How do metal alloy fuel fast reactors respond to rapid reactivity insertion events?
Update: (Posted Feb 21, 2015 at 7:22) The title has been modified after initial discussion indicated it was incomplete. Other related updates are in blue font.
Fast neutron spectrum reactors offer one answer to the trump question that is often used to halt informative discussions about using more atomic energy to reduce our excessive dependence on burning hydrocarbon based fuels — “What do you do with the waste?”
In a fast neutron spectrum, very little of the long-lived radioactive material is unusable waste; instead, it becomes a valuable fuel resource. There are detailed explanations available on Atomic Insights and throughout the web that explain why that statement is true and why it is so important to mankind’s future prosperity.
The simple version is that we currently produce about 11-12 million barrels of oil per day equivalent in nuclear plants that use only 0.5% of the initial potential energy of mined uranium. The rest of the mined uranium is often misclassified as “nuclear waste” because we are not putting it to good use. It is being stored as either depleted uranium or as the major component of reusable nuclear fuel — a valuable energy reservoir that many publications still refer to as “spent nuclear fuel.”
Fast spectrum reactors offer the long term potential to use much closer to 90-99% of the potential energy, consuming the portion sometimes demonized as “waste.” In the process, they produce a stream of byproducts that might have commercial uses, but even if they are not used, they have half lives that are generally substantially less than 30 years. That means they decay away in less than a few hundred years rather than in tens of thousands of years.
Though there is plenty of refinement remaining before fast reactors can be widely used as commercial power plants, their technical potential is incredibly exciting for those of us who believe that human society needs as many sources of reliable, abundant, ultra-low pollution power as our creative minds can invent.
I like providing power to the people; my interest in nuclear energy stems from my belief that we should use “the best of the above” for a given energy application like propelling ships or generating industrial heat in preference to equal measures of “all of the above.”
The above paragraphs are background information for the remainder of this post.
On numerous occasions when Atomic Insights has mentioned using fast spectrum reactors as an additional tool in our quest for improving society’s power options, there have been comments that question their response in the event of rapid reactivity insertion accidents. They fear that fast reactors can suffer failures that carry the risk of harming the public.
Some of the commenters pose their reservations in a credible way, suggesting that they have done some serious research and been unable to reassure themselves that there are good answers. Because they have been unable to find answers to questions that are worth worrying about if unaswerable, they are adamantly opposed to fast reactors. Here is a concluding quote from a note submitted by Bill Hannahan.
Supporters of solid fueled fast neutron reactors have not published computer modeling of a high reactivity rate accident in a full sized core. Have they no curiosity? I suspect it is because they know or suspect that the results would terrify the public, and their beloved IFR would never be built.
I really enjoyed reviewing the Bethe calculation because he writes like a brilliant teacher who wants to share his knowledge with others. He lays out all of his assumptions and simplifications, AND, explains the underlying reasons each of them.
NUREG 1368 is the exact opposite. The initial conditions, the assumptions, the sequence of events are all hidden away, we must just take their word on everything.
I wish we could bring back Bethe in his prime and give him the knowledge and the tools we have today to investigate criticality accidents in a full size IFR. I think he would say ‘forget the IFR, mass produce simple MSR’s for this century and design breeder MSR’s for the long run’.
QUESTIONS;
1… If a slow gravitationally induced, low reactivity rate accident in a 91 liter reactor yields 160 kg TNT equivalent, what would the yield be for a 6,000 liter core given the same reactivity profile?
2… What would the yield be for a 6,000 liter core given a high reactivity rate accident?
3… Why don’t proponents model a high reactivity rate accident?
4…Why don’t proponents model multiple criticality accidents or prove that they are impossible?
5… Do you agree that a higher reactivity rate of insertion will increase yield resulting in more prompt fatalities?
I’ve found answers to those questions for a specific metal alloy fuel design that I hope will reassure both the specific people who have expressed their concerns to me and anyone else who might have heard or read something about the issue from other people who have been unsuccessful in finding answers.
I’ve be aided in my quest by a number of experts who have done ground-breaking work in the field, including Yoon Chang, Dan Meneley, John Sackett, Neil Brown and Arthur Goldman. Though I gratefully acknowledge their assistance in providing answers and technical document references, this post is my own interpretation of the material, which is applicable to the alloy fuel used in the EBR-II as part of the Integral Fast Reactor project. If I get anything wrong, it’s my responsibility, not their’s.
The first part of the response is to remind Mr. Hannahan that Bethe’s 1956 vintage Hypothetical Core Disruptive Accidents calculation was based on an assumed fuel form, not a real one that had been refined based on years worth of fuel testing to be more resilient to various influences like overheating, excessive reactivity addition rates, and insufficient cooling.
The modern fuel form that was developed for the IFR included such innovations as sodium bonded metal alloy inside multiple tubes covered with a protective cladding. The sodium bonding provides good heat transfer paths in lightly used fuel, while being able to move to allow the fuel to expand inside the cladding as it grows under irradiation.
Each tube has a plenum above the fueled portion of the rods to allow space for fission product gases to accumulate and to provide a place for sodium to migrate as it is squeezed out by the expanding fuel alloy. In addition to the refined, resilient design of the fuel pins, the core is protected against seismic compaction by both a heavy structure lower grid plate and two sets of seismic load pads along the length of the hexagonal ducts that constrain radial tube movement during seismic events.
The combination of those features is called the core restraint system; it physically keeps fuel rod spaced apart to limit the maximum reactivity insertion under even beyond design basis events.
Of course, those design features cannot just be assumed to work if you want to actually build such devices in the United States. They had to be tested in ways that were adequate to convince both skeptical, questioning-attitude designers and regulators.
A series of tests were performed in a facility called TREAT (Transient Reactor Test Facility) between 1984 and 1987. These tests were carefully designed to determine the behavior of the proposed fast reactor fuel forms in the event of severe overpower conditions caused by rapid reactivity insertion. They demonstrated consistent behavior that would not be altered by different methods of inserting reactivity. (Ref: Behavior of Modern Metallic Fuel in TREAT Transient Overpower Tests – 4 MB PDF.)
The tests were performed with an exponential power increase at an 8 second period. (A reactor period is the time it takes for power to increase by a factor of e – roughly 2.72). Destructive analysis of the fuel rods was performed to determine their physical responses to the overpower condition and their mode of failure.
The tested rods withstood power that was about 4 times the operational maximum before they failed. Before reaching a point where the cladding failed, the rods expanded in the axial direction with fuel moving towards the plenum that held fission product gases. As the fuel began to melt inside the cladding tubes, fission product gases that had been dispersed in the fuel matrix pushed the fuel inside the cladding in an extrusion response that would be familiar to a cake decorator.
