Nuclear fusion: it's time for a reality check
theguardian.com
theguardian.com
Disclaimer-- I did my masters on fusion reactor simulations.
They say the fuel is cheap. It's not (tritium and He-3 are hard to come by), but even if it were, keep in mind that a fleet of uranium-fueled fission reactors like the one in France spends about 5-10% of its costs on fuel.
Certainly nuclear tech was subsidized originally by the Manhattan project, but very soon after, bomb making and nuclear power went their separate ways.
You could much more strongly argue that the nuclear reactor development made to power submarines helped develop the commercial nuclear power industry. This checks out much more cleanly, as Shippingport was led by Rickover himself.
Simply too expensive for civilian use and they never had a dual use that benefited generating material for nuclear bombs. Commercial reactors today are based on minor variations of nuclear fission that was also strategically relevant for getting plutonium. Different reactors but same mineral supply chains and a lot of overlap in R&D. As soon as the cold war ran to an end, funding for all that fell off a cliff. Including building new nuclear power plants. Coal and gas plants were vastly cheaper to build. And they still are. And back then, nobody cared about pollution.
Only now that people are interested in small reactors are these decades old fission reactors getting more attention.
The submarine reactor uses highly enriched uranium and, I believe, burns like a cigarette, which makes it easy to start and stop the reactor without worrying about [2]. It doesn't produce a lot of Plutonium because it doesn't have much U238 in it. I believe the fuel elements are horizontally oriented. It's built tough because it might get knocked around in a war.
The commercial reactor has vertically oriented fuel elements and tries to burn the fuel uniformly along the axis because this minimizes cost, because minimizing cost is a matter of maximizing power for a certain sized reactor core, and maximizing power is a matter of maximizing heat transfer out of the interior of the fuel rod and to the coolant, so you want to be producing and removing heat from the whole volume of the core. Temporal uniformity helps with spatial uniformity so the reactor is not so good at starting and stopping, but it can follow loads [3]
Notably the west (where natural gas is available) quit building coal burning power plants at the same time it quit building nuclear power plants, partially because of pollution [4], partially because a 100% steam turbine powerset as used in coal and nuclear plants can't compete with a gas-fired combined cycle powerplant which combines a very-low capital cost gas turbine with a steam turbine that is about half the size (and cost)
The small LWR is a pipe dream because LWR economics improve as the reactor gets bigger. One reason why we can't build reactors on an N-th of a kind basis is that every time we build one we decide the economics weren't that good and they might get better if we scale the design up. Had NuScale ever built a plant they probably would have tried scaling it up too.
It is claimed small reactors could be easier to build but until somebody actually builds one, this claim is hard to take seriously.
One path that makes sense is to pick an optimized design for a large reactor and "get good" at building it, it could be an advanced Western design like the AP1000 or something like the Chinese Hualong One which is an improved version of the reactors that France built in the glory days. The Russians are really good at building the VVER, their version of the PWR.
The other path is to build a reactor that works at much higher temperatures such as the fast breeder reactor, high temperature gas cooled reactor, or molten salt reactor and couple it to a gas turbine powerset or use the heat to produce hydrogen directly. [5] This is still decades away.
[1] Pressurized Water Reactor, there is also a Boiling Water Reactor in which the water boils in the core but it's not a vast difference: these are both called LWRs (Light Water Reactor)
[2] https://en.wikipedia.org/wiki/Xenon-135
[3] https://www.oecd-nea.org/upload/docs/application/pdf/2021-12...
[4] https://en.wikipedia.org/wiki/Acid_rain (SO2 from coal burning plants, ironically, masked the warming effects of CO2 until we cleaned up coal plants)
Conversely the graphite moderated Hanford reactors that made plutonium never led to commercial reactors. Closest you got was Hanford N but that was kind of a one-off.
If it matters, I'm a professional nuclear engineer with a lifelong interest and expertise in reactor history.
Graphite reactors have a history for power production: see [1] and [2]. They might even have a future [3] but the German experience makes me think the prismatic type core makes more sense [4]
[1] https://en.wikipedia.org/wiki/Advanced_gas-cooled_reactor
[2] https://en.wikipedia.org/wiki/RBMK
[3] https://www.world-nuclear-news.org/Articles/Chinese-HTR-PM-D...
Your response amounts to "the government was mistakenly subsidizing civilian nuclear power". Which might be true, but doesn't refute the fact that they were.
