Integrating six decades of learning about fast reactors
I learned some important new concepts yesterday from two of the leaders of the Integral Fast Reactor (IFR) project – John Sackett and Yoon Chang.
Among other things, they informed me — as a member of a group of about 35 other attendees at a workshop titled Sustainable Nuclear Energy for the Future: Improving Safety, Economics, and Waste Management organized by the Global American Business Initiative — that the “Integral” part of their project’s name refers to the fact that the IFR creators were aiming to produce a highly evolved system that integrated lessons from a number of separate fast reactor learning experiences.
It also meant that the leaders believed an important part of their project’s success was creating a situation where all of the disciplines required for a complete reactor power plant system were in one place where their special knowledge could be integrated with that of other specialists to produce the best possible system concepts.
Though I’ve written and thought quite a bit about the IFR project over the years, I had a mistaken impression about the reason that “Integral” was chosen to be part of the project’s name.
Because the project combined the Experimental Breeder Reactor 2 (EBR-2) with a closely associated fuel recycling facility that used pyroprocessing to produce new metallic fuel elements, I thought that ‘Integral’ meant that the project leaders envisioned that each IFR installation would include both reactors and a fuel recycling facility.
That never made economic sense to me; it would substantially increase the initial capital cost and eliminate some economies of scale that would accrue if recycling was done at a specialized, regional facility for a large number of power plants. It also seemed to inherently limit the potential market reach of the system; many potential customers for a safe, reliable nuclear power plant would not want — or be allowed — to get into the fuel recycling business.
Now that I have a better understanding of the ‘I’ in IFR, I’m a stronger fan of the concept and of the various system design iterations that fall under the umbrella of Integral Fast Reactors.
There are several key choices that make the IFR different from other fast reactors that have met with mixed success or outright failure. These choices were made as a result of a focused effort to apply lessons learned, something that happens more quickly and permanently as a result of evaluating a failure. Systematic learning can be inhibited in a situation where moderate successes more firmly establishes a path that has inherent limitations.
An IFR includes the following:
- Metal alloy fuel vice oxide fuel
- Fuel element design that provides space for gaseous fission products to accumulate without damaging cladding
- Low/no pressure sodium coolant
- Pool vice loop for sodium coolant
- Inert gas blanket
- Double walled tank to hold the coolant and provide leak detection in inert environment
- Double walled steam generator tubes
Within those basic choices, there are a wide variety of iterations that can provide specific solutions to customer needs.
It is important to recognize that the IFR design choices are not just conceptual. They were proven through 30 years worth of experience with an operating [not paper] power plant (EBR-2) and pilot scale fuel recycling facility. The system was reliable, experienced few sodium related challenges, demonstrated passive safety through a well planned series of physical experiments, and produced low worker radiation exposures. That last advantage was a result of the virtually non-existent corrosion of internal surfaces even after 30 years in a hot, sodium-bathed environment.
As Sackett and Chang informed the workshop, the commonly held perception of sodium as being a difficult and dangerous coolant has been proven wrong by experience. Despite the fact that sodium reacts violently when exposed to water or air, no one has ever been injured as a result of sodium leaks. All instances of sodium and water interactions in steam generators have been readily contained and all instances of sodium leaks from piping have been mitigated by standard response processes made easier by the fact that there is no pressure forcing the material out of piping or tanks. If there is a leak, it is a drip or a steady stream, not a gusher.
As I listened, I could not help but compare the experiences Sackett and Change described of working with low pressure sodium to the experience of working with high pressure, “live steam.” Even though water is not normally thought of as explosive, steam explosions were the cause of numerous fatalities in the era before the American Society of Mechanical Engineers and the Hartford Steam Boiler Inspection and Insurance Company joined forces to develop pressure vessel codes and standards.
Even in recent years, after we have had 150 years to become quite skilled at producing high quality piping, valves and pressure vessels, there are instances where people are severely injured or killed by accidental exposure to live steam. Pipes that start as high quality, high integrity components can deteriorate as a result of corrosion or erosion, and valves can either fail or be mispositioned.
