an exciting new concept - the fluoride salt based "fast" MSR

Threads relating to the design of liquid-fluoride reactors.
Lars
Posts: 3060
Joined: Jul 28, 2008 10:44 pm

Re: an exciting new concept - the fluoride salt based "fast"

Post by Lars » Dec 05, 2013 1:41 pm

My recollection is that the challenge is helium embrittlement. The helium is generated in a two neutron capture process so it will follow a quadratic law in time for a given neutron flow. There are two distinct paths for this. One for slow neutrons and a different one for fast neutrons. Some chemistry is involved in trying to trap the helium where it is generated. Naturally it will migrate to grain boundaries which makes the Hastalloy tolerance surprisingly low unless this is addressed. I believe toward the end of the program ORNL did find a way to change the minor elements in the hastalloy to help prevent the helium from migrating to the grain boundaries.

The French and Russians have given hints of different vessel compositions to reduce the problems. But I've seen no data so I'm thinking the plan for a 1.5 or 2 fluid reactor should be that the first wall is considered as flow control. It should not be considered a fission product barrier nor a structural element. I would plan on keeping its shape simple and replacing it every four years or so. As I mentioned the process is quadratic so making it last the lifetime of the reactor (say 64 years) is 256 times as hard as making it last four years.

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Cyril R » Dec 05, 2013 1:52 pm

darryl siemer wrote:This discussion has drifted off topic again -the MSFR would not have carbon in its core. I'd like to try to get us back on track by asking if anyone knows how to go about predicting the number of "displacements per atom" for real-world nickel as a function of time, neutron flux, & neutron speed. Are you aware of any experimental work relevant to estimating how long we could expect an optimized MSFR's "barrier" to last?
It has been theoretically calculated for the MSFRs reflectors.

http://www.gen-4.org/GIF/About/document ... elpech.pdf

They cannot even get the inner reflector to last 20 years. They are considering adding tungsten to improve the resistance, obviously a major metalworking and neutronics penalty there...

The flux through the barrier will be 10-20x higher so maybe you can get 1 year out of a nickel barrier reliably and maybe 2-3 years if you like risks.

The French group is using graphite as the barrier. They appear to be oblivious as to the plumbing and fragility problems of using graphite. A remarkable oversight, considering it was the main difficulty that ORNL experienced with two fluid designs and indeed the difficulty that led them to consider single fluid only in later phases of their multi billion dollar, multi brilliant scientists programme...

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 05, 2013 3:29 pm

fab wrote:Sorry for the off-topic.
I don't know if it ....
And in pages 11 ; 12 ; 13 and 14 :

http://www.gedeon.prd.fr/ATELIERS/mars_ ... ucotte.pdf


Thanks very much. It looks like MS Lucotte has given me what's needed to support my contention that routine replacement ("maintenance") of a properly designed reactor's core tank would obviate its durability "issue". Her slide 11 points out that there would only be 0.75 dpa worth of damage/year at the center surface of the reference MSFR's axial metallic reflectors which, are probably assumed to be made with or plated with nickel. Since that point on its reflectors should experience roughly the same neutron flux as would the core tank of a more "advanced" reactor, and graphite is still just shrinking* (not swelling yet) with 2 dpa's worth of damage, it seems reasonable to expect that a core tank made of/with natural nickel should last at least 2-3 years.

Again, the only way to actually prove such things is to do tests. Does DOE still have a "lead NE Lab"?

* http://web.ornl.gov/sci/physical_scienc ... EADJ~2.PDF
Darryl Siemer

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Lars » Dec 05, 2013 3:35 pm

The flux through the barrier is significantly different if you are a 1.5 or 2 fluid reactor. A isobreeding reactor using 2 fluids implies roughly 1/2 of the neutrons goes through the barrier. With a 1.5 fluid reactor you can have 80-90% of the fertile captures happen in the core rather than traveling through the barrier so using a 1.5 fluid you can increase the life of the barrier (at the cost of either lots of thorium waste or considerable chemical challenge in separating the fission products from the thorium).

I think the idea of making the first barrier easy to change easy. Hence, I think of it as a flow guide. It doesn't have to be leak proof, it doesn't have to be strong.

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Cyril R » Dec 05, 2013 4:56 pm

Damage with nickel is very different than damage with graphite. Graphite damage is about physical dislodging of atoms in the lattice; with nickel the limiting factor is helium generation. This makes nickel considerably more neutron sensitive than graphite.