Fuel extrusion was greater than would have been caused by thermal expansion. It inserted a large amount of negative reactivity, limiting the peak power achievable. The cladding failures occurred at the hottest point in the fuel rods – which was at the top of the fuel column or in the plenum just above the fuel column. The pressure of the molten fuel tended to squeeze the fuel into the flowing sodium, resulting in even more dispersal and more negative reactivity insertion.
All of the fuel movements that resulted from the overpower experiments moved fuel outward from the reactor center, reducing reactivity and limiting peak power. None of them moved fuel closer together.
Here is a quote from Chuck Till and Yoon Chang’s Plentiful Energy, an excellent resource for people who are interested in a summary of the lessons learned during the 20-year-long IFR research and development effort. Plentiful Energy is also available in PDF format.
The TREAT tests demonstrated this further dispersal when fuel is molten. Both the pre-failure extrusion and the post-failure dispersal found in these tests provide intrinsic large negative reactivity feedbacks. They terminate overpower transients no matter what their cause.
(Emphasis added.)
So, going back to the questions that Bill Hannahan asked:
1… If a slow gravitationally induced, low reactivity rate accident in a 91 liter reactor yields 160 kg TNT equivalent, what would the yield be for a 6,000 liter core given the same reactivity profile?
2… What would the yield be for a 6,000 liter core given a high reactivity rate accident?
ANSWER: The reactor core design assumed for that hypothetical event bears no resemblance to a modern fast spectrum reactor. There is no yield, even in the event of a rapid reactivity insertion. It’s possible to experience core damage, but the fission products are retained by the coolant, the coolant container, and the containment.
3… Why don’t proponents model a high reactivity rate accident?
ANSWER: They do, and they base that model on physical test results from a series of carefully designed experiments.
4…Why don’t proponents model multiple criticality accidents or prove that they are impossible?
5… Do you agree that a higher reactivity rate of insertion will increase yield resulting in more prompt fatalities?
ANSWER: The test results show that overpower conditions are limited by strong negative reactivity feedback. They disperse fuel; they do not concentrate it as is done in an implosion device. There is no yield and there are no prompt fatalities. In fact, there are also no delayed fatalities; fission products are retained in the vessel.
The public can be adequately protected from radioactive material releases from fast spectrum reactors even in the event of a rapid reactivity insertion. Those results shouldn’t terrify anyone; they should result in celebratory rejoicing.
Physical testing at the IFR in April 1986 — mere weeks before Chernobyl — also demonstrated passive safety in the event of a loss of cooling flow without a reactor scram and in the event of a loss of all electrical power. It is demonstrably possible to design and build safe fast spectrum reactors.
I’m not in favor of a crash program to build them, but I strongly support implementing policies that enable their near term deployment, including favorable investment tax credits, accelerated depreciation appropriate for early generation machinery, and dedicated regulatory resources that can provide effective and efficient oversight of this specialized branch of nuclear reactor technology.
Thanks for the research, Rod
I recently received my ANS simulation fuel pellet, and it brought to mind a question: how does it compare to the alloy fuel slugs designed for the IFR? I suppose they would look rather metallic, but are they bigger? With the concave ends? I can’t seem to find an image of one.
It’s nice that we’ve found a way to build sodium fast reactors in a manner which prevents them from blowing sky-high. Problems that once could involve the evacuation of cities now will only involve the loss of plant, and maybe not even a total loss at that, potentially only requiring a decade worth of repairs.
Still, we have the other wonderful advantages that sodium cooling and fast spectrum brings to a reactor. Relative difficulty to get at the fuel or even see the core compared to the LWR, smaller distance between the beginning of delayed criticality and the beginning of prompt criticality, pyrophoric coolant, positive sodium void coefficient, potential for coolant flow blockage and resulting fuel failure, general twitchiness and inclination to have problems that result in years-long outages… what’s not to love?
It seems that like moths to a flame, large parts of the nuclear engineering community are inexorably drawn towards the SFR in spite of it’s well documented history of problems, failures, and decades-long maintenance periods. And like moths to a flame, once they manage to assemble some sort of new sodium reactor and, as usual, parts fail or the design is flawed and consequently the reactor has to spend 7 years offline in maintenance, these engineers get burnt.
Insanity is doing the same thing over and over again and expecting a different result.
@Anonymous Coward
Insanity is doing the same thing over and over again and expecting a different result.
True. However, sanity often involves reviewing history carefully, learning from it, and avoiding making the same mistakes that have plagued others in the past. When there are demonstrated right ways and wrong ways to design something like sodium fast reactors, sane people stubbornly choose the right way.
Please review the history of the EBR-II before you condemn the technology. Don’t keep pulling out tired stories of plants like Fermi 1, which experienced a fuel channel blockage more than 40 years ago as a result of a hastily designed core catcher inserted as an design afterthought to placate a regulator who didn’t really understand what he was doing.
Good article I read on the fly while trying to do it. Where else can I post a nuclear relevant though non-topic head’s up on a nuke blog??
Just now on the 7:30 am on Fox and Friend News then later CNN, Joshua Katz of Enright Group. former Homeland Security advisor, stated that foreign hackers HAVE “broken into” the _actual controls_ of nuclear power plants around the country. You pros know the facts BUT the public doesn’t — where it counts!
ANS & NEI, Time to step up to the PR plate!!
James Greenidge
Queens NY
Different reactor, I know, epithermal neutron spectrum, but here’s one rather spectacular example of what a rapid reactivity insertion can do.
The reports I linked there sound a bit like what you said (quoted? It’s not clear) above: The test results show that overpower conditions are limited by strong negative reactivity feedback. They disperse fuel; they do not concentrate it as is done in an implosion device.
The integral fast reactor was a documented history of success, until 1994, when anti-nuclear Democrats cancelled all US fission reactor research.
This “news” did this make the news – not even the anti-nuke news such as “Democracy Now”. I imagine Joshua Katz knows nothing of nuclear plant control systems. Why would you trust him on this?
Rod, Sandia has published a 2 part study on fires in Liquid sodium reactors.
[PDF]Metal Fire Implications for Advanced Reactors, Part 2 …
prod.sandia.gov/techlib/access-control.cgi/2008/086855.pdf
SANDIA REPORT SAND2008-6855 Unlimited Release Printed September 2008 Metal Fire Implications for Advanced Reactors, Part 2: PIRT Results Tara J. Olivier, Thomas K …
[PDF]Metal Fire Implications for Advanced Reactors, Part 1 …
prod.sandia.gov/techlib/access-control.cgi/2007/076332.pdf
This report documents the results of the initial stage of the “Metal Fire Implications for Advanced Reactors … for Advanced Reactor Design Based on EBR-II …
I am not opposed to IFRs, but I think that we need to do two things, before nwe adopy fast reactor technology, 1. determine the safest reactor design consistant with reasonable codts.
