Natrium 'advanced nuclear' power plant wins Wyoming permit
wyofile.com
wyofile.com
The fission cross section for such energetic neutrons is much lower than for thermal neutrons. Therefore, there has to be a much greater density of fissionable material in the reactor core.
The lack of a moderator also means rearrangement of the core in an accident is potentially much more dangerous. If the fuel itself rearranges to become more compact, say by melting and flowing, the reactivity could increase. This is not possible in (say) a light water reactor, where such a rearrangement would reduce reactivity.
The nightmare scenario for any fast reactor, warned about by Edward Teller in 1967, is a rearrangement that causes the core to become supercritical on prompt neutrons alone (that is, on only the neutrons released promptly at the moment of fission, not on those + the delayed neutrons emitted by some fission products as they decay). A fast prompt supercritical configuration could potentially explode with great violence, greater than Chernobyl. An atomic bomb is a prompt fast supercritical system.
I will want to see how the NRC does or does not license their design, a process that has just started. I will not be surprised if their approach ends up being unlicensable in the US because safety cannot be assured by analysis under accident conditions.
> An atomic bomb is a prompt fast supercritical system.
Yes, it is, but more precisely it is a fast hypercritical system. The difference is this: in a supercritical system the ratio of neutrons in a new generation vs the old generation (generally denoted by k) is higher than one. For example 1.001 qualifies as supercritical. In a hypercritical configuration, k is higher than 2. In nature things generally evolve in a continuous way. To get to 2 you need to first get to 1 and 1.001, for example. For a bomb, it is very important to get to 2 extremely quickly, so that the assembly does not start the chain reaction while k is 1.001, or 1.1. To do that, you use either a gun or high explosives, to quickly reassemble a subcritical configuration into a hypercritical one, before a stray neutron has a chance to start a barely supercritical reaction (also known as a "fizzle").
With a reactor, there's no way to get an atomic bomb effect, because you go through the point of k=1.001. When you get there, there are plenty of neutrons around that this results in a "fizzle". Heat is produced and the core dilates and the ratio becomes low again. This results in "prompt excursions", each one lasting less than one second, each one some sort of a "bomb fizzle". These excursions are annoying, but they are not Hiroshima.
> I will not be surprised if their approach ends up being unlicensable in the US because safety cannot be assured by analysis under accident conditions.
What you are saying is a Catch-22. NRC can't approve a new design because it doesn't know how it performs under accident conditions, but then you can't know how something performs under accident conditions if you don't build it, and you don't build it if NRC does not approve it.
The fact is, Russia has been operating the BN-800 sodium-cooled fast reactor for 10 years [1], and the smaller BN-600 for 45 years. So sodium-cooled fast reactors can work, and Russia likes them enough that it plans to build a bigger version, BN-1200. They don't explode. You could say "not yet", and sure thing, this means that the NRC needs to work hard to cover all the bases. But this is not impossible.
> What you are saying is a Catch-22. NRC can't approve a new design because it doesn't know how it performs under accident conditions, but then you can't know how something performs under accident conditions if you don't build it, and you don't build it if NRC does not approve it.
Even if you build it, and even if there is an accident, you don't know how it behaves under accident conditions. You just know how it behaved in that particular accident.
It might be easier for developers if we could build thousands a fast reactors and let some of them explode to get some statistics on what works, but that's not how the NRC is arranged, and even that would not ensure all the catastrophic risk had been retired.
Nuclear engineers know how k behaves, they don't need to make assumptions. Here's a 101 on reactivity feedback effects by the NRC [2].
The US has built all sorts of nuclear reactors, sodium-cooled included. Even one that was sodium cooled and used molten (hot) plutonium as fuel [3].
The experience with prompt criticality excursions is (for the better or worse) extensive [4].
The NRC absolutely has the ability (and duty) to analyze all possible accident conditions for this reactor. This is the page it maintains for Natrium [5]. They have already received and reviewed thousands of documents for it. We would not be at the point that they allow the building to start if they think there's a slight chance the reactor would be "unlicenseable".
[1] https://terrapower.github.io/armi/
[2] https://www.nrc.gov/docs/ml1214/ml12142a130.pdf
[3] https://www.osti.gov/biblio/4206527
[4] https://en.wikipedia.org/wiki/Prompt_criticality#List_of_acc...
