Economics of LFTR

General discussion of thorium and nuclear energy.
User avatar
Charles Barton
Posts: 251
Joined: Jun 24, 2007 10:43 am
Location: Dallas, TX
Contact:

Re: Economics of LFTR

Post by Charles Barton » Sep 03, 2011 6:32 am

jagdish wrote:Economics vary with time, place and circumstances like regulatory expenses. Currently, costs are lower in China and India. Outsourcing of MSR/LFTR development to China is more logical, even more than the AP-1000. Indians, logical but slow in nuclear development, have not made a start with molten salts. Wishfully, they could make a start with pyroprocessing, a combination of chloride volatility and electro-refining.
Outsourcing to China does not look like a good idea to me. First, the Chinese are not particularry good at developing advanced technology, and often outsource advanced technology development to other countries. Secondly, the chinese have announced that they are interested in intellectual property rights with respect to LFTR development. We do not want to turn development over to people who want to do that.

martinburkle
Posts: 15
Joined: Jan 26, 2010 12:05 pm

Re: Economics of LFTR

Post by martinburkle » Sep 03, 2011 9:28 am

@Lars
Here is a quote from the AP1000 document on canned pumps:
"There are two pump journal bearings, one at the bottom of the rotor shaft and the other between
the upper flywheel assembly and the motor. The bearings are a hydrodynamic film-riding design.
During rotor rotation, a thin film of water forms between the journal and pads, providing
lubrication.
The thrust bearing assembly is at the bottom of the rotor shaft. The pivoted pad hydrodynamic
bearing provides positive axial location of the rotating assembly regardless of operating
conditions."

I assume that introducing water as a lubricant would be a bad thing for a molten salt reactor. So each time I read "hydrodynamic" I am wondering if the molten salt could replace water and be come "saltdynamic". Of course, heaters would be required to melt the lubricant at startup.

How would pump bearings bearings work in the molten salt environment?

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 03, 2011 9:52 am

martinburkle wrote:@Lars
Here is a quote from the AP1000 document on canned pumps:
"There are two pump journal bearings, one at the bottom of the rotor shaft and the other between
the upper flywheel assembly and the motor. The bearings are a hydrodynamic film-riding design.
During rotor rotation, a thin film of water forms between the journal and pads, providing
lubrication.
The thrust bearing assembly is at the bottom of the rotor shaft. The pivoted pad hydrodynamic
bearing provides positive axial location of the rotating assembly regardless of operating
conditions."

I assume that introducing water as a lubricant would be a bad thing for a molten salt reactor. So each time I read "hydrodynamic" I am wondering if the molten salt could replace water and be come "saltdynamic". Of course, heaters would be required to melt the lubricant at startup.

How would pump bearings bearings work in the molten salt environment?
Such bearings have been developed for molten nitrate salt pumps. Seems to work fine, its a question of the right materials choice and chemistry control. With nitrate salts there is a hard protective oxide layer which is not there on the fluoride salt version. So you need to use a fairly hard alloy. Hastelloy N or TZM should be fine. The problem is the seals at 1000 Kelvin. With no seals the biggest problem is gone.

Lars
Posts: 3060
Joined: Jul 28, 2008 10:44 pm

Re: Economics of LFTR

Post by Lars » Sep 03, 2011 9:58 am

ORNL did do some work with salt lubricated bearings. I don't know the end result. Their initial design as used in MSRE used only a bearing above the impeller and it was hydrocarbon lubricated. They did have to keep the speed of the impeller down to avoid vibrations. It seems like a bearing below the impeller would be a big help with vibrations.

Dr. Peterson said that one of the challenges with fluoride salts is that they clean the metal very, very well. So clean in fact that you can get contact welding. So fluoride salts can be complicated to use for lubrication.

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 03, 2011 2:48 pm

Lars wrote:ORNL did do some work with salt lubricated bearings. I don't know the end result. Their initial design as used in MSRE used only a bearing above the impeller and it was hydrocarbon lubricated. They did have to keep the speed of the impeller down to avoid vibrations. It seems like a bearing below the impeller would be a big help with vibrations.

