactly cheap to scale either. In fact the only reason it has gotten so cheap is precisely because the need for it has lead to plenty of advancements that have also driven down the cost.
So fusion in this form of reactor might prove to be a dead end but at some point humanity will end up with some form of fusion. Even if it’s not in our life time. And assuming we haven’t had an extinction event first…
> Usually those novel technologies that don’t become viable is because it’s either a technology we don’t want (like a sentient toaster) or there is another technology that can solve the same problem but cheaper.
I'm sure there must be academic studies looking into reasons for tech not getting to production. I don't know them, so I can only hazard a guess that this isn't quite right. I think many technologies just don't prove feasible outside the lab. One cause of infeasability is of course price (no route found to a true industrial scale learning curve). But sometimes it's safety or inherent limitations that weren't understood before trying to scale, or any number of other causes not predictable from basic science. You can't know 'till you know, and for fusion, we don't know yet. It's a bet (which to be fair is the same with nearly any new tech). A good one, perhaps.
it's not that simple. context (timing, culture/politics, market conditions, regulations) is very important.
there's a lot of interest nowadays in "progress studies" which tries to tackle this (or a very similar set of) questions
yes, sure commercial fusion is definitely one of the toughest nuts, but fission based small modular plants are viable, yet it's barely being done. and there are many products and technologies that are economical yet not widespread. (let's say public transport.)
Agree on the complexities. 'Viability' has to be conceived more widely than just a technical sense. Fission a good example, as you suggest. A dead duck in most of the world. As is wide-scale decarbonisation! We know that (a) it is technically feasible, and (b) it is politically infeasible. So decarbonisation is actually not viable, despite appearing technically possible from a narrow perspective.
This also makes commercail fusion very unlikely to happen in this civilisational iteration. It requires large advanced economies with highly functional and specialised supply chains. These conditions won't exist once climate refugees sweep the world in their billions, and wars break out everywhere. If, indeed, given Russia/Ukraine, and the joint determination of China & the US to go to war, we even last that long.
But then one still has to be skeptical about tokamaks. Plasma instabilities for unknown reasons alone is enough for that.
And then there is the flux of energetic neutrons with order of magnitude higher energies than that of a fission reactor. Spherical tokamaks are especially bad in that respect as the central thin handle cannot be protected sufficiently.
At least with stellarator designs there are no instabilities and the neutron flux can be managed with a sufficiently thick blanket, so those are safer bets than tokamaks.
But they also need to be larger for a given power output, so frontrunners are trying tokamaks first because if they can get those to work at reactor relevant levels then they will have a cheaper / more competitive product. If tokamaks don't work out and the price of electricity is sufficiently high, then stellarators will likely make the most sense.
They're not even sure tokamaks can be run steady-state and not pulsed. That's hand-waved as an "engineering problem" like everything else tokamak-related. Stellarators at least look like they could, since you don't need to induce a current to generate the magnetic field via induction.
Pulsed reactors is not a showstopper. If you pulse heat into a coolant blanket there is plenty of thermal energy buffer. Hitting 90+% duty cycle is an engineering challenge with no physics hurdles.
Just running the steam turbine part of a power plant struggles to compete with renewables. I don't see how attaching a complicated high-tech gizmo to the steam engine can help the economics unless it also prints money as a side effect to producing heat.
>Just running the steam turbine part of a power plant struggles to compete with renewables.
Disingenuous. Your bicycle costs less than a car, but can you use it to move 500 pounds at 60 mph in the rain? Talk about renewables + seasonable storage.
Fusion is a high-capex form of energy. If your only niche in the market is the two windstill, dark weeks in the year you'll have a hard time recouping your investments.
The problem with natural gas turbine is the energy cost of fuel. One looses a lot extracting it from Earth. And then there are unknown, and, as such, potentially catastrophic, consequences of pumping more and more CO2 into the atmosphere.
Fusion reactors has none of that. But the cost of gizmos as seen currently is a way too high indeed.
Fission is theoretically amazing in many cases, but is hampered by branding problems. If fusion can be marketed as "clean", it gets a comparative advantage.
As I understand it, fusion's radioactive contamination and waste is a lot shorter-lived than fission reactors', and there's no chance of a runaway meltdown reaction (not that there's any chance of that with a modern fission reactor either... and fission reactors generate a tiny bit of waste in exchange for free unlimited energy...)
That's one of the points of the spherical tokamaks - they're MUCH smaller than the standard tokamaks (I can't remember the numbers but it's vastly smaller). That brings the cost way way down.
The modern HTS magnets are also more powerful, even in the Tokamak systems (like CFSs) which again get the size way down - I think it's something like cubed size reduction for doubling magnet strength (?).
You'd be right for something of the silly size of ITER.
It doesnt have to be cheaper, the bar is set somewhere around 'lets not fuck up this planet more than we need' level, price is secondary.
A lot of renewables require massive amount of energy, raw materials and pollution to get done, and then last decade or two at best. I have yet to see a bigger wind turbines fields who are all fully working at given time, usually mote than half doesnt move even in stronger winds due to ie maintenance
The cost of materials may be low for fission but convincing people to have a nuclear powerplant near there homes and engineering it to be safe are not.
Fusion on the other hand has no chance of a run away meltdown and the materials that are used are much easier to dispose of.