The government spent money researching it early on, such as for the space program. Then the government has outright subsidized solar installations on roofs, etc. Part of the solar miracle of the last 15 years have been huge demand-side subsidies by China because they want to corner the market.
That said, recognize that there's an important difference here. Virtually all solar subsidies are on the demand side of the equation. They pay homeowners to buy solar, and act to increase the size of the market. They don't make solar cheaper to produce (except via second order effects due to the resulting scale).
Nuclear wasn't like that in the 60's and 70's. The government was handing the checks to the reactor builders, not the consumers.
Solar doesn't have that property. Solar is in most areas reasonably competitive already, and the subsidies work towards increasing the size of its market.
The hope, then, was to quickly obtain an industrial breeder reactor in order to considerably reduce (or even suppress any) dependency towards uranium providers (there is no known similar trick for oil/methane).
It failed, and it may be a major cause of the current nuclear demise.
"identified uranium reserves recoverable at US$130/kg ((...)) At the rate of consumption in 2017, these reserves are sufficient for slightly over 130 years of supply" Source: https://en.wikipedia.org/wiki/Uranium_mining#Peak_uranium
In other words the already precarious and degrading case of nuclear as a sound investment further degrades with each and every new reactor, up to the catastrophic "no more uranium for a reactor not yet amortized".
Nuclear now produces, worldwide, about 10% of electricity which is less than 25% of final energy: nuclear produces about 2.5% if final energy.
Build reactors to have it reach about 7% and there is not enough guaranteed uranium for all reactors to produce 100% of their theoretical output for about 40 years => from an investor's viewpoint it is moot.
Exploration is a bet: the uranium bubble (2007) triggered an impressive exploration raise, which had very insufficient results (reserves: + ~15%).
To reduce greenhouse gas emissions we have to electrify numerous applications and therefore will need at least 2x more electricity: those ~7% (of current electricity demand) will then be ~3%.
Tripling the nuclear output this way means maintaining the (already old) fleet of reactors and building approx 600 more. Before 2050?!
In the event of a nuclear 'renaissance', those limited uranium deposits will be subjugated by superpowers. In the event of a serious accident all hope to amortize may be instantaneously lost.
No industrial breeder => game over.
You seem to think nuclear failed because we didn't go with breeder reactors. But the reality is the opposite: we didn't get breeder reactors because nuclear failed. The justification for breeder development was that uranium was going to run out real soon (with everyone building cheap-as-dirt nuclear power plants), and breeders would then have a big market as uranium prices rose.
But nuclear was too expensive, so the build out was aborted. Uranium has remained at a price level at which breeder reactors (or even reprocessing to make MOX fuel for burner reactors) cannot be justified. In the time scale of operation of any existing reactor, or any reactor we could build today, uranium is projected to remain cheap enough for this still to hold.
That even France has mothballed its breeder program, saying publicly there would be no market for such reactors before the second half of this century, should be evidence breeders weren't some sort of panacea for the problems of nuclear. They were a solution to a hypothetical problem that failed to arise.
And other disasters like Three Miles Island very insignificant.
Do you have a source for that? I'd be very interested in a study that quantifies the damage of coal plants in terms of famous nuclear disasters.
If it's not the important issue justifying nuclear over coal, why bring it up?
Nuclear was the only viable green option back then. It failed because of mass psychosis and because it was more expensive than coal and gas.
Of course it’s too late now. But had nuclear plant construction continued to grow at the same pace as in the 60s and 70s CO2 emissions from power generation would have be been solved by the 2000s.
> adding pollution control to coal plants
Even if it was 100% effective that would do nothing to slow down climate change.
But the mass psychosis after Three Mile Island and then Chernobyl made nuclear expansion very expensive politically and financially (of course not saying that all of the additional safety regulations which inflated costs were not necessary). It would be rather silly to claim that it didn’t tip the scales at all.
You're spinning a comforting story to allow you to evade unpleasant realities. I would have thought "everyone is crazy but me" would have set off red flags for you, but apparently not.
I repeated that cost (of course only if externalities, that hardly anyone was fully aware of back in the 80s, are excluded) was the primary reason why nuclear failed. Do you really think that public opinion had no impact and did not accelerate it, though?
> "everyone is crazy but me"
Is that how you feel? Must be interesting..