2.Determine which fast reactor technology offers the greatest flexibility for use.
Mark, I haven’t seen anything current on the news either, so this might be a recent interview comment from him. But hey, he’s in the security business. There have been cyber attacks on S Korean reactors but I haven’t followed details. He may be stretching the 2003 Slammer worm attack at Davis Besse (well documented, just google it). It actually did get right into the plant and crashed the SPDS software and took down the monitoring device on the plant operating control panel. And it also killed the Plant Process Monitoring computer. No small feat, but neither of those is a “control device.” The folks in “that business” like to claim, if they can get in that far, they will eventually get to an analog control device. But it is one of the problems the NRC is concerned with when dealing with upgrading old analog control systems to digital. As I understand the DBNPP problem it is one of tradeoff between using a tight firewall on the plant systems network and allowing access to plant employees from external computers outside of the plant network. This worm got in through an external infected computer. But it did not affect an actual plant control device.
HI Mark.
Thanks for replying on this off-topic looking for a home. My sole point is not OUR trusting him but the MEDIA — vis-a-vis the public does — which FUD and frets effects nuclear policy our any new plants. My beef is that those nuke orgs already geared-up to beat back media FUD are just standing on the sidelines or else we’d seen a lot more media pro-nuke stuff! Heck, here we’ve Indian Point and “nearby” Oyster Creek under anti attack via media clips as well but crickets are doing nuclear PR, the lessons of the other nearby Yankee plants likewise gone to the wind.
James Greenidge
Queens NY
Anonymous above & Mr. Hannahan seem to be strong partisans of the MSR, and this influences their criticism of SFRs. Problem is MSR/LiFTR designs are at least a decade behind the IFR-SFR, specifically the GE S-PRISM which could be made available as standardized commercial turn-key units by 2020; not so the LiFTR.
I too believe the LiFTR will eventually supply the bulk of the world’s non-hydrocarbon primary energy, possibly as soon as 2050. In the meantime we’ll have accumulated ~2000 tons of LWR bred actinides worldwide which could fuel ~200GWe of mid-sized first generation SFRs. A French proposal calls for utilizing the SFR’s surplus neutron spectrum to breed U-233 which in-turn would allow a rapid build-out of thorium LiFTRs. The LiFTRs high neutron economy allows for a rapid build-out provided you have a supply of U-233.
The operating theory of the SFR & LiFTR are complimentary in other ways, e.g. optimized SFRs could have standing-wave in situ breeding “deep burn” fuel cycles of several decades with periodic fuel suffling, possibly lasting the life of the reactor. In contrast an optimized LiFTR’s fuel is constantly being processed and circulated in and out of the core. By the time any of these designs are actually constructed highly efficient SC-CO2 PCUs should be available.
Cheryl,
Is your point that it is possible to have an explosion if you put too much power to a specific reactor design? We know that. There were several designs that were driven to failure in the 1960’s as a part of the testing process. Great reading. I am not sure what that has to do with the engineering of the IFR and the possibility of that design exploding like the one in your interesting article.
You know, that’s the funny thing about the differences between reactivity excursions in thermal reactors and fast reactors:
Things happen a lot faster in a fast spectrum reactor.
Specifically, if we look at experimental results of power pulse profiles in each case, we see that generally the time scale for fast neutron bursts is plotted in microseconds, while thermal neutron bursts are plotted on a scale of milliseconds.
Why is that important ?
Because the total energy released by the pulse is the integral or area under the pulse profile curve.
In simple terms, that energy or area is calculated as pulse height multiplied by “half-width” — meaning the the width of the pulse at half the peak power.
Since fast neutron pulses are so much shorter than those in a thermal neutron spectrum, the total energy release is very much less in the fast case, for the same reactivity insertion.
This simple fact explains, in large part, why the famous Kiwi-TNT (or “Transient Nuclear Test”) exploded in such spectacular fashion, while fast neutron bursts of much higher peak power in test assemblies like HPRR, Godiva, Jezebel, etc. merely warmed by a few hundred degrees.
Why is there such a large difference in pulse width between fast and thermal spectrum systems?
Because the thermal system relies on neutron moderation — for example the graphite in Kiwi-TNT.
That, combined with the slow transfer of heat from fuel to moderator, means that the pulse can proceed for quite long, before the moderator heats to the point where it no longer performs its function.
By contrast, in a fast reactor there is no reliance on a moderator, and the heating of fuel has virtually instant effect on reactivity (via Doppler and thermal expansion).
This by the way is nothing new. It has been studied and proven to death.
What is new is people failing to take advantage of the knowledge gained from a great deal of work done decades ago.
Oh, what misery !
Sure, if you really want near term how about the ESBWR? It’s finally just been NRC certified after about a decade-long delay. The Enhanced CANDU-6 could also readily breed U-233 for future MSRs from LWR UNF via DUPIC.
I was wandering through the internet and saw Rod’s post. Since I had recently written about that particular example of rapid reactivity insertion, it caught my eye.
I think that Rod’s purpose is to show that IFRs are safe, even with rapid reactivity insertion, but it’s really hard to tell. We’ve got to do better than that, because someone googling “rapid reactivity insertion” will very likely come upon the KIWI-TNT event. Responding that “IFRs are different” doesn’t have much persuasive power.
Nor do the quotes, particularly the one I repeated. One of the things I found amusing as I wrote my post was the language in the two reports I linked. It was the kind of engineer-speak that was parodied as “A divergent instability developed. A search for the parts was initiated.” Also, one of those reports went out of its way to insist that the explosion was not nuclear, whereas the other (which I find more credible) noted both chemical and nuclear characteristics. I suspect the first of those reports was written out of a mindset like Ron’s when he wrote the top post – wanting to prove that the reactor was safe.
I’ve thought about that KIWI-TNT experiment a lot over the years and knew some of the people involved. I never asked them about it, although I think the KIWI-TNT designation is a clue. If you’ve ever stood next to a Rover reactor mockup or looked at the numbers for the concentration of U-235 in the core, you have to wonder how much like a bomb it really is. So the logical next step is KIWI-TNT. Which it was possible to do at the time.
To go back to that engineer-speak: In the quote I repeated, doesn’t “They disperse fuel” indicate something like that KIWI explosion? It certainly sounds like it. And if that’s not what is meant, perhaps it should have been said differently.
An 8 second period is nothing, it is not even prompt critical.