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 05, 2013 7:53 pm

Cyril R wrote:Damage with nickel is very different than damage with graphite. Graphite damage is about physical dislodging of atoms in the lattice; with nickel the limiting factor is helium generation. This makes nickel considerably more neutron sensitive than graphite.
What does "neutron sensitive" mean? Does it render Ni plating leaky? How leaky? How soon?
Darryl Siemer

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Lars » Dec 05, 2013 9:28 pm

It makes the metal brittle. If the helium migrates to grain boundaries and you put stress on the first wall (say by draining the blanket but not the fuel salt) then the wall might break. IIRC they added a tiny bit of carbon which distributes throughout the metal rather than at the grain boundaries. The helium would migrate to the carbon and stay. This dramatically increased the Hastalloy resistance to helium embrittlement. But we still have a problem.

So, my approach is two fold.
First take away the stress. By making it an operational requirement that the fuel and blanket salts have about the same density and when draining one you drain the other you remove the stress on that first wall. It isn't a fission product containment in the first place since there will be some fission in the blanket (not much but some). If we don't depend on that first wall to serve any structural purpose other than to guide the fuel salt flow (and not shed chunks into the salt) then we reduce the requirements on the wall significantly. This allows us to make the wall thinner - which reduces waste and neutron losses.
Second, replace it periodically. This problem is quadratic in time so cut the time in half you reduce the problem four fold.

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 06, 2013 4:31 pm

Lars wrote:It makes the metal brittle. If the helium migrates to grain boundaries and you put stress on the first wall (say by draining the blanket but not the fuel salt) then the wall might break. IIRC they added a tiny bit of carbon which distributes throughout the metal rather than at the grain boundaries. The helium would migrate to the carbon and stay. This dramatically increased the Hastalloy resistance to helium embrittlement. But we still have a problem.

So, my approach is two fold.
First take away the stress. By making it an operational requirement that the fuel and blanket salts have about the same density and when draining one you drain the other you remove the stress on that first wall. It isn't a fission product containment in the first place since there will be some fission in the blanket (not much but some). If we don't depend on that first wall to serve any structural purpose other than to guide the fuel salt flow (and not shed chunks into the salt) then we reduce the requirements on the wall significantly. This allows us to make the wall thinner - which reduces waste and neutron losses.
Second, replace it periodically. This problem is quadratic in time so cut the time in half you reduce the problem four fold.
Do you (anyone?) have any concrete suggestions along the line of the sketches I've previously ATTACHED?
Darryl Siemer

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 06, 2013 5:00 pm

cyrill
The flux through the barrier will be 10-20x higher so ...
Why would the neutron flux experienced by the center of the reference reactor's reflectors be 10-20 x higher than that seen by its "barrier wall"? It's a right circular cylinder which strongly suggests that those points are equidistant from its "hottest" (central-most) zone. Anyone?

btw it's nice to be able to actually respond to "concerns" - technical journal Editors often don't give you a chance to do so when they consider your submission to be too politically incorrect.
Darryl Siemer

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 06, 2013 9:42 pm

Here's one more question.

I'm trying to ball park how much extra it would cost to make a MSFR's "barrier" out of nickel that doesn't have much 58Ni in it.

The biggest part of that extra difference boils down to an enrichment cost. The way I've gone about it is to assume that it'd be done with a gas centrifuge like the one that Dr. Forsberg discussed in his recent ANS talk (ATTACHED) & that doing it would cost about the same/SWU as does enriching U. If you plug in an initial "non 58Ni" conc of 30% & a tails (the product in this case) conc of 1% into http://www.wise-uranium.org/nfcue.html, depending upon how much raw Ni you choose to process, it looks like it's going to cost on the order of 6000 SWU's per tonne of <1%58-Ni product; however, it's more apt to be about twice that much considering that SWU/mass is probably inversely proportional to the molecular weight of the gas being separated & that figure would probably be a lot less for NiX than for UF6. If so, then assuming that a SWU costs $100, our barrier is made of Hastelloy N (about 75 wt% Ni) & weighs 4 tonnes, that would add about 4*.75 *6000* 100 or $3.6 million to the price of what would otherwise be a $20,000 barrier material cost .