We see plans for nuclear go up and down based on economic conditions. Public opinion doesn't seem to have any effect on that. The nuclear renaissance was going great when natural gas seemed like it was running out. It collapsed when fracking made gas cheap again. Today there's concern for AI power demand and nuclear seems to be getting traction. Public opinion isn't driving any of that.
There was a period when being a pro nuclear liberal/central-left politician was politically suicidal.
> connected the dots and shown that had any tangible effect.
It’s not my job to put in significant amount of effort to try and change your opinion (which is worth exactly as much as mine is).
> The nuclear renaissance was going great
That’s an exaggeration. Even back then nuclear was hardly seen as cost effective without a lot of squinting. The number of new reactors was hardly significant.
> Public opinion isn't driving any of that
You mean currently? Perhaps. Don’t see how is that an argument when talking about the 80s.
> Public opinion doesn't seem to have any effect on that
That obviously not true. Unless you ignore how, when and by whom the decisions to close the remaining nuclear reactors in e.g. Germany were made. Especially considering that then (before cheap renewables were available) we already knew that nuclear was cost competitive if negative externalities are accounted for (which was only partially the case back in the 80s).
You cannot show any good evidence for it, though. You can merely assert correlation (if a bad one, given that nuclear was in trouble long before TMI or Chernobyl).
It's clear what's going on here: the actual situation, that nuclear has failed economically, is completely unacceptable to you, so you are confabulating a fantasy that something else caused nuclear's troubles. It's wishful denial of reality to preserve the illusion that your energy waifu has any chance.
Surely you must be trolling? Or you just can’t read? How many times did I repeat that?
> It's wishful denial of reality
I mean you are literally looking and what I wrote and seeing words that are not there.
e.g. NIH estimates 460k excess deaths (just in the US alone) between 1999 and 2020 were caused by emission from coal plants. And that’s the period when coal became relatively “clean”.
So it must have been much worse back in the 70s and 80s.
The highest estimate for Chernobyl is 60,000. But 16,000 is probably more realistic.
Nope: https://en.wikipedia.org/wiki/Chernobyl:_Consequences_of_the...
It may. I don’t think that’s the consensus view though and again something being well-known doesn’t make it the consensus.
I’m not saying that WHO’s and IAEA‘s figures are perfect and flawless but they aren’t widely challenged either.
Of course “environmentalists” and most others back in those days preferred cheaper coal and gas.
Transport fuel and combustion-based heavy industry are the hard nuts to crack, and reactors don't help there.
Lab breeder reactors emerged during the 1950's and were satisfying in the 1960's.
Big projects aiming at obtaining an industrial reactor were aplenty, worldwide, during the 1970's through the 1990's.
They all failed, and remaining efforts are quite sparse. https://en.wikipedia.org/wiki/Breeder_reactor#Notable_reacto...
Obtaining something adequate (raw size, reliability, safety, ease of use...) isn't always easy.
The quest of an industrial nuclear fusion reactor may tank this way.
Instead, I think need to find some way of generating a lot of power over extended periods of time if we want to crack space travel beyond lobbing items into Earth orbit. Particularly for human space travel.
We need significantly more energy density than what chemical motors can provide, and nuclear is the easiest and most "within-reach" that we know of. Things like anti-matter and even more exotic forms of propulsion are pure sci-fi for the time being (in terms of practicality). I suppose fission could work in a very primitive way of basically detonating small nukes behind a pusher plate, but fusion feels so much more elegant (at least to a layman). Mostly in the sense of not having to use effectively the same nukes we stick in weapons. Something tells me that this will never fly (excuse the pun), for obvious reasons.
(and after it leaves the earth's atmosphere, radioactivity ceases to be a concern)
(EDIT: thrusters are rocket engines, just very small ones. All those missions used small chemical rockets for maneuvering, using storable propellants like hydrazine.)
A fusion power generator coupled with ion drives could potentially generate a lot of thrust for little mass since the fusion fuel also has very little mass.
Can we shrink a fusion reactor small enough to make the rocket as a whole feasible? I don't know. My hope is that for space travel, we can make certain tradeoffs that are not feasible on Earth - e.g. expelling radioactive waste products into the environment. A rocket engine doesn't need to be all that efficient at generating electricity either, it mostly needs to be efficient for propulsion, which might make it's design much simpler.
https://www.nasa.gov/general/pulsed-fission-fusion-puff-prop...
Sure, if the fraction is 1/10000. Very high exhaust velocity is not optimal since it increases the power needed per unit of thrust.