[5] https://www.nrc.gov/reactors/new-reactors/advanced/who-were-...
I do wonder how you think all possible accident conditions could be evaluated. All possible geometries of melted fuel? The possibilities are vast, indeed exponentially vast.
Edward Teller in his 1967 comment expressed skepticism that the consequences of a serious accident in a fast breeder could be evaluated (yes, I know Natrium is not quite this configuration, uranium having a considerably higher bare core critical mass, but it's similar):
"For the fast breeder to work in its steady-state breeding condition you probably need something like half a ton of plutonium. In order that it should work economically in a sufficiently big power-producing unit, it probably needs quite a bit more than one ton of plutonium. I do not like the hazard involved. I suggested that nuclear reactors are a blessing because they are clean. They are clean as long as they function as planned, but if they malfunction in a massive manner, which can happen in principle, they can release enough fission products to kill a tremendous number of people.
...But, if you put together two tons of plutonium in a breeder, one tenth of one percent of this material could become critical.
I have listened to hundreds of analyses of what course a nuclear accident can take. Although I believe it is possible to analyze the immediate consequences of an accident, I do not believe it is possible to analyze and foresee the secondary consequences. In an accident involving a plutonium reactor, a couple of tons of plutonium can melt. I don't think anybody can foresee where one or two or five percent of this plutonium will find itself and how it will get mixed with some other material. A small fraction of the original charge can become a great hazard."
But the problem with this argument is that it can be applied to anything. In particular, if you really follow this argument, we should shut down all the current operating reactors, because, you know, there might be disaster scenarios that we didn't consider.
I hear you, I hear you. This argument only applies specifically to fast reactors, because this is what Dr. Teller said in 1967. But why did Teller say that?
In 1961 a nuclear reactor that the US Army commissioned with the stated intent to be virtually idiot proof, went boom. It became prompt critical. What happened was one of the most (if not the most) life-beats-fiction events in the history of nuclear reactors. It appears one of the operators decided to commit suicide by pulling a control rod as fast as he could [1]. Of course, since the guy died in the following half a second, we'll never know his true motives.
Another thing is that Teller was profoundly anti-communist. In 1957 another mind-bending accident happened in the Soviet Union, [2]. It had to do with plutonium, it went something like this: the Soviet Union was producing lots of plutonium to build bombs. Plutonium is extracted by chemical separation from spent uranium. After the separation, various steps are taken to increase the concentration, and eventually to purify the plutonium metal, but some chemical tailings that contained plutonium in low concentration remained. They were dumped in large underground tank. The solution contained ammonium nitrate, but it was mostly water. Water is a good moderator, so there's a good chance the solution would from time to time undergo prompt criticality excursions. Nothing major, no explosion happened. Yet. But because of the periodic heating, the water evaporated, and in time the ammonium nitrate became more concentrated, and it became a huge amount of (chemical) explosive. At some point this ammonium nitrate decided to go boom, like it happened in Beirut in 2020. The explosion was totally non-nuclear, but it put in the atmosphere a huge dirt cloud full of plutonium and other radioactive nasty stuff. The Soviet Union kept this a secret for many, many decades. But some people in the West learned some tidbits about this, and Teller, obsessed as he was with the Soviet Union, might have been one of the guys in the know. Plus, Teller (the father of the hydrogen bomb) had access to the highest level of top secret information in the US in regards to nuclear things. If anyone in the US knew anything about the Kyshtym, then Teller knew that too.
But why did Teller say that about fast reactors and not thermal reactors? Because that was the context. He was asked to testify if the US should build a sodium-cooled fast reactor. And he said it's a bad idea. If you know anything about Teller, is he said everything with a supreme confidence. Everything was very black and white for him, and he always knew with utmost certainty what was black and what was white. He thought for example that we should use hydrogen bombs to dig canals. Or that we should put anti-ballistic-missile thingies orbit that would zap the incoming missiles with a beam of X-rays produced by a nuclear bomb explosion [3]. This is not a joke.
