The Ergosphere
Friday, October 22, 2010
 

Why the Integral Fast Reactor had to die

Over at The Oil Drum, I left this comment, not really grasping the enormity:

Fast breeders can turn any isotope of uranium into energy (as well as burning all the transuranics), so you might want to look at this 2008 paper:

During the 50 years that the Federal Government controlled the U.S. uranium enrichment enterprise, DOE generated over 700,000 metric tons of depleted uranium hexafluoride (DUF6).

About 10% of that is suitable for enrichment to LWR fuel, but fast breeders or converters can use 100% of it.  At 0.8 tons of metal/GW-yr, that is 590,000 GW-yr of generation or 590 years at 1 TWe without mining another gram of uranium.  The USA is currently using about 450 GW average.

A Gen IV nuclear reactor will probably have a thermal efficiency of about 40%; therefore, 1 TWe = 2.5 TWth.  What I didn't realize at the moment is how this compares to total US energy consumption.  The USA uses approximately 100 quadrillion BTU per year, give or take.  This is just a bit over 3.3 TWth.  The conclusion is that uranium already in the US inventory, burned in Integral Fast reactors, could power the USA for over 400 years at current rates of consumption!

I'm sure you see the implications.  The coal-mining industry:  gone.  The gas-drilling industry:  gone.  Coal transport by rail and barge:  gone.  Gas and oil pipelines:  gone.  Oil refining:  reduced to supplying the chemical industry.

"What's the problem," I hear people asking.  "Wouldn't we rather be without gas-line explosions and oil spills in our rivers and seas?"  Well, it depends who you are.  If you are one of the monied interests in the carbon-based fuel sector, these things represent huge amounts of wealth wiped out, and the rest transferred to regulated monopolies.  All of the power they create today, everything that lets those industry heads rub elbows with senators and play golf with heads of state... gone.

They can't let that happen, even if it would take 50 years.  That's why the Integral Fast Reactor had to die.

 
Comments:
Rod Adams has come to the same conclusion that you have, and he regularly posts articles about this under the “smoking gun” tag on his Atomic Insights blog.
 
"100 quadrillion BTU per year, give or take. This is just a bit over 3.3 TWth."

1 Joule = 1 Ws (Watt second)

1 BTU = ~1055 Joules

100 quadrillion = 10e17

100 quadrillion BTU = ~ 10e20 Joules

or 100 Exa Joules.

1 KWh = 3.6 MJ

1 TWh = 3.6 PJ

3.3 TWh = ~10 PJ

100 EJ = 100,000 PJ

so 100 EJ = 10 PJ * 10,000

1 year = ~ 10,000 hours (8,760 exactly)

Did you mean 3.3 TWh?
 
The USA uses over 4000 TWh of electricity alone every year, and you don't even have the units right.

100 quads = ~1.05e20 J

1 yr = 3.16e7 sec

100 quads/yr = ~3.3 TW
 
Now I understand, you converted BTUs to Watts by dividing out years.

In your inventory of existing actinides for use in IFRs, do not forget the so-called waste from light water reactors. It contains about 95% of the energy of the pristine fuel.
 
I'd never forget that, but as there's only about 60,000 tons of it in the USA it amounts to a rounding error in the big calculation.
 
It's probably not all conspiracy theory though. The IFR has greater cost than LWRs, and the latter were already proven. Plus there appeared to be plenty of fissile material in the ground available at low cost. This meant there was too little industry and commercial support. It was mostly a government pet peeve project. That's usually a recipe for disaster.
 
I've read that the US had to refund Japan's contribution to the IFR experiment, which was more than it would have cost to build and run it.  If that's true, the cost argument doesn't wash.  The IFR also eliminates heavy pressure vessels and large containments required to hold vented steam, so the cost might well fall below that of LWRs.

There are costs beyond the immediate price of fuel.  The issue of waste disposal is pertinent, and Kirn Sorensen notes that spent LWR fuel becomes an excellent source of plutonium after about 300 years.  Kicking that can down the road only makes sense if you have a high discount rate, and we arguably have made lots of serious mistakes that way already.
 
Actually dry storage in modular canisters on the surface is very cheap and works well for 300 years (might need a few times repackadging). Kirk is getting a project starting to test fluorination equipment for Teledyne Brown Engineering.

Discounting is simple and real. You have some money, about 0.1 cent/kWh charge for nuclear kWhs, and put in in a fund of government bonds in perpetuity. You get reasonable discount rates this way, 2-4 percent above inflation. The US decommissioning fund is around 29 billion or something like that and still increasing. That's too much to dismantle all plants, in fact its even plenty for 300 year dry storage using existing module tech. Dry storage is in fact so simple and cheaper than underground parking areas, it's almost funny that people worry about it. But its sad that people worry about little things and end up doing the big things wrong, like building more coal plants.