Bethe calculated a reactivity insertion rate of 40 to 50 dollars per second at a gravitationally induced velocity of only 0.24 m/sec. One dollar is the reactivity required to get to prompt critical, the point beyond which delayed neutrons play no significant roll.
If the core were crushed by explosion at 1,000 m/s the reactivity rate would be 190,000 $/sec.
The title uses the words “rapid reactivity insertion” but none of your references actually looks at that condition.
Your unsupported conclusions actually contradict NUREG 1368 which predicts an accident worse than Chernobyl for a low reactivity rate gravitationally induced criticality accident.
India performed analyses and tests of worst case scenarios for their PFBR project, which is nearing completion of construction.
https://dl.dropboxusercontent.com/u/11686324/Indian_SFR_core_accident_test.jpg
Cheryl,
I agree that engineer speak is often confusing and covers over issues. Yes, people who are searching for that specific phrase would come across the KIWI-TNT experiment, but I doubt that any lay person would be researching that phrase. I would not have been, (I am very much a lay person in this regard). But your point that we need to make a clearer case is cogent. The audience for Rod’s article are technical people with the capacity to understand enough to see that their technical objections have good technical answers in the improved fuel form.
You said, “I think that Rod’s purpose is to show that IFRs are safe, even with rapid reactivity insertion, but it’s really hard to tell.” I am a bit confused about the sentence. Do you mean it’s hard to tell if IFR’s are safe with rapid reactivity insertion? Or on the other hand do you mean it’s hard to tell if Rod’s purpose is to show that?
From the Paragraph “I’ve found answers to those questions that I hope will reassure both the specific people who have expressed their concerns to me and anyone else who might have heard or read something about the issue from other people who have been unsuccessful in finding answers.”
It seems clear that Rod is working to show that the fuel design makes rapid insertion of reactivity safe.
Do you have actual concerns about the IFR safety or just the way that it’s actual very safe operation might be misunderstood by the public?
For comparison, the KIWI medium speed neutron reactor mentioned above was driven $8 beyond prompt critical at $300/sec.
http://www.tandfonline.com/doi/pdf/10.1080/18811248.1970.9734632
Bethe calculated his small fast neutron reactor exploded about 34 cents above prompt critical at an insertion rate of $45/sec.
The authors of NUREG 1368 keep their numbers secret.
On the other hand, experimental boiling water reactors running on slow neutrons have survived reactivity insertions more than $3 above prompt critical during chugging experiments.
David,
It’s hard to tell from what Rod wrote whether IFRs are safe. Or for me, anyway, and I think of myself as being just a little bit technical. Rod’s (or anyone else’s) assurance alone is not enough for me. I don’t have time right now to research those 1980s experiments.
The engineer-speak that Rod quotes is far from reassuring. The added reactivity is not a problem because it causes disassembly rather than assembly of a critical mass. That’s what happened in the KIWI-TNT experiment too.
Experiments from the 1980s are insufficient to demonstrate IFR safety in any generic way. People have made claims that previous experience was sufficient for safety in similar ways, and they have been wrong. I’ve been working on an article on one of those times, and I hope to see it in a national magazine soon.
As I say, a number of things I’m working on came together, I thought that readers here would enjoy the KIWI-TNT story, and I see a number of problems in this post and comments that nuclear advocates have made for themselves too many times.
Various experiments and analyses reported a simple mathematical relation (formula) for reliably estimating fast neutron fission burst yield as a function of peak reactivity, or multiplication factor ‘k’ – and NOTHING ELSE !
The formula can be adjusted to give yield results ‘Y’ in units of kilograms or tonnes of TNT equivalent explosive energy, and looks like this:
Y(tTNT) = 7.246E-27 * exp (4 * ln (k – 1) + 74)
Examples of a few calculation results are as follows.
Note that for Uranium the delayed neutron fraction is 0.0064, so “one dollar of reactivity” above “delayed critical” is equivalent to k = 1.0064 and $10 is k = 1.0640.
k.…………………Yield
1.001………0.001 kgTNT
1.002………0.016 kgTNT
1.005…..……0.62 kgTNT
1.007…….……2.4 kgTNT
1.01……………10 kgTNT
1.04….……..2550 kgTNT
1.1……….……..100 tTNT
1.2……………..1.6 ktTNT
1.3……………..8.1 ktTNT
1.4……………25.5 ktTNT
So the question is not whether an energetic fission burst is theoretically feasible, but rather what the maximum possible reactivity in any given design might be.
The rate of reactivity insertion is also important because, as illustrated by the examples above, it doesn’t take a lot of excess reactivity to disperse the fuel into a configuration which can not regain criticality again.
For example, fuel slumping as a result of overheat and melting is a relatively slow process, resulting in a low-energy fission burst from LESS than $1 of excess reactivity – i.e. below prompt critical.
Larger and faster reactivity insertions are unlikely in power reactors, for the simple reason that the core is maintained at very small excess reactivity.
The latter is a direct result of the fact that fast reactors are relatively insensitive to fission product poisoning (especially Xenon), compared to thermal spectrum reactors, which carry lots of excess reactivity (suppressed by boron) in order to override poison buildup during power swings.
Agreed, I would love to see hundreds of ESBWR’s and AP 1000’s and other designs that are well proven to be safe being built around the world.
@Cheryl Rofer
I make no apologies for engaging in “engineer speak.” The determination of whether or not a particular design — with detailed specifics about the way its particular structures, systems and components behave in both realistic and reasonably proposed hypothetical situations — is adequately safe is an engineering problem. It is not a science problem and not an emotional problem. There is no way to demonstrate safety in a generic way; that is why there is a detailed process of reviewing proposed reactor plant designs before they can be built and operated.
There is a major difference between “dispersal” and “disassembly.” The specific challenge that Bill Hannahan continues to pose is that there is some kind of mechanism that may result in core compaction that is sufficient to cause a fast reactor core to achieve something beyond prompt criticality (again with the engineer-speak.) He claims that this hypothetical event — the causes of which involve some wild speculation on the order of an asteroid strike — could cause an explosion with a yield of several hundred tons of TNT.
The testing performed in the 1980s demonstrated that modern (at that time) fast reactor fuel does not behave as Hannahan postulates. There is no mechanism that can cause the core to become prompt critical. If enough reactivity is inserted to cause a severe power excursion, the fuel expands in the axial direction (gets longer). It does not collapse toward the center. At the point when metal alloy fuel rods fail, the fuel material squeezes out of the top of the rods and gets dispersed into the sodium coolant. That spreads the fuel material out, cooling it and preventing and further fission reactions because there is no way to achieve a critical configuration where a self-sustaining reaction is possible.