Ok. Let's leave Dr. Teller aside for now. Are sodium-cooled reactors too risky to build. Well, patently not, considering that Russia is operating 2 right now and plans to build more. But let's say you are not happy with this. Will there be a point where you would concede that the experience the world has with building and operating such reactors is sufficient for the NRC to start granting such licenses? If not, what makes fast reactors any different that pressurized water reactors?
And by the way, the Natrium reactor is not designed to burn plutonium. It is not designed to be a breeder reactor. And the issue about prompt neutrons is actually less of a problem for such a reactor: U-238 has the highest ratio of delayed neutrons to prompt neutrons of any fissionable isotope and for fast reactors a larger proportion of fission events involve U-238 compared to thermal reactors [4].
[1] https://en.wikipedia.org/wiki/SL-1
[2] https://en.wikipedia.org/wiki/Kyshtym_disaster
[3] https://en.wikipedia.org/wiki/Project_Excalibur
[4] https://www.nuclear-power.com/nuclear-power/fission/prompt-n...
I understand the risks around sodium, but the "passive natural circulation cooling" I don't understand. Is it more feasible with this design and why?
" Pros:
high temperature means we can use process-heat which is a much more efficient use of heat.
fast spectrum neutrons means we can burn importantly troublesome parts of nuclear waste.
fast spectrum is also better for breeding new fuel, significantly increasing how much energy we can extract from uranium/thorium.
passive natural circulation cooling is much more feasible.
fast spectrum is a little more complicated to control.
fast reactors require high enrichment.
inspection of the plant is very difficult with liquid metal.
high temperature liquid metal doesn't play nicely with metal pipes.
sodium burns in air and is explosive with water.
we simply do not have nearly as much experience with sodium as we do water and that really cannot be understated.
"In large reactors, these features have not been enough to compensate for the disadvantages and sodium-cooled reactors have not been successful, coming in more expensive than light water reactors for a given power output. France, which had been developing fast reactors, has recently mothballed the effort with no plan to restart before 2050.
Wouldn't you just design the shape of the reactor so that if it got too hot for any reason, the shape it would melt into would result in a less compact geometry that would slow down rather than speed up the reaction?
About the only approach I'd be comfortable with would be dissolving the fuel in molten salt (probably chloride salt). This is not Natrium's approach.
Melting of fuel is not a theoretical problem -- it has actually happened at two fast reactors in the US (EBR-1 and Fermi-1, the latter the reactor in the hyperbolically titled book "We Almost Lost Detroit"). No explosions occurred, but it's very troubling the fuel melted at all. The NRC will surely insist on analysis of the consequences of partial fuel melting accidents.
> For example, how do you prevent melted fuel from flowing into the cooling channels that go through the core?
Expect that to happen and use a geometry that doesn't cause the reaction to increase in speed if it does, e.g. because fuel flowing into the cooling channels would make the fuel less rather than more compact.
Cooling typically means things like maximizing surface area, minimizing the thickness of the object being cool, etc.
Maximum neutron density presumably happens in a sphere, which is coincidentally the shape that minimizes surface area.
The whole point of a nuclear reactor is that it heats up, and you can convert that heat into useful fuel. Presumably that means you need to carry quite a lot of heat away per volume. Presumably that means putting the fuel into a spherical shape really doesn't work that well.
Uranium oxide melts at 2865 C; uranium metal at 1132 C, plutonium metal at just 639 C. In contact with iron, plutonium forms a eutectic with a melting point of just 410 C, below the melting point of zinc. There was a crazy reactor at Los Alamos, LAMPRE, that used molten eutectic Pu-Fe in tantalum tubes as the fuel.
Or use uranium carbide. High melting point with still decent thermal conductivity.
Next time you complain that waiting for a code review approval till next week is excruciatingly long, think about these turnaround times.
Also this is why we can't quickly build many reactors to ramp up electric generation for millions of new electric cars, etc.
But on the good side, we _didn't_ have a hydrogen explosion due to some late-night troubleshooting, great work Carl!
But we learned a lot, our alerts were confusing, and our metrics didn't surface coolant level or the relief valve that was stuck open.
So next sprint, let's focus on monitoring and cross-team knowledge transfer! 3 points, right?
"We" here means any country on the planet. Nuclear plants are simply inherently more complicated and more risky to build that the alternatives. Even China, who often rightly, gets criticized for pushing through individual rights on such projects can't build them quickly.