As long as we don't attempt needlessly expensive and irrational fear mongering technologies such as Yucca Mountains, whose only major achievement is scare the public in thinking waste storage is actually a big problem, the fund will be plenty for both decommissioning and 300 year spent fuel storage.

Some of the spent fuel will likely be reprocessed much earlier, due to radio-isotopes value (for medicine and diagnostics, eg Cs-137, Sr-90, I-131, Sm-151 etc).

Fluorides reactors seem much more attractive to me than IFRs. The reprocessing becomes so much easier with better seperation factors. The IFR repo systems have never done much better than 2% actinide losses. That makes a bunch of actinide waste. Fluorination and distillation are much better.

I'm not sure if you are referring to the Monju breeder in Japan, but that plant was a disaster. It was expensive, poorly run and mismanagement in communications regarding the sodium fire. Then it was shut down 15 years. For investors this is a nightmare, no way convincing in the business case.
 
The workability of dry-cask storage isn't at issue; the problem is that after 300 years or more, 99.9% of the stuff that "protects" the SNF from diversion has decayed and you have what amounts to a very high-grade plutonium ore.  Sorensen says that this is a good reason to reprocess immediately, and I agree.

The useful radioisotopes like Cs-137 are no longer present in significant amounts after 300 years.  If we want them, we need to take them from relatively fresh fuel.  Again, this supports Sorensen's proposition.

In an era when governments are defaulting on debt and inflation-protected bonds are selling for more than their face value at maturity, do you think the idea of government bonds as a store of value lasting centuries can be taken seriously?  Do you think that some pols wouldn't raid the fund, knowing that the problem wouldn't have to be dealt with until after they were long dead?  I'm not interested in concepts based on unrealistic views of human nature.

I agree that LFTRs are highly attractive, but they operate with a thermal spectrum.  This makes them unable to effectively destroy actinides, and they can't make use of the 470-odd thousand tons of depleted uranium we have lying around.  Fast neutrons will burn actinides; even Pu-238 and Pu-240 will fission with neutrons over 1 MeV.  A fleet of IFRs could be started using our spent LWR fuel and run for centuries on nothing else, leaving essentially nothing but fission products.

"Pyroprocessing" is essentially the LFTR reprocessing method too.  How can it have such a high rate of actinide loss in one application and not the other?  Both actinides and fission products wind up in the reprocessed IFR fuel, which constitute its protection against diversion.

The IFR had nothing to do with the Monju plant, which uses MOX fuel (the IFR was going to use metal fuel).  Japan had contributed to the cost of the IFR experiment, and the money had to be refunded when Hazel O'Leary persuaded Clinton to cancel it; the refund came to more than the remaining cost of the experiment.  That was a fiscally idiotic move, which could only be justified by ideology:  the threat wasn't that a technical problem would prevent the IFR from working, but an ideological problem that it would.

Last, the LFTR has to be started on something.  I've seen the suggestion that two-fluid LFTRs be started on LEU, build up their inventory of U-233 (taking roughly 2 years), and then switch fuels.  The partially-used LEU could then be taken to start another plant, with some makeup material added.  But what do you do with the actinides created in the LEU?  A system of IFRs could burn them instead of disposing of them as waste.

I view these two systems as complementary.  Per David LeBlanc, a fast-spectrum reactor needs a fissile inventory of 10-20 tons per GWe.  If it burns 0.8 tons per GW-yr and has a breeding ratio of 1.2, it will take between 60 and 130 years to double its inventory; this is not going to grow fast enough to meet our needs.  But we have about 60,000 tons of SNF in the USA (at 1.0% Pu and other actinides, enough to start 30 to 60 GWe of fast-spectrum reactors) and our LWR fleet will be cranking it out for the next 60 years or more.  If we stop building LWRs by 2030 and make LFTRs instead, we'll have a need for a certain amount of LEU (assuming no excess U-233 bred in other LFTRs) to start each one, and get a smaller but still significant amount of actinides needing disposal.  The IFRs can dispose of that along with the existing SNF and operate on an independent fuel supply, depleted uranium.
 
Okay, warning:  your comment was published, but was later flagged as spam by Blogger and "disappeared".  I caught this, checked the spam box, and was able to get it re-published.  If you see this happen again, let me know.
 