If I've done something wrong in deriving this estimate, please point it out and work out/post what you feel is a more accurate figure.
Attachments
ANS 2013 Depleted Nickel Alloy View Transaction.pdf
(1.63 MiB) Downloaded 137 times
Darryl Siemer

jagdish
Posts: 2304
Joined: Apr 19, 2008 1:06 am

Re: an exciting new concept - the fluoride salt based "fast"

Post by jagdish » Dec 07, 2013 5:42 am

Corrosion and neutronic damage are among major challenges of fast MSR:-
http://info.ornl.gov/sites/publications ... b29596.pdf
The possible solutions are
SiC cladding tried out in the PBMR fuel pebbles might be useful for reactor vessel and piping.
Chloride salts with Cl37 may be more manageable.
The concept could be used to recycle used LWR fuel or run a fast Th-U233 breeder reactor. The latter would be a fast version of LFTR without the challenges of high isotopic purity 7Li or short life graphite moderator.

camiel
Posts: 358
Joined: Feb 28, 2011 10:10 am

Re: an exciting new concept - the fluoride salt based "fast"

Post by camiel » Dec 08, 2013 6:12 pm

darryl siemer wrote:Here's one more question.

I'm trying to ball park how much extra it would cost to make a MSFR's "barrier" out of nickel that doesn't have much 58Ni in it.

As I am reading the paper of Dr. Bloore & Dr. Forsberg, it occurs to me that isotope-separation technology is going to play a very important role if MSFRs are to be developed, not only for producing alloys with a low nickel-58 content for the barrier material, but also for producing the lithium-7. Although it possibly can be done, it is going to be a drawback in terms of economics.

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Cyril R » Dec 08, 2013 7:42 pm

I'm not sure if they are net economic drawbacks. Li salts have superiour heat transfer properties, which will reduce equipment cost, pumping power electric bills, and such. Plus the neutronics are better than other salts so a lot of additional valuable fissile is created. All things considered, Li7 may still be the cheapest option...

Might be similar for nickel enrichment, though there are possibly more attractive barrier materials such as silicon carbide based composites if we want to talk fancy.

darryl siemer
Posts: 317
Joined: Jan 21, 2008 9:12 pm
Location: idaho falls

Re: an exciting new concept - the fluoride salt based "fast"

Post by darryl siemer » Dec 09, 2013 12:32 pm

darryl wrote: I'm trying to ball park how much extra it would cost to make a MSFR's "barrier" out of nickel that doesn't have much 58Ni in it.
Since my much-simplified MSFR implementation scenario would discard roughly 1 kg of pure 7-Li/day, I've also been trying to come up with a cost number for it. As is all too often the case these days that search revealed that the USA has given up its ability to produce its own (US PWR's use/discard about 300 kg pure 7-Li per year - see http://www.gao.gov/products/GAO-13-716) & has therefore become totally dependent upon Russia & China. No realistic hard cost number seems to be available because that cost is both relatively small & totally determined by supply & demand in a business that is guaranteed a "reasonable" profit, not by the intrinsic cost of manufacture. However, the last time that a US manufacturer was planning to actually make it (circa 198...Benedict, Pigford & Levy, "Nuclear Chemical Engineering" 2nd Ed., p 639), it was supposed to cost $3/gram.

In other words, throwing away a kg of 7-Li/day would cost the owner/operators of a 1.5 GWe nuclear reactor about $3000/day.
Last edited by darryl siemer on Dec 09, 2013 2:11 pm, edited 1 time in total.
Darryl Siemer

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

Re: an exciting new concept - the fluoride salt based "fast"

Post by Cyril R » Dec 09, 2013 12:45 pm

You mean PWR, not BWR. BWRs don't use Li7.

Also, the 300 kg per year is for all US PWRs. So more like throwing away 45 dollars a day per PWR.

You can't even wipe the desks in the offices of a PWR for 45 dollars. Peanuts.

The total 300 kg/year market in the US would be swamped by even a single prototype MSFR.

I full well understand why an industry of less than a million a year can't survive domestically. You can't operate a domestic li7 facility for a million bucks a year. You couldn't run the sales and marketing office with a million dollars.

I also find the talk about a shortage for Li7 for PWRs to be complete nonsense. The amount used is tiny and can easily be replaced by sodium if necessary. The fact that it isn't suggests it isn't necessary.

Post Reply