One of my former shipmates had the life changing experience of being the commanding officer of a ship that experienced a steam line rupture. Unfortunately, some of the sailors involved experienced a life-ending experience.
We may all be more comfortable with water than with liquid sodium, but power plants don’t use benign, well-behaved forms of H2O, they use high temperature, high pressure forms that are at least as hazardous as hot, low pressure sodium.
Sackett pointed out a maintenance advantage to sodium that I had never thought much about. Since sodium freezes at 98 degrees C, maintainers can easily create a freeze plug to isolate a valve or a pipe section when the plant is shutdown for maintenance.
From my water-cooled reactor experience, I’m familiar with using freeze plugs, but they are not easy or cheap to create or maintain. They require a continuos supply of refrigerant to keep the water well below room temperature. In a sodium reactor, freezing happens naturally as long as there is no effort to add the heat required to maintain sodium well above room temperature.
In the 21 years since President Clinton and current Secretary of State John F. Kerry (who was then a Senator) joined forces to kill the IFR project, creative scientists, engineers and administrators have managed to continue to develop and prove out some of the planned innovations — especially in fuel recycling — that had not yet been completed. People who recognized the unique value of the IFR have also continued to refine their designs, publish papers, publish books [ex: Plentiful Energy and Prescription for the Planet] and give talks around the world to increase understanding of the potential for nearly infinitely sustainable nuclear energy.
GE-Hitachi’s PRISM reactor is one of the more well known commercial variations on the IFR concept, but Terrapower’s current iteration of the traveling wave reactor seems to qualify.
Another intriguing variation is the ARC-100, a small, simple, long-fuel life (20 years between refueling) version that I first learned about when I was preparing to retire from the Navy in 2009-2010. I plan to learn more about its current status and the company’s development plans in the coming weeks.
Despite the impression that the above photos might provide, yesterday’s audience was fairly diverse and included a number of people young enough to make IFRs a reality. I’m more optimistic about our future energy choices today than I have been for quite a while. I’m looking forward to the next GABI workshop and want to express my appreciation for their continuing efforts to provide excellent learning opportunities that make it worthwhile to drive to DC every once in a while.
Wow – for over a decade I thought the “Integral” in IFR referred to the closed-loop fuel cycle of the design – whether pyroprocessing was local or regional. Part of me suspects some kind of “rebranding” but that may be unkind, and it doesn’t matter to the big picture.
Since I first heard of the IFR, I have believed it to be the most rational method to reuse the used fuel rods from LWRs – namely to destroy/reduce the TRUs while generating useful amounts of electricity. I envision a relatively small fleet of IFRs that run off the used fuel rods we already have, but not using IFRs for whole scale replacement of coal and natural gas. Disclosure: I am an advocate of all rational fission power plants but lean heavily to SMRs, especially MSR/LFTR, for replacing coal and natural gas.
Thank you for showing that IFR is not dead, as I had feared!!
Glad to hear John Sackett is still involved in fast reactors. He was a great speaker when he was on a panel for NAYGN many years back and it is good we still have such experienced folks to show the way…
…if we are smart enough to go forward and not stall over the politics.
I’ve long wondered why sodium-cooled plant designs include steam generators at all. That’s the part I’d like to be rid of, if possible, & as I understand it, the “closed-cycle” Aircraft Nuclear Propulsion Program reactor concepts used sodium or Na-K as an intermediate coolant to transfer heat to the air turbine.
I’m sure it would mean a loss of thermal efficiency, since sodium-cooled reactors operate at a relatively low temperature for modern gas turbines, but (as you so often point out, Rod) fission heat is cheap, & eliminating the whole steam loop would reduce the capital investment substantially, in addition to gains in safety/reliability/capacity-factor.
I thought Sodium plants weren’t quite high temperature enough for gas-cooled turbine systems? I could be wrong though – I’m a little fuzzy on those sorts of technical details.
Rod: Question about your idea of regional recycling facilities – I can see how, for most types of material streams, regional facilities makes sense – economy of scale – one large plant instead of dozens of small plants.