Well, fertile material has never been the problem. There's plenty of Th-232 and U238. The problem might be fissile, but more in terms of bottlenecks than actual resources. Mining capacity, enrichment, or spent fuel reprocessing rates will dominate scale-up. Fast reactors require a lot more fissile than thermal or epithermal. The French work on Thorium Molten Salt Reactors started with plutonium in a fastish spectrum showed it can do very well in deployment, better than just LWR-IFR. It can iso-breed, CR=1. The IFR had to reduce the spectrum to get better reactivity coefficient for safety reasons. They initially thought breeding of perhaps 1.2-1.3 would be possible, but this is only possible in the higher eta region of the tranuranics. Because of the slower spectrum modifications (eg by adding some BeO moderator) the breeding was considerably reduced, IIRC something like 1.04 is the aim now. This means it takes nearly forever to transition purely from LWR spent fuel to IFRs.

Fast fluid fuelled reactors have easier deterministic safety, so one wonders if a fast chloride or fluoride machine wouldn't be a better candidate. Ottewite and Taube believe that the breeding could be as good as 1.3, with fully negative dilatation coefficients. Too bad their work is largely theoretical.

The IFR will probably use electrolysis reprocessing in molten chloride baths. I'm an idiot when it comes to chemistry but apparently the seperation factors and resulting losses of plutonium were not very good in actual experiments. They lost 1 or 2 percent every pass.

The cool thing about a fluoride reactor is that you can have a vacuum distillation unit and this is simple, compact and doesn't do anything chemically. Because of the big differences in the fluoride volatilities, the fuel and carrier salt comes out very cleanly, very little uranium fluoride is wasted. Plutonium is harder and will mostly stay in the still, another reason the thorium cycle is attractive - very little plutonium is generated and it is mostly Pu238, no good for weapons. An IFR could do something similar but would have to first fluorinate the metal fuel, distill, then reduce the fluoride or chloride back to metal again, so more chemical steps. But then they can't run well on U-Pu since distillation wastes most of the Pu. For metallic fuels, other processing methods might be attractive. Metal zone refining might be more attractive for the IFR, no chemical conversions required. This saves cost and messes. It is quite an advantage to keep the fuel in its original chemical form during processing.

The Monju breeder had a total cost of about 7.5 billion, a lot of money for a 280 MWe reactor, more than 26 dollars/Watt. I know this is a prototype but its also reasonably sized and there should be some economy of scale already. There were a lot of problems with the Monju machine. Sodium is low heat capacity, reactive and opaque, not ideal traits for a coolant.

Do you have good estimates on the cost of an IFR?

Regarding governments stability, without that no nuclear scheme, or any advanced society for that matter, could endure. I don't recall the US government ever not paying out on their bonds. If they did that they'd lose credibility and would have to buy their money at higher rates (risk premium). Not what they want! But you could spread out the fund over a number of private banks, even different currencies, if you'd like. My point is that discounting is real up to a certain point. Agree that discounting beyond interest on bank accounts, bonds etc. is dubious.
 
Well, fertile material has never been the problem. There's plenty of Th-232 and U238. The problem might be fissile, but more in terms of bottlenecks than actual resources. Mining capacity, enrichment, or spent fuel reprocessing rates will dominate scale-up. Fast reactors require a lot more fissile than thermal or epithermal. The French work on Thorium Molten Salt Reactors started with plutonium in a fastish spectrum showed it can do very well in deployment, better than just LWR-IFR. It can iso-breed, CR=1. The IFR had to reduce the spectrum to get better reactivity coefficient for safety reasons. They initially thought breeding of perhaps 1.2-1.3 would be possible, but this is only possible in the higher eta region of the tranuranics. Because of the slower spectrum modifications (eg by adding some BeO moderator) the breeding was considerably reduced, IIRC something like 1.04 is the aim now. This means it takes nearly forever to transition purely from LWR spent fuel to IFRs.

Fast fluid fuelled reactors have easier deterministic safety, so one wonders if a fast chloride or fluoride machine wouldn't be a better candidate. Ottewite and Taube believe that the breeding could be as good as 1.3, with fully negative dilatation coefficients. Too bad their work is largely theoretical.

The IFR will probably use electrolysis reprocessing in molten chloride baths. I'm an idiot when it comes to chemistry but apparently the seperation factors and resulting losses of plutonium were not very good in actual experiments. They lost 1 or 2 percent every pass.

The cool thing about a fluoride reactor is that you can have a vacuum distillation unit and this is simple, compact and doesn't do anything chemically. Because of the big differences in the fluoride volatilities, the fuel and carrier salt comes out very cleanly, very little uranium fluoride is wasted. Plutonium is harder and will mostly stay in the still, another reason the thorium cycle is attractive - very little plutonium is generated and it is mostly Pu238, no good for weapons. An IFR could do something similar but would have to first fluorinate the metal fuel, distill, then reduce the fluoride or chloride back to metal again, so more chemical steps. But then they can't run well on U-Pu since distillation wastes most of the Pu. For metallic fuels, other processing methods might be attractive. Metal zone refining might be more attractive for the IFR, no chemical conversions required. This saves cost and messes. It is quite an advantage to keep the fuel in its original chemical form during processing.