Dr. Peterson said that one of the challenges with fluoride salts is that they clean the metal very, very well. So clean in fact that you can get contact welding. So fluoride salts can be complicated to use for lubrication.
How do you get contact welding when the fluoride salt cools the bearings and housings? The salt would go in the pump at inlet temps, which are around 550-570 degrees C. That is a long way from the melting temp of the alloy, and if you have fluoride salt immersion right next to the surface of the metals, it would be hard to get peak temps on the surface up a lot.

Lars
Posts: 3060
Joined: Jul 28, 2008 10:44 pm

Re: Economics of LFTR

Post by Lars » Sep 03, 2011 3:05 pm

You don't need high temperatures for contact welding - just very, very clean ones. Check out contact weld in wikipedia.

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 04, 2011 5:10 am

Lars wrote:You don't need high temperatures for contact welding - just very, very clean ones. Check out contact weld in wikipedia.
Cool, this appears to be the process that makes sedimentary gold nuggets!

http://en.wikipedia.org/wiki/Cold_welding

I've never heard any engineer suggest this is a problem, but since fluorides are such excellent fluxes, Dr. Peterson could be right (as usual of course!).

If this is a problem, an obvious solution would be to use dissimilar metals for the bearings and the housings. Possibly a SiC ball or journal bearing would be interesting to combine with Hastelloy N housings. Though SiC is much harder than Hastelloy.

If the contact welding is still a problem I suspect journal bearings are out of the question. They depend much more on a good lubrication.

cloa513
Posts: 78
Joined: Jun 30, 2011 7:10 pm

Re: Economics of LFTR

Post by cloa513 » Sep 04, 2011 6:27 pm

Probably use induction coldwelding. There is a reasonable video of it on Youtube.

Also inertia welding and a few other cold welding processes available.
http://www.welding-technology-machines. ... elding.htm

User avatar
zeropoint
Posts: 10
Joined: Jul 08, 2011 10:01 pm
Location: Austin Tx

LEU startup

Post by zeropoint » Sep 07, 2011 12:48 am

Thanks Charles for the pointer. I read some posts on Nuclear Green as well as some threads on the forum about possible LFTR and LCTR (fast chloride reactor) startup fuels.

Several people have suggested that LFTRs can start with LEU (19.75%) instead of U233. For a 1 GWe plant we need 1 tonne of LEU. The price for 1 kg of LEU is approximately 8,854 USD (got the number from here and here) which comes to 8,854,000 USD for 1 tonne.

What is the engineering drawback of using LEU? If LEU works for startup fuel then why would we care about building Floride/Chloride fast breeders that only have a breeding ratio of 1.1?

I read here and here that LEU is not made by commercial enriching companies but is made by the DOE by downblending HEU (High Enriched Uranium) and they sell it to research reactors.
Can LEU be purchased for commercial reactors not just research reactors?
If the DOE was not willing to supply 10 tonnes of LEU a year, a commercial enriched uranium supplier could supply the 19.75% LEU correct?
- ZeroPoint Energy

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 07, 2011 2:53 am

It is not feasible to run as a breeder on thermal spectrum with LEU. Not enough neutrons. You need thorium for that and there’s not enough place for it with all that U238.

There is another problem in the U238. There’s lots in LEU so you get lots of plutonium. We don’t have technology to process plutonium out, and many experimental results are not encouraging. If you lose the plutonium with LEU started reactors your neutron budget gets abysmal.

Fast fluoride with a faster salt and no moderator other than the salt is one possible solution. The faster spectrum means you need not process that much fuel so don’t lose that much plutonium, and you will get a bit more neutrons from the faster spectrum (with LEU). Major downsides are more fissile startup needed, more materials issues from higher fast neutron flux, and more leakage of neutrons due to not being able to have an outer undermoderated neutron cushion (since there is no moderator).