However, isn’t there a problem specifically related to the fuel coming out of a nuclear reactor – that it is HIGHLY radioactive for months or years? I thought the advantage of on-site recycling was that you didn’t have to transport highly radioactive material, which brings it’s own set of costs (particularly, political costs as you will have antis banging the drum to the public of just how radioactive the partly used fuel is, and that it’s traveling through their town on rail or truck).
Would it be safe to transport such highly radioactive material? Or, would you just let the material cool off for a decade onsite before transporting it to the regional recycling center?
With onsite processing, you just use robots to move the fuel out of the reactor, through a tunnel, and into the recycling center, yeah?
I thought with IFR, the real idea to get economy of scale is that you build several reactors at a single site (like maybe 5 or 6), and that way, you can share a single recycling center between them?
The European ASTRID project officially switched from steam-Rankine to nitrogen-Brayton a couple of years ago — for safety reasons.
https://dl.dropboxusercontent.com/u/11686324/ASTRID_Brayton.jpg
It seems unlikely that IFR/PRISM are going to make much progress without matching ASTRID in this respect. The sooner the better.
The other important trend in recent decades, for SFR design, has been a trade-off of fissile breeding ability for safety.
This is a direct trade-off, because it involves making the core more leaky for neutrons, such that a coolant void incident (sodium boiling in one or more fuel channels) has a negative reactivity effect, rather than a positive one.
In the ASTRID project, this trade-off results in a reactor that is slightly below iso-breeding.
It seems unlikely that an IFR/PRISM without similar core design would get licensed.
TerraPower’s TWR looks to be following similar safety design philosophy — avoiding the top part of the breeding blanket, that characterized EBR-II for example.
https://dl.dropboxusercontent.com/u/11686324/GE-Hitachi_PRISM_tech_brief.jpg
Thanks for the response Jaro. I have a followup question:
I have read in several sources, including “Plentiful Energy”, that the IFR reactor has a negative temperature coefficient of reactivity based on some physics principle (doppler something, if IIRC), that causes the reactor to automatically and passively shutdown if the temperature gets above a certain point, and that certain point is well below the boiling temperature of sodium.
So, the scenario you mention, sodium boiling, is impossible, isn’t it? Why should a physically impossible contigency be regarded as a licensing problem?
The article gives me the impression that the piping problems with liquid Sodium have been dealt with. It made me curious as to how much easier it would be to pump liquid Sodium than molten salt.
I guess you can use magnetic pumps for Sodium.
http://demec.ufpr.br/reterm/ed_ant/17/artigos/science01.pdf
I would suppose these pumps would have less wear and tear than those with large impellers. Seems like this may be a big advantage over a molten salt reactor. Is this wrong?
I also read Plentiful Energy ( as well as Tom Blees’ ‘ Prescription for the Planet’ , plus numerous blog entries from Barry Brook, who first convinced me that nuclear was necessary.) They give a very favourable assessment of sodium cooled reactors, for me only called into question by two things-
1/ The Enrico Fermi unit 1 reactor was, I think, the only one, apart from EBR II and the Fast Flux Test Facility, to use metal fuel, and that was switched to oxide fuel after its coolant blockage and partial meltdown. The Europeans, the Russians, the Japanese, the Chinese, and the Indians are all sticking to oxide fuel. Do they know something I don’t? ( I think the South Korean KALIMERA design follows most of the EBR II recipe, but it’s only a design, not a reactor.)
2/ The Russians, who have by far the most practical experience of running sodium cooled reactors, with the BN600 and BN800 both putting watts into the grid, and the BN350 having run for thirty years, seem to have called a pause on their proposed follow-up BN1200, and are continuing to work on nitrate fuels, and on the supposedly much more corrosion-prone lead-cooled reactor.
Right. That passive shutdown mechanism refers to reactivity increase due to withdrawal of an absorber rod. That causes a rapid rise in fuel temperature, which has the negative reactivity effect you describe.
The whole transient is very quick — a power pulse with a high peak but narrow half-width, which means very low total energy release ( = peak power x half-width).