The high energy density of the fuel is irrelevant if the reactor is very massive and if only a tiny fraction of the reactor's mass in fuel can be fused over the lifespan of the reactor. IIRC, it would take something like 700,000 years for ITER to fuse its own mass in fuel, if it could be operated continuously forever (which it can't, of course.)
I'm not sure how much of the design on Earth is bulky and cumbersome because it needs to be vastly safer and more complicated than what a rocket engine needs to be. Taken to the extreme, nuclear propulsion in space can be very dumb: just detonate nukes some distance behind the rocket and ride the shockwave with a pusher plate. It's a radically different design than anything we'd attempt for terrestrial applications.
The Orion spacecraft concept you mention will use fission bombs on any vehicle we might build soon. Thermonuclear driven interstellar Orion has been proposed, but the vehicle would be enormous (and take 400 years to reach the nearest star.)
The neutron corrosion alone is very serious problem that I doubt can be fixed. There is only one material that can withstand it, its called neutronium and we all know why we cant even use it on earth, not to mention about producing one. The Li-7 trick they are currently trying to use is not convicing me. Whats the point of it if you need to stop reactor every few years to replace entire Li-7 shielding. Lithium itself aint cheap and we basically will destroy it in such reactors (non-recycleable).
The hope lies in Fission Reactors, breeder one. Unfortunately, there is big tabu here because those can be used for enrichment, and so, for nuclear weapons.
It can be fixed, because some fuels don't produce neutrons. But aneutronic fusion is like two orders of magnitude more difficult than Deuterium-Tritium. Still, it might be necessary for any reasonable application. One of my professors was a big proponent in aneutronic fusion even for Tokamaks.
I assume this refers to the sun. Other responses don't seem to have made the same conclusion.
As I said in my comment at the time https://news.ycombinator.com/item?id=42850472, the cost of solar panels has dropped in recent years to the point where it's dramatically cheaper to generate electricity with solar than from a thermal power plant.
Gas has the advantage that gas turbines are less inefficient than conventional steam, though ultra-supercritical steam (mentioned in the EIA paper) could conceivably take that away, and of course it's easier to ship around over short distances.
In particular, solar and wind won't produce any fuel at night when no wind is blowing, and may not produce enough fuel during the rainy season or winter.
[1] https://en.wikipedia.org/wiki/Electrofuel
[2] https://blog.sintef.com/energy/thermochemical-energy-storage...
[3] https://en.wikipedia.org/wiki/Fischer%E2%80%93Tropsch_proces...
I feel like oxidizing methane to methanol shouldn't be rocket science, especially if you have an unlimited amount of energy to use. NADPH does it at scale already. A few more cheap process steps would give you nontoxic ethanol instead.
The problem has always been cost. It typically costs 5x more than conventional oil. Still oil is a much denser store of energy than batteries and so for many applications it might be worth using it despite the much higher costs.
Presumably most of the cost of synfuel has been in the energy inputs, right? And solar will drive the cost of intermittent energy down by at least an order of magnitude and possibly two.
As Handmer's post says:
> (...) up until this point no-one has considered scaling these up as a fundamental source of hydrocarbons, because doing so would be cost prohibitive. Why? The machinery is not particularly complex, but the energy demands are astronomical.
Hypercane
From Wikipedia, the free encyclopedia (Redirected from Hyperhurricane)
A hypercane is a hypothetical class of extreme tropical cyclone that could form if sea surface temperatures reached approximately 50 °C (122 °F), which is 12 °C (22 °F) warmer than the warmest ocean temperature ever recorded.[1] Such an increase could be caused by a large asteroid or comet impact, a large supervolcanic eruption, a large submarine flood basalt, or "incredible" global warming.[2] There is some speculation that a series of hypercanes resulting from the impact of a large asteroid or comet contributed to the demise of the non-avian dinosaurs.[3] The hypothesis was created by Kerry Emanuel of MIT, who also coined the term.[4][5][3]
This large size has devastating implications. The nuclear island of a fission power plant is maybe 12% of the cost of the plant. Increase that cost by an order of magnitude and you've doubled the cost of the plant. Commercial nuclear power is already not competitive; double the capex and it becomes ludicrously uncompetitive.
If someone presents you a fusion proposal, the first thing you should ask is "what's the volumetric power density of your reactor?"
If you look at the arguments purporting to show DT fusion could have a chance they do things like assume they can build everything for 2x the cost of materials. Yeah, good luck with that, especially when that cost estimation technique would give wildly low results if you applied it to fission power plants, which are much less complex.