To visualize this, think about the way a fire behaves when its fuel is spread out far enough and is forcibly cooled with something like water or AFFF so that it cools off and can no longer sustain combustion.
Sure, if a fast reactor experiences a criticality event that overheats the core material in this fashion, it will be damaged enough so that the installed core will have to be removed and the sodium will have to be purified or replaced. There will be no damage outside of the plant boundaries. It might even be possible to clean up the primary system enough so that the plant can be restarted.
A little known or understood fact, by the way, is that Fermi 1 was repaired and operated for two more years after its infamous partial core melt event.
PS – In the event that anyone proposes building sodium cooled fast reactors in the US, there will be more testing and review by competent regulators. There has already been some groundwork laid to provide the capability to test even though political action and budget challenges resulted in the purposeful destruction (not really, but that is the subject of a different post) of the Fast Flux Test Facility. https://www.inl.gov/article/futurix/
@Jaro
For example, fuel slumping as a result of overheat and melting is a relatively slow process, resulting in a low-energy fission burst from LESS than $1 of excess reactivity – i.e. below prompt critical.
And the TREAT testing of proposed fuel forms shows that “fuel slumping” is not the event that actually occurs as a result of overheat and melting. As the rods heat up, the fuel material expands in the axial direction. Once it gets hot enough to break through the cladding, it squeezes out of the top of each rod and gets dispersed into the sodium coolant. It does not magically melt into a more reactive configuration.
From Plentiful Energy page 156:
@Bill Hannahan
If the core were crushed by explosion at 1,000 m/s the reactivity rate would be 190,000 $/sec.
The only way I know of for explosions to crush something is to very carefully arrange lens-shaped charges with exquisitely timed detonators to make sure they all fire at the same time to create a shock wave that implodes the material inside of the shaped charges. In any other configuration, explosions tend to disperse material, not compact it.
BTW – Exactly what is your speciality? Have you studied explosives or weapons systems designs?
Cheryl,
So you are working on an article that shows how “People have made claims that previous experience was sufficient for safety in similar ways, and they have been wrong.” Why? Are you somehow convinced that good engineers and business people want to spend billions of dollars on a machine that will fail and ruin their reputation? You seem to have time to research other failures but Rod’s technical explanations do not seem to be enough for you. You don’t have time to look into the actual safety of the IFR.
It seems to me that you are not open to being convinced, or that you set the standard for Nuclear “safety” at a level much higher than for other energy sources. I don’t think anyone is saying the word “impossible” but “improbable” is adequate. I read Rod’s article and his reply to you as showing that the new fuel form adequately addresses your concerns and you have not given a technical reply that demonstrates he is wrong.
Hi Rod –
What you’re doing in this comment isn’t “engineer speak” in the same way that is illustrated in the two reports I linked in my post, nor in your quotes above from those 1980s reports. Those reports obfuscate what was observed. And it’s not necessary to do that to write a good article. Editors of scientific journals have praised my plain, but accurate writing. Simply using technical terms is not what I am referring to as “engineer speak.”
Hanrahan’s concern is a real one that was a big part of safety calculations for light water reactors back in the 1960s. I’m well aware of the desirability of a fuel “pancake.” Flow can be a real mystery – one of my colleagues used to cite thermocouples in reactor coolant tubing within several inches of each other that registered tens of degrees difference. That was in the bad old days, of course, without the instrumentation and modeling capabilities we have today.
Many calculations were done back then, and more need to be done now – glad you agree with me on that!
Your condescension is unnecessary. It’s a form of personal attack, so I won’t answer it and will be merrily on my way!
Since Cheryl has decided to depart this conversation, I need help from others. Can anyone guess from a careful reading of my comments what the basis was for this last line in her comment?
Your condescension is unnecessary. It’s a form of personal attack, so I won’t answer it and will be merrily on my way!
I really need the feedback. We all need critical reads of our work now and again.
@Bill Hannahan
Can you point to the section(s) in NUREG 1368 that make the prediction that you claim?
I just read through section 15.6.8 Hypothetical Core Disruption Accident (HCDA). The conclusion of that section does not support your interpretation. Though it indicates some uncertainty that could be reduced with further testing, the NRC’s Preapplication Safety Evaluation Report for the Power Reactor Innovative Small Module (PRISM) Liquid Metal Reactor (aka NUREG 1368) makes the following statement on page 15-26:
That is not a slam-dunk, everything-is-perfect assessment, but it sure does not indicate to me there are any fundamental concerns that would take IFR type fast reactors off of the list of available options for helping to solve many of the challenges that human society faces.
John GaULT, MY REVIEW OF IFR R&D RESEARTCH LITERATURE suggests that there were acknowledged safety issues, and the Sandia repost that I pointed to offered an even stronger case for safety issues. Is the reason GE does noy nuild a cxomercoal prototype of the S-prisum simply a commercial decision, or is it due to a lack or “ready for prime time technology?”
This may be a bit snarky, but.. what if a hydroelectric dam were crushed by explosion at 1,000 m/s? Or a coal-fired boiler? Or a natural gas-fired turbine? Or a wind mill? Or combined solar power or photoelectric panels??
JohnGalt:
What exactly are you implying here? Merely that more uncertainty exists with 2nd generation SFRs than 3rd or 4th generation LWRs; or that SFRs are actually inherently more dangerous than, say, NG fired power plants and their assoc. pipelines & storage, which happen to explode all the time (and which seem to have relatively few regulatory difficulties)? Or that SFRs are somehow even more dangerous than the routine risks of mining and atmospheric fallout (emissions) of burning pulverized coal?
Here’s the conclusion of the authors of NUREG-1368 “On the basis of the review
performed, the staff, with the ACRS [Advisory Committee on Reactor Safeguards] in agreement, concludes that no obvious impediments to licensing the PRISM design have been identified.”
Hello Jaro,
Interesting equation, do you have a reference that explains its derivation?
Just to be clear, are you saying that a 60,000 kg core with k=1.01 will produce the same yield as a 60 kg core with k=1.01?
It seems to me that the larger core would require a much larger dimensional change to become subcritical, requiring much more energy to move the larger mass over a greater distance in the same amount of time or even over a longer time.
You wrote; “Larger and faster reactivity insertions are unlikely in power reactors, for the simple reason that the core is maintained at very small excess reactivity.”
That is one of the great advantages of an MSR, especially a fast neutron MSR.
A Solid Fuel Fast Neutron Reactor (SFFNR) contains a coolant, usually sodium, that is a neutron poison, like water in an RBMK 1,000 design. I have seen estimates of the coolant worth ranging from -$5 to -$17, depending on the details of the design.