Look at the shape of these curves [1], nuclear is flatlining while renewables are accelerating at incredible speed.
[1]:https://cleantechnica.com/2023/02/06/renewables-in-china-tre...
No. The reason is that they have not been price competitive with renewables.
And so there isn't the volume of approvals that gives regulators experience which in turn reduces approval times.
Commercial solar takes months to approve by comparison.
And right now renewables continue to get better and cheaper so it will be hard for nuclear to get much traction.
You take so much for granted.
If these markets produce their version of 2008, building renewables and no storage, and then we have blackouts - will you stand by those words if the power grid collapses and people freeze to death?
With the ‘markets’ not ask for government intervention and bailouts at my expense?
We have identified a few key areas where pricing energy is not enough and therefore price available of physical properties.
Not sure what point you think you are making.
Following their troubles they strengthened ancillary services and created new regulations to ensure stability.
All this made them quite easily manage the last winter storm in January 2024.
https://www.spglobal.com/market-intelligence/en/news-insight...
Says who? This is an opinion, not a fact.
All of the newer SMR companies are pitching their design as costly at start, but then cheaper because of mass production. Well the entire gaggle of companies won't all be able to mass produce their SMRs in the numbers required to get the scale down since there isn't the appetite for it.
> there isn't the appetite for it.
A distinct without a difference.
I agree on regulatory experience but that comes more from panic over nuclear proliferation rather than any economic reason. Whether or not you view that as a legitimate concern is open to debate.
The primary cost of running solar plants is no longer the panels themselves (ie if they drop to 0 the $/KW wouldn’t change a whole lot).
Maybe people should read something from the actual domain experts instead? The latest CSIRO report specifically responded to the claims made about extending the time period.
> CSIRO said accounting for a 60-year life reduced the costs of nuclear by about 9% compared with calculating over 30 years, but found other technologies such as solar and wind saw similar cost reductions of about 7% under the same approach.
You yourself seem a bit confused about facts after using these sources.
1. Nuclear is not anywhere near the same price as solar. (I'm hoping you typo'd and were really trying to make the much saner but still wrong claim that renewables plus storage has costs comparable to nuclear which this document tries to claim)
2. It's not just the panel costs that have dropped, the delivered cost of solar energy has been falling. It dropped 12% in 2023 alone. This has lots of causes, like the safe and predictable success of solar allowing for lower interest rates and cheaper panels allowing for other assumptions to be rethought (e.g. ditch trackers that used to maximise output per panel and use the money and space saved to buy more panels).
Well, nuclear is 24/7 base load. Nuclear can't peak that well. Also, when nuclear goes offline it does so for days.
You didn't check, you decided to cook the presentation of the facts.
Indeed, solar is not available 24/7. But it's not individual installations which need 24/7 availability, it's the grid as a whole; plus, energy use during the day and night differs, so even the grid as a whole doesn't need the same generation capacity 24/7. And - the grid has many different energy sources (i.e. not just solar and nuclear), some renewable and some non-renewable, which aren't daylight-only.
Which is to say there is plenty of room for solar installations irrespective of energy storage solutions. And indeed, those installations are taking place, and they are rather cheap.
The peak to trough is ~30% reduction if I recall correctly. It’s different but not massively so. And long distance energy transport is insanely expensive. It’s being done but mainly to stabilize the grid rather than arbitrage. And your entire solar capacity in the US goes offline for ~12+ hours. That’s a lot of energy you’re going to have to supplant in other ways regardless of long distance.
> And indeed, those installations are taking place, and they are rather cheap.
Because the cost of storage and the destabilization of the grid it causes the provider are treated as externalities.
When your basic assumption is that an entire industry is doing something dumb, it’s probably wrong. If nuclear was actually that cheap electric companies would constantly be proposing new projects, but it’s not.
> 24/7 peak availability
The industry abandoned nuclear because electricity demand stopped increasing decades ago. It simply isn’t cost competitive when facing both significant curtailment and the need for backup generation for the multiple weeks to months downtime nuclear power plants have semi annually.
That’s the core issue nuclear needs both 24/7 demand, and the grid to also be perfectly fine when it goes away for months. That only possible when it makes up ~30% of supply which is roughly where it’s been for decades.