As I see it, there's only a couple of ways fusion might make it. The first is an approach that makes minimizing reactor size the primary goal. This would involve very small plasma configurations and one where all plasma-facing surfaces are covered with thick layers of flowing liquid lithium so areal power density can be very high, at least an order of magnitude higher than most concepts. Among DT efforts, only Zap looks like it might have a chance here (but there are parts on one end of the reactor that are still exposed to neutrons, unshielded.) The experience with liquid sodium in fission reactors should give one pause about making this work, though.
The other approach would be to avoid non-nuclear parts than fission power plants have by exploiting direct conversion of plasma energy to work, so the apples-to-apples comparison of heat sources does not apply. Helion is the front runner here, using more advanced fuels that put most of their energy into the plasma, not into neutrons. I consider Helion the least dubious of all current fusion efforts.
Even if fusion reactors can be made reliable, they'll still require a huge amount of investment with a payoff decades into the future. Meanwhile, you buy some cheap solar panels and can get returns on your money almost immediately.
It's important for clean energy to be cheap because that creates an economic incentive for people to invest in switching.
The good news with renewable energy and battery storage is that we're basically there. Texas which produces lots of oil also produces lots of clean energy. Because oil refineries require a lot of energy to refine their fuel. And of course using your own product at a premium when clean renewables do the same job would be madness. Think about it, the very companies that are expecting others to buy and burn their fossil fuel are investing (heavily) in renewable energy because that's cheaper than using their own cheap supply of oil and gas for the same purpose. It's more lucrative to sell that than to use it for them.
That's why records are being broken with amounts of solar, wind, and battery coming online on a regular basis. It's cheaper. So, people do more of it as fast as they can scrape the funding together to do it.
Fusion if it happens at all is probably going to be really expensive for quite some time. Nuclear has the same problem. Clean but just super expensive. And slow to plan and execute. You need to decide decades in advance that you need it in order to have it ready. Most reactors coming online this year (hint: not a lot) were planned when solar and wind were still expensive (1-2 decades ago). In other words, they probably did not take into account current price levels of new clean energy. In terms of profitability these things are kind of dead on arrival. Fusion faces the exact same challenges. It's not completely hopeless. But we need 1-2 orders of magnitude reduction of cost per energy generated this way.
Cheap and expensive in dollar terms is meaningless. Dollars are numbers in computers. Joules and watts are real.
What I can agree with we need good enough EROI energy and then adjust the policies regarding fusion, nuclear and fossil joules, watts and their related numbers in computers.
Because most of the numbers coming out of the nuclear sector for this are too good to be true. To put it mildly. It's tempting to use more harsh language here. Like "fabricated nonsense" or "a scam". And some people do.
Probably the biggest challenge for nuclear proponents aside from demonstrably high cost, overruns, wildly optimistic timelines, etc. for past projects is the credibility of their predictions about cost for future ones. The lower those predictions, the smaller the chance that those are anything more than wishful thinking mixed with a lot of hopium and based on very flawed assumptions.
When cost overruns in nuclear projects are routinely measured in hundreds of percents, something is very wrong. It would be nice to get that down to two digit percentages. That would still be terrible but a lot better than the industry norms for this. Especially compared to renewable projects.
This credibility challenge is a large part of the reluctance of investors to fund new efforts in this space. Most of the projects that actually still happen feature governments and tax payers taking most of the risk. Which makes it more likely that the huge budget gaps get funded rather than resulting in investors cutting their losses and bailing.
If you want numbers, look at numbers of GW coming online for nuclear, wind, solar, and energy storage projects. There's a stark difference there. Renewables are outpacing nuclear at such a pace that you could legitimately wonder why to bother with such projects at all.
Indeed, especially as renewable will more and more eat into nuclear 'baseload', lowering its load factor and therefore raising its total production cost. https://www.youtube.com/watch?v=udJJ7n_Ryjg
A fusion reactor is going to create and consume large quantities of tritium, which is difficult to control because it is an isotope of hydrogen and hydrogen can go right through metals. It might not really be a problem but it is always going to be in the paper that somebody detected tritium around a reactor.
Also, because fusion reactors produce large quantities of neutrons, you could put U238 or Th232 into the blanket and breed either Pu239 or U233. Back in the 1980s some engineers thought that a hybrid fusion-fission system could have better economics that either alone if you used the bred fuel in thermal reactors. Those materials could just as well be used to make bombs, although, unlike a fast breeder cycle, a fusion reactor doesn't require plutonium processing technology in order to work.