One economical solution is to start a converter reactor that does not process fissile material online, only remove the noble metals and gasses by filtering and sparging. This will allow you to run a long time. You can run on LEU only which then allows you to use LWR enrichment levels (<5%). So you can buy that from commercial enrichment companies. You also don’t make any U232 since no thorium is present. Its true that you can’t isobreed like this, and the conversion ratio will be not much better than LWRs, but the uranium requirement will be much lower due to the high burnup and high thermal to electric efficiency.

Dr. LeBlanc’s presentations and PDFs are highly recommended on this matter. Google them.

User avatar
Lindsay
Posts: 821
Joined: May 24, 2009 4:42 am
Location: Calgary, Alberta
Contact:

Re: LEU startup

Post by Lindsay » Sep 07, 2011 4:25 am

zeropoint wrote:Several people have suggested that LFTRs can start with LEU (19.75%) instead of U233. For a 1 GWe plant we need 1 tonne of LEU. The price for 1 kg of LEU is approximately 8,854 USD (got the number from here and here) which comes to 8,854,000 USD for 1 tonne.
Depending on the design, I think that you'll need more fissile than that for a 1 GWe plant.

ORNL-TM-7207 Concept Design of a DMSR (in the EFT doc repository) requires 3.45 t of U235 as 17.45 t of LEU, but one can do better than that. As I understand it, running LEU curses you to have to carry more fissile due to the absorptions in U238 as Cyril mentioned as well as wrecking the breeding potential (which for some is a significant benefit in terms of enhancing proliferation resistance).

You can have LEU and low Pu though, but only if the spectrum is very thermal, this enhances the relative reactivity of Pu while hopefully minimising the resonant absorptions of U238 in the epithermal spectrum. DMSR does this, but due to the very high fertile content in a single fluid design it has to carry a reasonable quantity of fissile.

As far as breeding goes, one needs to ask whether it is really necessary, fissile is cheap and many of these designs can run as efficient converters requiring very small fissile additions to keep running. For example DMSR only needs ~165 kg of U235 a year on average to keep going. With a few tweaks here and there one should be able to get that down below 100 kg/year, I think that's pretty economical.

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 07, 2011 6:38 am

There is a trade-off between fissile startup and breeding ratio. You can make do with less thorium which reduces the fissile required due to less absorptions in thorium, but that reduces the breeding ratio for the same reason - less thorium to absorb neutrons to breed new U233. So if you want a high breeding ratio then you need more fissile.

Like Lindsay said, ORNLs DMSR used 3450 kg of U235 which, at 19.9% U235 enrichment, means over 17 metric tonnes of LEU.

http://www.thoriumenergyalliance.com/do ... eBlanc.pdf

You then add roughly 1000 kg of LEU per year to keep the reactor critical.

If you have a more compact reactor, maybe 1.5 or 2 fluid, then you can have a lower fissile startup and a higher breeding ratio but you need processing, which is complicated by the use of LEU.

I think it is better to either run a DMSR with uranium only and <5% LEU or a similar reactor with thorium and plutonium from spent nuclear fuel. Currently I am wondering if we can fluorinate the spent oxide fuel from LWRs and use the fluorinator bottom ‘gunk’ as startup for the latter type of reactor.

David
Posts: 909
Joined: Mar 07, 2007 11:02 am
Location: Ottawa

Re: Economics of LFTR

Post by David » Sep 07, 2011 7:47 am

Just a quick recap....

MSR is a broad term that can cover all sorts of molten salt reactors from breeders to converters, Single Fluid, Two Fluid, running on just thorium, just uranium, a mix, etc, etc.

LFTR is much narrower which I believe means a breeder (or breakeven, make just enough fuel for yourself) and running off thorium (only thorium?). Going a little further, most assume LFTR is a Two Fluid which means one blanket salt for thorium and one fuel salt for the U233 it produces. If there is any thorium with the fuel salt it changes the dynamics completely and was called a One and a Half Fluid by the original designers at Oak Ridge. Of course, Single Fluid means everything in one fluid.