Certainly not enough to boil the coolant.
But it is presumed that coolant boiling can occur in other ways.
The question then becomes which reactivity effect “wins” — the positive reactivity effect due to voiding, or the negative reactivity effect due to Doppler.
The answer may depend on details of the design.
Evidently the ASTRID and TWR design groups are taking a conservative safety approach in making their reactor core leaky, with a negative void reactivity coefficient (but reduced breeding ratio).
The IFR/PRISM group have traditionally been much less forthcoming in their disclosure of design details, so we may never find out — until some regulatory agency reviews the design and publishes their findings.
In the case of the UK proposal for a plutonium burner, this should not be an issue, since the reactor is supposed to be a burner in any event.
It may become an issue for anyone wanting a fissile breeder however.
A sodium plant can power a nitrogen driven turbine. It’s a lot safer because nitrogen is unreactive, like an inert gas. Yet it is less efficient than steam driven turbines. Chemically, Na + CO2 requires an induction time at 500°C or less, due to CO formation, and the resultant reaction between Na and CO. “At higher temperatures, a fast global reaction occurs”. I believe the expected operating temperature of the IFR is 550°C, and that 650°C or more is normally considered right for supercritical CO2 Brayton cycle. [ Excuse my engineering ignorance but, for me, ultra-safe N2 driven turbines get my vote for the IFR. ]
These slides have a nice diagram comparing the efficiencies of N2/CO2/steam on page 9: slides. Here’s a paper with the same comparison: pdf download (2014). As expected N2 Brayton loses a lot of efficiency over steam of CO2, but it is so many orders of magnitude safer; being so unreactive.
As clarification, sodium is a strong reducing agent and very benign in a reactor environment.
Sodium reacts with spectacular vigor with water, but not so much with air. Sodium spontaneously ignites in presence of air, but it burns with a (relatively) cool, smoky flame, not at all as it does with water.
Lead-cooled reactors have their own trade-offs. Pure lead melts at 325 C, PWR’s run at about 305 C max. To obtain something more workable, lead is usually cut with bismuth to form a lead-bismuth eutectic (LBE) which melts at 124 C. But molten lead is corrosive, which presents a design consideration possibly affecting lifetime of plant.
There is another trade-off with neutron activation products. Sodium absorbs neutrons to give activated sodium-24, a strong gamma emitter with a half-life of 15 hours — short enough not to be a disposal hazard, as it will all go away within a week of reactor shut-down.
In contrast, Bismuth-209 absorbs a neutron to produce (after beta decay) Polonium-210, a potent alpha emitter with half-life – 138 days – long enough to be a disposal concern. Not an insurmountable concern — alpha is not penetrating radiation — but still an additional something to take into account.
Polonium-210 has both useful (makes a great RTG heat source for moderate life batteries) and sinister applications (Litveninko).
To me, one of the beauties of the IFR concept is the simplicity of the reprocessing step. This simplicity makes local reprocessing possible.
The metallic fuel elements are put into a bucket, and the contents of the bucket electroplated into a molten salt bath. The current is reversed, and the actinides (fuel) are plated onto a rod to be fabricated back into metallic fuel elements.
The trickiest part seems to be the distillation of the molten salts to remove fission products.
@donb
You neglected to mention the hot cell handling requirements. Ever seen remote manipulators, observed the challenges of training operators, or thought about the life cycle costs of such facilities?
Would N2 going through a heat exchanger with activated sodium make any Carbon 14?
@John ONeill
No. The activation of N2 to C-14 requires neutrons. Those are only available inside the primary shield that stops reactor generated neutrons.
@John ONeill
I believe that other nations are sticking to oxide fuel rods due to being conservative and sticking to what is familiar. Metal fuel rods are better at containing fission products, especially gaseous FPs, without cracking or releasing. They also expand with temperature more, which IIRC reduces reactivity which is safer. And the heat transfer is somewhat better, so the internal peak temperature will be a bit lower – I think an annular fuel rod design would help either type of fuel form. Of course the down side is that the metal fuel rods melt at much lower temperatures.