We can make fission reactors now that have no chance of run away meltdown. Good luck convincing the public to live next to one anyway. I don't expect Average Joe is going to be willing to differentiate between fission and fusion in that regard.
Depends on the reaction. Most efforts are around deuterium-tritium fusion (that's the pair that's easiest to make fuse), which emits most of its energy as high-energy neutrons. So for these, yes, like the peer comment said, the neutrons are hitting something that heats up (that's the "blanket," which might be made of molten salt or metal), and then you can get the heat out of that with a heat exchanger. The company described in this article is aiming for D-T fusion.
There are other reactions one can pursue that produce charged particles instead of neutrons. With these reactions, there are alternative energy conversion pathways that turn the kinetic energy of these charged particles directly into voltage, and you can skip the turbine. Of the fusion startups, Helion is probably the most prominent pursuing this kind of approach, with a D-He3 reaction. See https://en.wikipedia.org/wiki/Direct_energy_conversion for more info on this general approach.
That has nothing to do with steam turbines. 1 Ton of coal produces 1979 kWh of electricity, which at the current $179/ton is $0.094/kWh just for the fuel, then you have capital costs, labor, profit margins, and that still only gets you to wholesale.
For comparison, fusion fuel is $0.00014/kWh, equivalent to coal at $0.27/ton. There's no where on Earth where coal is that cheap. You couldn't move it from the mine to the powerplant for that cheap even if they were next door to one another. Even if all the costs got subsidized away, you couldn't count on the fuel remaining cheap for the lifetime of a powerplant.
And of course there's that pesky CO2 issue that makes investment in coal unattractive even if it could be economically competitive.
There are places where there is abundant brown coal, which cannot be economically transported long distances and is therefore basically loaded on a conveyer belt directly into power stations.
New-build brown coal isn’t economically viable either.
Natural gas power plants are gas turbines, not steam turbines.
Some have a secondary steam turbine as an exhaust energy recovery system but as I understand it that’s becoming uncommon on new builds because new gas turbines are expected to be used less now that there’s a lot of renewable energy on a lot of grids.
If I remember correctly, most natural gas is used in combined cycle turbines. Kind of like hooking up a turbine from an aircraft directly to a generator. Then the excess heat from that turbine is used to generate steam for another turbine.
So even if the fusion reactor reduces all of those costs to zero, the capital costs would only be 20% cheaper than new-build coal.
As I mentioned earlier, coal isn't economically competitive with renewables pretty much anywhere, even places with very cheap coal resources that can't be traded on the global market. There's heaps of lignite in Texas, for instance, but nobody's building new plants to burn it.
As such, it's hard to see how fusion reactors turning steam turbines are going to be economically competitive either.
The simplest answer that's part of the plan for most DT reactors is "as neutrons".
The fraction of the energy that is retained as velocity of particles with a charge vs that is lost as the velocity of neutrons turns out to be conveniently just about where you'd want it to be. So, to capture the energy you need to surround your reactor with something that effectively converts fast neutrons to heat, such as a blanket of molten lithium, which you then use as a heat source. Lithium is proposed because it would also breed the necessary tritium.
Assuming you need enough lithium to replace the tritium, and 100% energy efficiency, one atom of lithium makes one atom of tritium which makes 17.6 MeV of energy [0].
That's an energy output of 7 x 10^7 Wh per gram [1], or alternatively, a 70MW reactor needs one gram per hour, or alternatively, a very large 3.5GW plant needs 500 grams per hour, roughly the equivalent of the water drunk by a single thirsty person.
You need slightly over 1 lithium atom per fusion event. So for ~40% heat to electricity efficiency, you'd need about 7 grams of lithium for 150 MWh of electricity. A 1 GWe plant would consume about 403 kg of lithium per year, about as much as in 5 Tesla Model 3 Battery packs.
More than likely a plant will have a lifetime supply of lithium and other blanket materials on site before it is turned on. The fuel costs really do drop to effectively zero on MCF devices.
Cost of cryogens, cost of steel, cost of working nitronic, cost of Beryllium, and cost of superconductors are the dominant sources to track.
Indeed reactors are not free. The irradiated materials situation isn't as bad as fission though since they are low level activation materials that become safe within a few hundred years, as opposed to trans-uranics which are radioactive for thousands of years and will always be toxic.
Since it is a complicated engineering problem there is space to turn engineering effort into reduced cost.
That's kinda on the 'we'll figure it out later' track.
Materials are being investigated that can withstand the heat and neutron bombardment to attach a steam turbine the old fashioned way.
But we ain't there yet, not without having to constantly change the 'walls'.
> Most of the energy produced inside a fusion reactor is emitted in the form of neutrons, which heat a material surrounding the fusing plasma, called a blanket. In a power-producing plant, that heated blanket would in turn be used to drive a generating turbine
One video that shows two guys inside the tokamak, installing a diverter, which gives you an idea of the internal dimensions of the toroid - 10 minutes into https://www.youtube.com/watch?v=YkkeCjgrG-0
There's already something called a "spheromak" (I'm not sure what the difference is except they've been around for a while and apparently they're not good prospects for energy production — they have other scientific uses)
Tokamaks have been asymptotically approaching viability for a very, very long time.
At some point they just have to give up on the design, because it just isn't working. But I suppose as long as there's funding there'll be people trying to get it to work.