---
As for solar and wind being "cheap" don't believe the hype. I mean, it is cheap if you don't mind that it is only available when the sun shines and when the wind blows. You see estimates for solar + storage that are deceptive and thought-stopping like Case 17 of [1] (an hour and a half of storage is significant, but it is not comparable to the 24 hour service that fossil fuels and nuclear can provide.)
I haven't seen a detailed analysis of the generation + transmission + storage costs of a 100% renewable system published in the last decade. If you try a back of the envelope calculation it usually winds up worse than Case 9 from that report, except that Case 9 is optimistic because (so far) it always costs a lot more than you estimate to build a nuclear plant. [2]
[1] https://www.eia.gov/analysis/studies/powerplants/capitalcost...
[2] to get some sense of how bad the problem is you need: (a) to get through the night (12 hours of storage? still might beat AP1000...) and (b) get through extended periods of bad weather (2 weeks of storage?) and (c) get through the winter (overbuild solar 2-3x? otherwise 6 months of storage?)
BTW, the battery cost estimates there in that link seem very high, not reflecting the recent collapse in LFP prices from China.
Australia updates their GenCost every year, which feeds into their Integrated System Plan. The draft of this year's is up for public comments:
https://www.csiro.au/en/research/technology-space/energy/gen...
> In this report, where we make a comparison between the costs of variable renewables such as solar PV and wind and the costs of other technologies we include the cost of firming those renewables which we call integration costs. These are the additional costs of ensuring supply is reliable when using intermittent energy sources. These integration costs are itemised in the report and include storage, transmission, system security and spilled energy
https://www.sciencedirect.com/science/article/pii/S030626192...
About the worst you can find in terms of small geographical area and having to deal with both Dunkelflautes and arctic conditions spilling south which often leads to -5C to -15C and near zero wind.
Of course excluding that storage is plummeting in price with the latest Chinese auctions with bids submitted December 2024 landing on $62/kWh installed and serviced for 20 years.
https://www.ess-news.com/2025/01/15/chinas-cgn-new-energy-an...
I wish SMR gets more attentions and focus. The UK has more Wind resources to power all four nations and it is on course to double its current 30% electricity from Wind to 60% by ~2030. I dont see why it cant be 90%+ by 2040 barring policy, politics and human factors.
All we need then are SMRs for base loads, while we work on energy storage that could be battery or Dams.
https://spectrum.ieee.org/the-forgotten-history-of-small-nuc...
Or just this recent summary on how all modern SMRs tend to show promising PowerPoints and then cancel when reality hits.
https://www.youtube.com/watch?v=XECq9uFsy6o
Simply look to:
- mPower: https://en.wikipedia.org/wiki/B%26W_mPower - NuScale: https://oregoncapitalchronicle.com/2024/10/29/the-rise-and-f...
And the rest of the bunch adding costs for every passing year and then disappearing when the subsidies run out.
Not sure why you would want to fill storage with extremely expensive nuclear power when we might as well just use cheap renewables.
https://www.gevernova.com/nuclear/carbon-free-power/bwrx-300...
Which is especially true considering how renewables and storage is plummeting in cost seeing incredible growth.
You know, just like Flamanville 3 was back in 2005 before they had a 7x cost over run and 13 year delay on a 5 year project.
Try get them to sign a fixed price contract for delivery at said competitive price. We all know they will run scared from the room because otherwise they would be selling their product using that method.
Today they are relying on absolutely enormous government subsidies, where the government both subsidizes the electricity and takes on the project risk.
Does it really need to be cheaper if it's more promising?
It needs to promise to eventually be cheaper, or otherwise better in some niche.
I can imagine it being useful for spaceships, for example.
But I can't see it ever being cheaper than eg solar, without some kind of breakthrough, like cold-fusion.
See also their recent commitment to using AI. Funding AI research and supporting AI startups: sure. Basing policy on the fact that AI can replace civil servants and other workers: there should probably be some evidence of that first.
1. https://orcutt.net/weblog/wp-content/uploads/2015/08/The-Tro...
Congrats! All the hassle of fission with more steps. Proceed to scream at me!
It just seems like a great way to turn a billion dollars into a pile of extremely hot, radioactive, radiation embrittled metal.