Just removing the sodium, without moving the fuel, would put it deep into the super prompt region. That would leave a core with a large volume of empty space in it, a high leakage configuration. Crushing the core into a solid metal mass would add a great deal more reactivity.
In a comment Rod refused to publish, you would have learned that Bethe calculated that his reactor would explode after collapsing only 2.91 cm under the force of gravity, with a maximum velocity of 0.24 m/sec, $45 / sec. It is not necessary to crush all of the sodium out of a SFFNR to get a high energy criticality.
In a fast neutron MSR, fuel and coolant atoms are mixed at the atomic scale. I cannot think of any way to rapidly eject coolant atoms while leaving fuel behind.
So all of the advantages of a SFFNR are possible in a fast neutron MSR without the criticality risk.
Rod,
The closest thing I see to “condescension” is this part:
“To visualize this, think about the way a fire behaves when its fuel is spread out far enough and is forcibly cooled with something like water or AFFF so that it cools off and can no longer sustain combustion.”
It is a great analogy, but maybe Cheryl thought you were “talking down” to her?
I admit that I’m reaching a bit — but that’s the best I can come up with.
SECOND TRY
Rod wrote;
“Can you point to the section(s) in NUREG 1368 that make the prediction that you claim?”
https://atomicinsights.com/kirk-sorensen-why-didnt-molten-salt-thorium-reactors-succeed-the-first-time/#comment-15142
Chernobyl did not come close to producing any offsite prompt fatalities, and the firemen would not have died had they been used carefully. See #4 below.
https://atomicinsights.com/russia-continues-sustained-fast-breeder-reactor-effort/#comment-97675
Random quotes from a > 400 page document are not helpful. Page numbers please.
My replies to your questions:
Q: “Interesting equation, do you have a reference that explains its derivation?”
A: The equation is an empirical one, fitted to the results of many experiments on various fast spectrum critical assemblies.
You can find these (including the formula, with minor changes in the constants) in publications such as Nuclear Science and Engineering, mostly in the 1960’s and 1970’s.
There is very little deviation in results between different assemblies. Deviations only become significant at very high yields, when a significant fraction of the fuel is consumed (fissioned) in a pulse. Obviously the equation does not apply at all when total yield exceeds the potential in a given mass of fissile material.
Similar equation formulations have been attempted for thermal spectrum assemblies, but the variability there is large, so no single equation is applicable.
The only generic observation is that thermal spectrum pulse widths are about three orders of magnitude wider, relative to pulse height, than fast spectrum ones (i.e. milliseconds versus microseconds).
Q: “Just to be clear, are you saying that a 60,000 kg core with k=1.01 will produce the same yield as a 60 kg core with k=1.01?”
A: For a similar ratio of neutron mean free path to core diameter, yes. Definitely. In other words, it is not the mass that counts, but rather the density (which sets the neutron mean free path).
In a high density core – such as metal – the mean free path will be a few inches to a few feet, again depending on material & lattice geometry specifics, which determine overall density, as seen by neutrons.
For your 60-tonne core, a metallic version will likely be many times larger than the neutron mean free path, so a reactivity excursion will not involve the entire core, only a part of it. The same would be true for a 60,000,000,000 kg core measuring a mile across.
In theory, you could compress entire 60-tonne core, such that a large number of simultaneous reactivity excursions occur throughout. Not sure if you’ve answered Rod’s question about that one.
Q: “A Solid Fuel Fast Neutron Reactor (SFFNR) contains a coolant, usually sodium, that is a neutron poison, like water in an RBMK 1,000 design. I have seen estimates of the coolant worth ranging from -$5 to -$17, depending on the details of the design. Just removing the sodium, without moving the fuel, would put it deep into the super prompt region. That would leave a core with a large volume of empty space in it, a high leakage configuration. Crushing the core into a solid metal mass would add a great deal more reactivity.”
A: There’s that “Crushing the core” business again. Are we talking asteroid impacts or what ?
As for reactor lattices containing sodium coolant, a few things to consider.
First of all, sodium (Na-23) has a remarkably low neutron absorption x-section in the fast spectrum – on the order of 0.0003 barns, compared to about 0.5 barns for hydrogen (water) in a thermal spectrum like that of RBMK.
For comparison, the U238 in LEU (used in both thermal and fast reactors) has an absorption x-section of 0.1 barns in the fast spectrum and about 5 barns in thermal spectrum.
So if you take the ratios of absorption x-sections s(U238)/s(H1) and s(U238)/s(Na23) in thermal and fast spectra respectively (ie. 10 and 333), you see that hydrogen is about 30 times more of a poison in a thermal reactor than sodium is in a fast reactor.
As you say, estimates of the coolant worth depend on the details of the design.
Not surprisingly, modern SFR designs aim to minimize the consequences of sodium voids due to boiling or other mishaps.
Typically, the active region of these cores have a low aspect ratio (i.e. a very short cylinder, almost a pancake), which results in sodium voiding having either no, or even a negative reactivity effect.
That is difficult to arrange, given the low neutron absorption of sodium on the one hand, and the increased neutron leakage out of the core as the coolant is voided, on the other. The effect may be augmented, if desired, by spiking the portions of fuel assemblies above and below the active portion with neutron absorbing material like boron.
Cheryl,
I think that I may have been condescending in my comment to you. Sorry for that, I said,
“It seems to me that you are not open to being convinced, or that you set the standard for Nuclear “safety” at a level much higher than for other energy sources.”
Which was a judgement about your intentions. I assumed them rather than asking you about them.
@Jaro
You wrote:
In a high density core – such as metal – the mean free path will be a few inches to a few feet, again depending on material & lattice geometry specifics, which determine overall density, as seen by neutrons.
I’m pretty sure that you understand the situation, but that specific phrase might give some people the wrong impression.
Metal cores don’t have to have a high density of fissile material, even if they use plutonium or highly enriched uranium. In the fuel designed and used at the IFR, the fissile material is alloyed with other metals. Here is a quote from page 40 of Plentiful Energy.
From page 44
From page 104
Even if it offends some scientists when I carefully distinguish between the fields of engineering and science and express my strong preference for the first over the latter, this is an example of where the application of scientific understanding (aka engineering) demonstrates its importance in the real world of building plants that can operate safely and reliably to produce a useful product.
A few days have passed and the media have shrugged off the massive explosion and fire of that oil-carrying train in Pennsylvania that required extensive evacuation of nearby towns and residences. Can you imagine the media reaction if it were a train carrying new or used uranium fuel, which in an accident would likely result in the result of…nothing? No raucous meetings of regulatory officials on that one, but the agitators were tearing the place apart in a recent meeting for Vermont Yankee’s PSDAR.