Now having not built nuclear for decades once off projects are constantly wildly over budget and behind schedule. Solar + batteries is currently beating even optimistic estimates for nuclear just about anywhere, let alone current boondoggle pricing.
The industry is doing the right thing but the reason it’s not choosing nuclear is because regulatory burdens make project success and prices highly variable and uncertain, not to mention a shift in political winds kills your project. The reason isn’t purely a price thing.
> The industry abandoned nuclear because electricity demand stopped increasing decades ago
There’s no way this is true considering that crypto and AI have started to push significant power demands.
> Now having not built nuclear for decades once off projects are constantly wildly over budget and behind schedule. Solar + batteries is currently beating even optimistic estimates for nuclear just about anywhere, let alone current boondoggle pricing.
That’s a choice we get to make. China is building their nuclear muscle because they realize it’s required for baseload for solar, way cheaper than batteries, and the only way you’re going to be able to charge all these EVs.
I was describing what happened when we stopped building nuclear power plants, not what’s happening now or in the near future.
From 1.7TWh in 1975 to 3.1TWh in 1995 is an 82% increase in demand in just 20 years, but the next 20 years only takes us to 3.9TWh in 2015 a 26% increase. That had massive knock on effects in terms of power plant construction, move forward to 2023 demand is only at 4TWh less than a 3% increase. https://www.statista.com/statistics/201794/us-electricity-co...
The point of that was simply context.
> The reason isn’t purely a price thing.
All that risk ultimately just equates to a cost, if nuclear was wildly profitable to operate nobody would care about those risks but it’s at best barely breaking even when nothing goes wrong.
Also, costs rose not because of regulations in a vacuum. 3 mile island happened because of a lack of maintenance, you don’t solve that problem without spending more money on maintenance. And so it goes for a huge range of issues like foreign material in cooling ponds. Suddenly there’s all this effort when working around cooling ponds to avoid a multi million dollar issue and that’s not free, but there’s no free lunch where you can avoid the expense and not also be cautious.
Thus nuclear power was always expensive, the costs where initially just hidden as risk and escalating maintenance costs.
Right. My point though is that that cost is because of political decisions we make, not inherent in the technology. We could also choose to make solar projects equally risky in such ways; force solar projects to require to be baseline generators, not risk grid stability, and stop subsidizing home panel installations and see what happens to the solar industry. I’m not saying we should do that because solar is an important partner in the energy mix, but nuclear is more important to decarbonize as we still fail to see a drop in green house emissions on any grid that’s invested purely in renewables with only nuclear showing that it can offload carbon inputs; renewables have only generally shown the ability to level off growth of carbon inputs at the grid level (unless you look relatively small grids not in the richest nations, but the richest nations are the biggest polluters).
> Also, costs rose not because of regulations in a vacuum. 3 mile island happened because of a lack of maintenance, you don’t solve that problem without spending more money on maintenance
Just not true on several fronts. First the consequences of 3 mile island were wildly overblown and generated unnecessary fear. The cleanup cost was $1B which is negligible. In fact the plant is being restarted on the other reactor by 2028. Secondly, reactor design is hugely important. Fukushima and 3 mile island are Gen II reactors while Chernobyl was a Gen I reactor and we can see the drastically smaller issues we’ve had from such reactors. Gen III reactors are much safer and are cheaper while Gen IV reactors are generally completely fail-safe (i.e. doesn’t matter the quality of maintenance on the safety of the reactor). So pretending like nuclear has a constant maintenance cost that’s ever increasing ignores our ability to solve problems like this at the design stage. The problem is that regulations that remain stuck responding to Gen I designs being constantly applied to newer generations.
“1 billion dollars” is not negligible when your company is paying it either directly or through insurance. They also lost a nuclear reactor (~$2.5 billion in 2018 dollars to build - depreciation) and lost revenue, so they were out far more than just 1B.
Worse this isn’t a one off event, every single nuclear reactor was facing the same issue.
Just from a pure economic perspective if someone says you need to spend say 20-30 m/year to avoid losing your nuclear power plant and paying 1B on top of this your going to do some internal calculations and figure out how much maintenance is worth 1B dollars and frankly it’s a lot of maintenance. But now every single nuclear reactor is now paying more every year on maintenance forever. The industry just got more expensive seemingly from thin air thus 3 mile island cost the nuclear industry a lot more than just what was spent on the island.