ITER seems to have a project called the Agile Robot Transporter which you can see some renders of that demonstrate how fiddly this problem is.
https://indico.iter.org/event/45/contributions/1090/attachme...
Fission plants are operated remotely. IE, that's what the stereotypical "control room" of a fission plant is.
There was a group that tried to analyze the expected uptime of DEMO, the pre-commercial follow on to ITER. With the expected MTBF (mean time between failure) and MTTR (mean time to repair) they estimated it would be working just 4% of the time. That's too low even for a research machine.
RAMI (reliability/availability/maintainability/inspectability) is one of the major issues for fusion, and even if nothing else stops fusion it alone will make the development process very long and iterative. Good RAMI requires long iterative improvement of individual components and processes.
1. The fuel is free to cheap, depending on the fuel; and
2. Stars can do it so surely we can.
So let's consider a mythical plant that costs $30B to build, has a lifetime of 30 years, costs $3B/year in maintenance and produces 1000MW of power. And the fuel. is free. Is the electricty free? Not even close. That's ~8.8GWh per year. No plant runs at full capacity 24x7 so let's call it 4TWh. That's about $1/kWh. That's ludicrously uneconomical. Of course you can throw in vastly different numbers and vastly different results but the main point is that fuel cost is a non-issue.
And as the article points out, in a D-T reactor, you still need to produce tritium because it's basically not naturally occurring in a way that can be gathered. That means a fission breeder reactor. That fuel isn't free.
Second, "the Sun does it" misses the point. As the article correctly points out, damage from neutrons ("neutron embrittlement") is a huge problem. Beyond that, it's a huge problem because every high-speed electron that escapes the plasma is energy loss from the system.
Stars use gravity for containment.
It doesn't require a fission breeder reactor. All DT schemes imagine a blanket where they breed their own tritium, and produce enough to start additional DT reactors. Some imagine a startup phase where they run just on DD for a few months, with Q < 1, to make enough tritium to start up. If fission had to be used to make all the T then DT fusion would be a complete nonstarter economically.
My position is they're the least dubious of all the fusion efforts, and have a very clever design that offers the plausible possibility of avoiding some of the engineering and economic showstoppers confronting DT fusion. Of course no pudding can be eaten before it's cooked, and no reactor design can be counted on before it's proved.
There may be a reasons we don't have fusion energy that have nothing to do with the physics or engineering.
What if I told you "fusion is not (perpetually) 20 years in the future... it's 60 years in the past and we missed it." ?
There's a story here, but it requires a bit of suspension of orthodox belief:
https://waterstarproject.com/read-the-waterstar-manifesto/
https://waterstarfoundation.org
-- Paul Schatzkin (actual name!)
https://news.mit.edu/2024/commonwealth-fusion-systems-unveil...
The fusion reactor would just fizzle out the moment the vacuum is breached.
Not sure I can say the same for the bomb.
--PS
In the same time the UK has been arguing over a third runway at its biggest airport, China has built over 100 new airports (20 years).
The hard part here is twofold: there are several orders of magnitude needed to be overcome (in efficiency) for it to work. Also the learning curve in fusion is not a curve, but more like a staircase with very tall steps
ITER is gearing up to be a great disappointment
I really don't feel like magnetic confinement is the way to go, unless we get some new superconducting technology and need to create electromagnets with 100x their current strength
Did you mean superconductor technology?
I don't think it's 100x, but it doesn't have to be. Tokamak output scales with the fourth power of magnetic field strength. Double the field, 16X the energy output. CFS and Tokamak Energy are building tokamaks with these.
Tokamak output also scales with the square of reactor size, which is why ITER is so big. These new reactors can be much smaller for the same output, so they'll be much faster and cheaper to build. The CFS design is about the size of JET, which was built in four years, with three of those being just for the building.
100x ARC would be ~2000 T. The magnetic pressure of such a field would be about 100x the detonation pressure of TNT. Experimentally, fields of such magnitude could only be produced very briefly in devices that near-instantly explode (for example, devices driven by intense laser pulses that produce large charge separations.)
> Tokamak output also scales with the square of reactor size
ITER is big because it wouldn't work at all if it were smaller (at the magnetic fields involved). Output scaling as (linear dimensions)^2 means you want your reactor to be as small as possible, to maximize power/volume. Square-cube law in operation. The higher magnetic field in HTSC tokamaks allows then to be smaller. Volumetric power density is still poor, though, just not as terrible as ITER.