Rod – There is no basis. All I can say is that … well … it’s Cheryl. I’ve been reading her stuff for years, and I’ve come to realize that she has some … er … personal issues.
I base what I say on a careful reading of her comments over the years. This is a very Cheryl-like thing to do. I guess she gets some kind of perverse pleasure from throwing people off-guard — like tossing a hand grenade and running.
As we can see here, the one thing that she won’t do is stick around to answer challenges to anything she writes. When she is challenged, she just doesn’t “have time right now to research” anything on the topic or she complains about stupid stuff like “engineer-speak” or she just up and leaves. She prefers to live in her own little world, which is why she has her own little website. I suggest you let her live there, and simply ignore her when she is here. Much of what she writes is completely incoherent or just plain wrong — as is immediately obvious to anyone with any sense — but she is simply not interested in anything that doesn’t support her own small-world view. I speak from experience.
If you disagree with her, it’s a “personal attack.” If you try to introduce relevant information, you’re being “condescending.” These are her problems. Don’t let them become yours.
ROD wrote “The only way I know of for explosions to crush something is to very carefully arrange lens-shaped charges with exquisitely timed detonators… In any other configuration, explosions tend to disperse material, not compact it.”
Rod, I agree if we are talking about something small like a bomb core or the football sized core of EBR-1.
A large core, like Superphenix, contains hundreds of critical masses and is loosely connected neutronically, as Jaro explained, allowing for multiple independent critical masses within the core volume.
Imagine you want to “disperse” a large 1 GWe solid fuel fast neutron power reactor core using high explosives, where would you put the charge?
Dead center was my first choice. Visualize the sequence in ultra slow motion.
The HE slowly converts itself into a bubble of superheated gas at extreme pressure. The initial pressure is many times the material strength of the fuel rods, spacers, fins and structural components of the core. Initially those things behave more like a liquid than a solid. Inertial forces dominate.
The sodium and fuel rods rapidly accelerate outward. The sodium accelerates faster due to it’s lower density creating an almost instantaneous sodium free void growing at the center of the core. At this point, it might produce a high energy criticality just due to rapid sodium voiding.
If not, the inner rods slam into and meld with rods farther out to form a rapidly expanding spherical shell of solid fuel. The shell thickness is rapidly increasing as layer upon layer of rods are added to the outer surface. It grows until the thickness is sufficient to support multiple criticalities throughout the shell. Due to the high velocity, high reactivity addition rate, each of these criticalities is more energetic than a single slow gravitationally induced criticality.
If this happens at a fairly small radius, the energy from the criticalities is added to the energy of the explosive, increasing the expansion velocity, resulting in the formation of a second shell expanding through a larger radius, at much higher velocity, much higher reactivity rate. The second shell rapidly builds thickness resulting in another round of multiple criticalities throughout the shell, each more energetic than those in the previous round.
You could put the explosive along side the core. That would produce a rapidly thickening hemisphere of solid metal into the side of the reactor. That makes the full diameter of the core available for chain reaction criticalities, which may yield more energy than a central explosion.
You could put the explosive on the top or bottom of the core. That would crush the core several orders of magnitude faster than the gravitational meltdowns that Bethe and NUREG 1368 analyzed, resulting in a much more energetic criticality.
Feel free to run this experiment with a water moderated core, light or heavy water. The result will be dispersal without criticality.
Feel free to run this experiment on a Chernobyl core. The result will likely be a medium energy criticality with yield limited due to the long neutron lifetime, similar to the actual accident.
I do not claim that SFFNR’s are unsafe, I claim that their proponents have not proven that they are safe.
Have them model a high velocity core crush, better yet, crush a real core at high velocity. If the result can be reliably contained I would be satisfied.
@Bill Hannahan
Imagine you want to “disperse” a large 1 GWe solid fuel fast neutron power reactor core using high explosives, where would you put the charge?
You have obviously ASSUMED away all of the layers of security and protection that prevent anyone from getting anywhere near any nuclear reactor core, including the self protective force field that the reactor creates for itself with just a few days worth of high power operation.
I have no interest in wild hypotheticals when there are so many real problems to solve.
The dead-center explosive event sounds very much like the problem faced by Teller and Ulam in trying to come up with a sustainable fusion reaction in the secondary of a thermonuclear device. Any kind of explosive process results in an expanding reaction front wherein radiation (neutrons in this case) leak away faster than the reaction process that is producing them, resulting in extinguishing of the reaction.
The sideways or top-bottom process helps mitigate this effect somewhat, but setting that up would take a lot more time and effort than simply setting a bomb in the center of the core. You’d likely have to have multiple explosive events occurring with a high degree of simultaneity, probably set up in layers with tampers and the like to “push” the materials together in the right way. Unless its a freshly loaded core at a reactor with no power history, it’s hard to imagine someone being able to rig something like this up. It would take time and working in the radiation fields near a core with any kind of power history would make it problematic. It’d be easier just to build a plutonium bomb using WGP that someone would sell you.
Jaro, consider a thought experiment.
Imagine a 60 kg reactor and a 600 kg reactor running as a steady 5 watts. Both have the same neutron lifetime. Increase k to 1.01 and plot power vs time.
Both reactors are deep in the superprompt region. The power curves overlap until heating begins to provide negative feedback. Since things are changing so fast, I would guess power is in the 10’s to 100’s of megawatts at that point.
For each degree of temperature rise in the big reactor, the small reactor gains 100 degrees. The power curve for the small reactor begins bending to the right from temperature feedback while the other continues climbing rapidly. As k drops through 1 in the small reactor, its power peaks and begins dropping. The big reactor peaks later and higher than the first reactor. The total energy produced by the big reactor is much greater than for the small one.
If you said the energy DENSITY, joules/gram, is the same, that would seem reasonable. Otherwise I suspect there are some assumptions or limitations that prevent your equation from applying to this case.
What do you think?
Correction a sentence in my las comment should read;
“For each degree of temperature rise in the big reactor, the small reactor gains 10 degrees.”
Bill Hannahan wrote that “You could put the explosive on the top or bottom of the core. That would crush the core….”
Let’s take a look at how practical that idea is.
First of all, the top part fast reactors is typically enclosed by an upper containment vessel, with maintenance ports bolted shut during normal operation.
No way to access the top deck above the reactor without going through the procedure of unbolting one of the maintenance ports.
https://db.tt/hNOb1w74
https://db.tt/qMxeFciu
Next, the “top of the reactor” is actually not the top of the active part of the reactor, merely the top parts of assemblies, of which the active portion is the lower part:
https://db.tt/SNBLam9q
https://db.tt/ngGSfMDw
Moreover, the top parts of the assemblies, below the reactor top deck, do NOT have openings through which one can simply drop explosives.