Getting back to the question, can you start a LFTR off LEU (19.75% U235 or lower). The answer is yes for any Two Fluid design and a conditional no for the others because the U238 would stick around for decades throwing off the cycle. The trick for a Two Fluid design is you run LEU in the fuel salt just long enough to produce and collect enough U233 in the blanket to start all over again with only U233 for the fuel salt. Depending on the design this could take as little as one tonne U235 (i.e. 5 tonnes 19.75% LEU) for softer spectrum designs but more likely at least a few tonnes U235. It will take a year to several years and at that point you remove any leftover LEU from the fuel salt for reuse elsewhere. You'll also have significant amounts of Pu in the used starter LEU fuel salt which is then a national choice how to handle (ideally burn off in the same reactor now running off thorium and U233). There really is not a great deal of difference in overall Pu and/or other transuranic production overall in this method.

Oak Ridge never discussed such a startup method but they assumed you could use highly enriched uranium if U233 or Pu was in short supply. Shipping and using HEU is a lot more contentious today. I threw this startup method into my patent application a few years back, no idea what will come of it.

David LeBlanc

Lars
Posts: 3060
Joined: Jul 28, 2008 10:44 pm

Re: Economics of LFTR

Post by Lars » Sep 07, 2011 10:13 am

One could also use 20% LEU to start up a thermal 1.5 or 1 fluid design. Roughly the ratio of fertile/fissile is around 30-50 in a thermal design so the 238U uses up 4 parts of the 30-50 parts fertile - the rest can be thorium so the design will convert the 235U to 233U in a reasonable time with no need to exchange the fluids and dispose of your initial load.

ORNL proposed something like this for DMSR but they added two additional constraints. First keep the uranium in the fuel salt denatured. This meant adding 238U strictly for the purpose of denaturing the fuel salt. That meant adding lots of extra 238U. Second, they avoided fuel processing until the fuel salt wouldn't burn anymore. If we put back in some processing and depended on fission product self-protection of the hot fuel then I suspect we could achieve unity breeding starting on 20%LEU.

Cyril R
Posts: 5037
Joined: Jul 14, 2008 3:12 pm

Re: Economics of LFTR

Post by Cyril R » Sep 07, 2011 3:17 pm

Lars wrote:One could also use 20% LEU to start up a thermal 1.5 or 1 fluid design. Roughly the ratio of fertile/fissile is around 30-50 in a thermal design so the 238U uses up 4 parts of the 30-50 parts fertile - the rest can be thorium so the design will convert the 235U to 233U in a reasonable time with no need to exchange the fluids and dispose of your initial load.

ORNL proposed something like this for DMSR but they added two additional constraints. First keep the uranium in the fuel salt denatured. This meant adding 238U strictly for the purpose of denaturing the fuel salt. That meant adding lots of extra 238U. Second, they avoided fuel processing until the fuel salt wouldn't burn anymore. If we put back in some processing and depended on fission product self-protection of the hot fuel then I suspect we could achieve unity breeding starting on 20%LEU.
From ORNL's DMSR it appears most definately that you can't isobreed on LEU startup alone. Take a look at ORNL -DWG 80-4267 from the main DMSR document:

http://www.energyfromthorium.com/pdf/ORNL-TM-7207.pdf

The conversion ratio starts at 0.8 and never gets above 0.85. Possibly with processing you could go up to 0.9 or something but getting to 1.0 is a big stretch. Looking at the French work they seem to confirm this, LEU isn't good enough even for their faster spectrum that gets more out of the plutonium. So the French suggested to start with reactor grade plutonium for the startup. That of course is expensive for a faster spectrum and gets into issues with trifluoride solubility. So I think a DMSR but fuelled with reactor grade plutonium and a well thermalized spectrum, without the denaturing constraint (which makes no sense for this no-fissile processing reactor at all), will get us off to a good start with a decent trade-off between waste burning, economics and RD&D needs.

Post Reply