Worse the same thing happens after basically every expensive incident. Almost never does the industry look at an expensive accident and say well that’s just the cost of doing business we don’t need to change anything. Which then means every year it just keeps getting more expensive.
Checked where? Lazard says it is one fifth of the price.
>There’s a huge difference in capability though because nuclear is 24/7 peak availability which solar can’t do
Not 24/7. They have to be turned off for maintenance and when that happens e.g. France burns an awful lot of gas.
Pair solar with wind and storage and it's still cheaper than raw, nuclear power and would be even if matching storage were 4x the cost.
Pretty sure >90% of France’s energy consumption is completely green. Renewable grids wish they could boast these kinds of numbers.
The rountrip efficiency of this is shit (~40% compared to 90% for pumped storage or batteries) but the important thing is it's cheap to store a lot of it for a long time and straightforward to use a lot of it all at once during those rare dark, windless spells. For natgas we even already have the infrastructure.
What's really startling is the cost of producing a kwh from stored syngas/hydrogen generated from solar and wind is similar to the cost of getting a kwh from a nuclear power plant.
So, worst case storage + solar + wind ~= average case nuclear power.
But even if nukes were somehow very cheap to build, the review process due to high safety requirements and a lack of a standardized reusable reactor design would make the process slow. To become cost-competitive, nuclear reactors need to be mass-produced, using a proven design that needs little review. The French sort of achieved that, with two mass-produced types of reactors.
Renewables as baseload are still problematic though. Solar cells and even wind turbines are not expensive, but the batteries are very expensive, and are a huge fire hazard. It would be interesting to estimate how much would a kWh cost from a nuclear power station, and from an equal-power LiFePO4 battery installation with the capacity of, say, 3 days worth of the rated max power. Batteries can be replaced gradually, but would need to be replaced much sooner.
[Update:]
If we take a modest nuclear generation unit producing 300 MW of electric power, I'd like to compare it with a battery that can store 300 MW * 3 days of energy from renewable sources. It's 21.6 GWh. With LiFePO4 typical bulk price of $100 per kWh, the upfront cost in batteries alone would be $2.16 billion. It's still lower than nuclear reactor equipment, but very much in the same ballpark.
Have you seen how the fire risk petroleum presents? It's crazy.
This requires very special approaches to the architecture of a battery installation, and likely rather special insurance premiums.
You need to factor in waste disposal and decommissioning costs for nuclear.
And if you are concerned with the fire risks of batteries you may want to think about the impact of serious events with a nuclear reactor. In almost all cases they end up being quite expensive.
Have you heard of non state actors? Would be a shame if US built a breeder reactor in each state and oath keepers or whoever got a hold of one.
States that want nuke tech can just get it from:
- China - Russia - Iran - North Korea - Pakistan - UK (soon, ~1 decade) - France (soon, ~1 decade)
Non-proliferation is a joke. Always has been.
https://wyofile.com/fate-of-natrium-nuclear-plant-may-depend...
"In the current stockpile, the average duration since a warhead was manufactured or refurbished is roughly 28 years." [DoE]. I suppose some of the aging warheads will be reprocessed as fuel.
[CNN]: https://www.cnn.com/2024/09/09/climate/nuclear-warheads-hale...
[DoE]: https://www.energy.gov/nnsa/us-nuclear-weapons-stockpile
(Previous version of this comment was incorrect.)
For southern states, or states that get a lot of solar power, they shouldn't have nukes.
For areas that have plenty of wind, they shouldn't have nukes.
For areas with lots of people, they shouldn't have nukes.
Areas of seismic activity, they shouldn't have nukes.
Research needs to be done on how to lower the cost of interconnect and installation of solar power... like billions of dollars of research rivaling nuke research.
Following nuclear news would be absolutely tragic if I didn't think it was all a pointless diversion from the real work being done. As such their lack of progress leaves me a bit more ambivalent.
dont know much about the subject but this will have more or less the same runway as gasoline
Huh? Is this something where there's multiple incremental steps in the process, and that date is just the final approval stamp, or does it actually just take more than 1.5 years?