Well if IFR is going to be built. it is either with no revisions or with some revisions.
I would advocate to move away from Na to Pb-alloy since it does not burn and “freezing” issue can be dealt with.
@JM
There are many measures of effectiveness associated with coolant choices. I’m not a fan of using molten lead
…
“There is less experience with using it compared to sodium”
I must admit one point that Sodium run reactors have longer running time.
But Running time of Lead- Bsimut cool reactors stem from 80 year of operating time in Alfa-class. Which is not insignificant. Only issue was freezing issue aka those reactors could not be restarted. But now they have soled this issue in planed reactors. Of 1GW size as well as SMR where freezing is bonus since you transport whole reactors frozen..
There would have to be Cost/ benefit analysis. But with Lead /Pb you can drop those double walled steam generators and fire hazard. Leaks of Na are not uncommon. French Phoenix suffered the same issue. On disadvantage side you have to solve metallurgy of vessel. Yes but if you solve it you solve for good. It is solved = done
With Na you pray that those fancy steam generators do not breach again once they are punctured somehow, the Inert gas cover stays there, there is no leak with water around… . Well I am big fun of Black Swan book. And with Na it is just black swan nest.
To reprocessing on site. Prime rational was ….drum roll… PROLIFERATION.
Yes it is stupid to have small expensive chemical plant next to every IFR site, but somehow Pu must stay there was great benefit.
“Its activation products are longer lived than those associated with sodium”
With U/Pu cycle you would still have some waste. At worst is could still be good 2nd loop coolant assuming that it will not get activated too.
IFR did not get licence so still time for improvement. If I have to go fast I would go for Dual Fluid Reactor 😀 Breeding ration of 1,6 and doubling time of 4 years is tempting as COMBO for DMSR for not trusted parties running mox fuel and DFR for Big boys.
@JM
But Running time of Lead- Bsimut cool reactors stem from 80 year of operating time in Alfa-class.
Perhaps. But the operational and technical details associated with those machines is less available to western countries that the operational and technical details of US navy nuclear plants.
In other words, we know they operated. That’s about it.
Mark Pawelek: Well, from a fuel efficiency standpoint, it’s very likely worth the tradeoff, I would think – lower efficiency than steam, perhaps, but the fuel (assuming you’re using SNF left over from light water reactors, maybe be ‘free’ or *even have a negative cost* [I think there’s an argument to be made that since light water reactors have been paying for decades into a waste disposal fund, that fast-breeder reactors ought to be elligible to claim those disposal fees for any waste they take; then they would in turn pay their own waste disposal fee to dispose of the waste they generate and ship out – that way, the DOE isn’t being paid multiple times for the same waste].
Of course, in order to use that waste, an IFR would have to process the waste, as discussed in this article, so the fuel wouldn’t really be free, but quite possibly very low cost, and more to the point – because of breeding and recycling, there is an absolutely enormous amount of potential energy in that fuel, so the cost per unit energy should approach zero. (That’s just fuel cost – not capital costs for the power plant, recycling center, etc).
But, I wonder about nitrogen turbines – efficiency aside, how would the cost compare to SuperCrit CO2 turbines? My understanding is that, on the one hand, SC-C02 turbines have to use highly specialized materials in their construction, which would make them more expensive, but they are also supposed to be a LOT smaller, which might make them ultimately cheaper than larger Nitrogen turbines?
Is it possible that it’s worthwhile going with SC-C02 not for the efficiency, but for the capital costs?
That’s an interesting possibility, is it not? Reducing the size of the turbine to fit in a pickup bed, while slashing the maximum operating temperature from the 1380°C of high-performance gas turbines to the ~500°C limit of CO2, is bound to cut a bunch of costs in both manufacturing and materials.
The problem I can see is that the small size requires high speed, so a direct-drive synchronous alternator is no longer feasible for grid power. Frequency conversion electronics running at hundreds of megawatts will add a bunch of capital cost, and O&M too. There’s always a mechanical motor-generator but those run into money too.