Rather, the top of the assemblies comprise grappling appendages, to facilitate removal and replacement.
During maintenance or refuelling, one or more of these assemblies might be removed, leaving a hole into which things might be dropped.
But during these activities the reactor is shut down, with all absorber rods inserted, preventing any sort of reactivity excursion.
https://db.tt/tRFaqK3Q
From your comments, readers might be left with the impression that anyone could simply toss a stick of dynamite in the general direction of a reactor and cause a disaster.
Personally I don’t see how you can justify that as anything more than cheap scare mongering.
Warning – this thread closes today at about 8:00 pm.
@Jaro
One more item to add to your list – in an IFR style reactor like the PRISM, the core is at the bottom of a rather deep pool of sodium.
http://nextbigfuture.com/2011/12/overview-and-status-of-smrs-being.html
Even if you figure out how to open up the top of the reactor vessel with some kind of magical shaped charge, several tens of feet of liquid sodium would appear to be a pretty substantial shock absorber.
Rod, your title asks;
“How do metal alloy fuel fast reactors respond to rapid reactivity insertion events?”
Your essay refers to a study that subjected up to three fuel pins to a power increase on an 8 second period. While 8 seconds is short for a power reactor, it does not take much reactivity or a high reactivity insertion rate to get to 8 seconds, well below prompt critical.
Based on this study, the essay generalizes the result to conclude that all criticality accidents are safe or impossible;
” The public can be adequately protected from radioactive material releases from fast spectrum reactors even in the event of a rapid reactivity insertion.”
But the study does not prove that at all, it is not a high rate criticality accident study. It does not answer the question posed in the title, or the questions in the essay itself.
The study shows that melted fuel will be swept away by coolant flow. What happens next? Uranium and plutonium are far more dense than sodium. When it exits the core the flow area expands and the velocity drops. Does the fuel fall out on top of the core plate, is it still liquid or resolidified, does it pile up into a critical mass, what happens to that material if the pumps stop?
What happens if fuel melts during station blackout and pumps are not running?
@Bill Hannahan
The reactivity insertion rate used was high enough to demonstrate how the fuel behaves when power increases rapidly. Your postulates of external events that you believe might have the possibility of inserting an extraordinary amount of reactivity in a short period of time are wildly unrealistic and do not demonstrate any understanding of how physical systems behave. I ask again – what is your area of expertise based on education, training and experience?
You’re getting a bit tiresome. However, I’ll address your last question.
What happens if fuel melts during station blackout and pumps are not running?
Apparently you have forgotten the results of the full scale test of the IFR that showed that a station blackout that stops the pumps from running will not result in damage to the core, even if the event happens and neutron absorbing rods are not inserted.
http://www.ne.anl.gov/About/hn/logos-winter02-psr.shtml
Here are the NUREG 1368 page numbers [ and PDF#’s] for the associated quotes.
http://pbadupws.nrc.gov/docs/ML0634/ML063410561.pdf
A-26 [378]….“Regarding accommodation of HCDAs, there is not sufficient data to confidently predict the size of an HCDA in a metal fuel ALMR.”
A-18 [370]…“The major contributors to core melt all lead to energetic core disassembly accidents and Release Category R4A.”
A-17 [369]…“For the R4A no-evacuation case, prompt fatalities were shown to increase from 7 to
124, and latent fatalities increased from 1,520 to 3,320.
15-17 [329]…“The PRISM design has been described as passively safe. On this basis, the designers contend that core melt and sodium boiling do not have to be considered in the design”
In graduate school I asked why a full meltdown was not a design basis accident. My instructors said the PRA’s proved that a meltdown was so unlikely that I would never see one in my lifetime.
Had a meltdown been a design basis accident, the Fukushima reactors would have had a core catcher, hydrogen recombiners and a meltdown rated vent filter. And we would be in the midst of a nuclear rebuilding campaign.
I believe that rummors of a Tooth Fairy are more likely to be proven true than the optimistic risk probabilities in NUREG 1368.
Jaro wrote;
“…readers might be left with the impression that anyone could simply toss a stick of dynamite in the general direction of a reactor and cause a disaster.
Good point, I agree it would be very difficult. I was giving an example in response to Rod’s implication that ordinary explosives cannot be used to cause a criticality.
My guess as to how a criticality might happen, in order of probability, high to low;
1… Meltdown leads to medium energy gravitationally induced criticality explosion worse than Chernobyl, as described in NUREG 1368.
2…Partial meltdown leads to medium energy gravitationally induced criticality explosion which crushes air and sodium out of a large section of unmelted core at high velocity, resulting in a high energy criticality that ejects the entire core, and perhaps a nearby spent fuel pool into the atmosphere.
3… Unknown mechanism. History has shown that the PRA’s of the 1970’s were wildly optimistic because the authors could not imagine all of the paths leading to an accident, eg. operators bypass multiple reactor protection systems.
The consequences of a high energy criticality in a Solid Fuel Fast Neutron Reactor (SFFNR) are so severe that they should not be built just to avoid the risk from unknown mechanisms, even if all other scenarios can be eliminated.
In the 70’s we had far more experience with light water reactors than we have now with SFFNR’s, yet we greatly underestimated the meltdown rate.
4… A disgruntled, smart, employee with access to vital areas of the plant;
4a……identifies a vulnerability in the control system and causes a max speed withdrawal of all control rods, that cannot be overridden from the control room.
4b…… Secretes a small explosive package into a new fuel assembly that explodes in the core during refueling or startup, rapidly crushing the fuel, ejecting sodium and concentrating the fissile fuel at a velocity orders of magnitude greater than 0.24 m/sec assumed by Bethe in his analysis of a small reactor.
4c…… Pulls several spent fuel assemblies out of their racks in the spent fuel pool, stacks them up on top of the racks, and crushes them with an explosive device or heavy weight dropped from considerable height.
4d…… Arranges new fuel assemblies into a near critical array in air and crushes them with a heavy weight falling through a large distance or using explosives, or by dropping half the assemblies on the other half from a large height, resulting in a much higher velocity, and yield, than in the Bethe calculation predicted.
5…Terrorism.
5a……A group of terrorists, perhaps with inside help, overwhelms a plant and implements one of the above scenarios.
5b……Aging Jumbo jet containing a massive tungsten penetrator or a long propeller drive shaft from a scrapped supertanker, packed with tungsten, impacts at over 500 mph and crushes the reactor core at high velocity.
Actually, a fuel fabrication plant, or a truck or train carrying new fuel assemblys would probably be an easier target.