The Monju breeder had a total cost of about 7.5 billion, a lot of money for a 280 MWe reactor, more than 26 dollars/Watt. I know this is a prototype but its also reasonably sized and there should be some economy of scale already. There were a lot of problems with the Monju machine. Sodium is low heat capacity, reactive and opaque, not ideal traits for a coolant.

Do you have good estimates on the cost of an IFR?

Regarding governments stability, without that no nuclear scheme, or any advanced society for that matter, could endure. I don't recall the US government ever not paying out on their bonds. If they did that they'd lose credibility and would have to buy their money at higher rates (risk premium). Not what they want! But you could spread out the fund over a number of private banks, even different currencies, if you'd like. My point is that discounting is real up to a certain point. Agree that discounting beyond interest on bank accounts, bonds etc. is dubious.

Hmm, blogger is giving me all sorts of error messages. Maybe you're right, perhaps I really am a troll ; )
 
BTW, I caught another false-spam flag.  I told Blogger the comment was not spam (trying to fix the Bayesian engine) but you probably want to delete the duplicate comment.

I've been looking for details on actinide losses in electrorefining and not finding any (this 2007 progress report doesn't mention the issue, though it does claim >99% recover of cadmium from cadmium pool electrodes).  Got anything I can read?  This JAERI/CRIEPI paper claims "Collection efficiencies for plutonium ... were nearly 100%".  There's a hint that remnant Pu trapped in compounds with reduced lithium could be retained by decanting salt, or the problem could be prevented by limiting the electrode voltage.

I'm getting tons of info on the IFR from the Q&A on Steve Kirsch's blog.  It has this:

    Q. How much would it cost to build a 1 GW IFR plant?

    The first one will probably cost around $1 to $2 billion.

Of course, the real problem is our regulatory system which can drive costs up without limit or reason.  There's no other reason that it takes us 10 years to build an AP-1000, and the Chinese can build one in 4 years for a small fraction of our cost.  Modular reactors with all the major pieces factory-built and trucked (not barged) to the site ought to slash the cost, unless the regulatory apparatus insists on protecting its employment and power.

The USA has already failed to pay on bonds, by abandoning the dollar peg to gold in 1972 and inflating the currency over extended periods.  If the purchasing power of the funds set aside for waste handling is depleted (and there is every incentive to do so), future generations will be stuck with a burden and not enough resources to deal with it.  Better to take care of everything in 20 years, if not 10.
 
Post a Comment

Links to this post:

Create a Link



<< Home
Talk largely about energy and work, but also politics and other random thoughts


Mail Engineer-Poet

(If you're mailing a question, is it already in the FAQ?)

Important links

The FAQ
Glossary
The Reference Library

Blogchild of

Armed and Dangerous

Blogparent of

R-Squared




The best prospect for our energy future:
Flibe Energy

ARCHIVES
January 1990 / February 2004 / March 2004 / June 2004 / July 2004 / August 2004 / September 2004 / October 2004 / November 2004 / December 2004 / January 2005 / February 2005 / March 2005 / April 2005 / May 2005 / June 2005 / July 2005 / August 2005 / September 2005 / October 2005 / November 2005 / December 2005 / January 2006 / February 2006 / March 2006 / April 2006 / May 2006 / June 2006 / July 2006 / August 2006 / September 2006 / October 2006 / November 2006 / December 2006 / January 2007 / February 2007 / March 2007 / April 2007 / December 2007 / January 2008 / May 2008 / June 2008 / August 2008 / October 2008 / November 2008 / December 2008 / February 2009 / March 2009 / April 2009 / May 2009 / June 2009 / July 2009 / August 2009 / September 2009 / October 2009 / November 2009 / December 2009 / January 2010 / April 2010 / May 2010 / June 2010 / July 2010 / August 2010 / September 2010 / October 2010 / November 2010 / December 2010 / January 2011 / February 2011 / March 2011 / April 2011 / May 2011 / July 2011 / August 2011 / September 2011 / October 2011 / April 2013 / November 2013 / December 2013 / January 2014 / February 2014 / March 2014 / April 2014 / July 2014 / August 2014 / September 2014 / October 2014 / November 2014 / February 2015 / April 2015 / October 2015 / March 2016 / April 2016 / May 2016 / June 2016 / July 2016 /


Powered by Blogger

RSS feed

Visits since 2006/05/11: