Energy From Thorium Discussion Forum

It is currently May 25, 2018 1:46 pm

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 41 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
PostPosted: Aug 13, 2009 10:48 pm 
Offline
User avatar

Joined: Dec 19, 2006 11:01 am
Posts: 395
Location: Knoxville, TN
David wrote:
Is that the right report # ORNL 3913? That report is a collection of chemistry papers.
My mistake, this was another reference in the paper.
The reference case scenario is ORNL-4528: Two-Fluid Molten-Salt Breeder Reactor Design Study

Quote:
I'm just curious as I've never seen Oak Ridge quote as low as 2.6 years even for their Two Fluid work. I assume they were talking about ORNL's Two Fluid design of interlacing fuel and blanket salts which evolved but remained about the same from 1965 or so to 1968. I guess that depends on what they assume is "ideal reprocessing", maybe they are also assuming lower starting inventories.
Indeed, U233 critical mass in the optimized design is nearly half of what it is in the reference case.

Quote:
The "plumbing problems" of interlacing fuel and blanket salts within the core is probably still nearly insurmountable if for no other reason than the shrinking and expanding graphite tends to dramatically change the ratio of the two fluids within the core (and the work on using metal looked even more hopeless). Hopefully we can convince them of the merits of a new simple geometry of tube within tube that only needs one barrier, not thousands. The old ORNL Two Fluid design also has a positive temperature reactivity coefficient for the blanket salt which they rarely mentioned. If your blanket salt is only surrounding a core of fuel salt, it ends up to also have a negative temp coefficient since it acts as a reflector (hotter=less dense=less neutrons reflected back) David L.
The graphite lifetime seems to be an issue, 1.42 years in the optimized design compared to 1.32 years in the reference case. It is defined as a time during which the six most central graphite fuel channels exceed 2.9e22 cm^-2 fluence for neutrons with energy >180keV.
The paper actually discusses the issues with the blanket salt.
Perhaps make a visit to Prague and present the tube in shell concept?


Top
 Profile  
 
PostPosted: Aug 14, 2009 5:23 am 
Offline
User avatar

Joined: Dec 19, 2006 11:01 am
Posts: 395
Location: Knoxville, TN
Here is the paper attached: Neutronic Analysis of Two-Fluid Thorium Molten Salt Reactor, J. Frybort, R. Vocka (NRI).


Attachments:
File comment: Neutronic Analysis of Two-Fluid Thorium Molten Salt Reactor, J. Frybort, R. Vocka (NRI).
9384 Paper ICAPP09.pdf [103.64 KiB]
Downloaded 1811 times
Top
 Profile  
 
PostPosted: Aug 14, 2009 10:20 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3068
Thanks for the paper.


Top
 Profile  
 
PostPosted: Aug 14, 2009 12:04 pm 
Offline

Joined: Mar 30, 2008 8:37 am
Posts: 145
Location: Stumbling out of a van, head full of zombie
How are doubling time, breeding ratio, and fraction of fuel ex-core related? Is there a closed form approximation?

_________________
Circular logic - see self-referential

Self-referential - see circular logic

imbibo ergo fnord


Top
 Profile  
 
PostPosted: Aug 14, 2009 4:35 pm 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
fnord wrote:
How are doubling time, breeding ratio, and fraction of fuel ex-core related? Is there a closed form approximation?

You also need to define the power density (watts/litre) and fissile concentration, or more directly watts (thermal) per kg of fissile, which gives you the rate of fissile consumption and hence fissile production rate. The baseline guestimate usually used here is ~2000 -2500 MW(th)/Te or 2-2.5MW/kg of fissile, which means it burns through its inventory in about a year. If the breeding ratio is 1.05 it takes (Burnup time)/(breeding ratio-1) = (1 year)/(0.05) = 20 years to generate a new startup load, i.e. doubling time 20 years.

The Czechs think they can do much better on all counts. Very low fissile concentration, and less out of core, lowers fissile inventory, raising the power / fissile to ~10 MW/kg, dropping burnup time to only 3 months. Optimised reprocessing and core geometry push the breeding ratio up to ~1.1. Doubling time is now ~ 3 months/(1.1-1) ~ 30 months.

Luke


Top
 Profile  
 
PostPosted: Aug 14, 2009 5:19 pm 
Offline

Joined: Jun 12, 2008 3:32 pm
Posts: 57
Luke wrote:
fnord wrote:
How are doubling time, breeding ratio, and fraction of fuel ex-core related? Is there a closed form approximation?

You also need to define the power density (watts/litre) and fissile concentration, or more directly watts (thermal) per kg of fissile, which gives you the rate of fissile consumption and hence fissile production rate. The baseline guestimate usually used here is ~2000 -2500 MW(th)/Te or 2-2.5MW/kg of fissile, which means it burns through its inventory in about a year. If the breeding ratio is 1.05 it takes (Burnup time)/(breeding ratio-1) = (1 year)/(0.05) = 20 years to generate a new startup load, i.e. doubling time 20 years.
Nitpick: with compound interest, the doubling time is about 15 years. (2 = 1.05^14.2)
That mightn't apply at first; 1/20th of a minimum fuel load doesn't get you 1/20th of a reactor. But if you've got twenty 250MW(th) reactors....


Top
 Profile  
 
PostPosted: Aug 14, 2009 6:28 pm 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
ondrejch wrote:
Here is the paper attached: Neutronic Analysis of Two-Fluid Thorium Molten Salt Reactor, J. Frybort, R. Vocka (NRI).

Fantastic performance from their optimised design, although at the cost of a lot of waste graphite. Their 'safety issue'
section is a bit worrying though. Not because the problem isn't solvable - I think you can fix most of it just by putting the whole reactor in a vat (Hastelloy?) of fertile salt. and having the salt inlet and outlet at the top - but because they've described the 'loss of fertile salt' accident as being equivalent to a positive void coefficient. If the regulators take the same view, they will never let it be built.

Luke


Top
 Profile  
 
PostPosted: Aug 15, 2009 1:35 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3068
They stated that they did not have the composition of the fuel change with time because the reprocessing would keep things clean. I wonder if they assumed the uranium stayed u233 or if they have an equilbrium composition of uranium nuclides. It makes a big difference. If the current calculations assumed the composition stayed pure u233 then just the uranium nuclide evolution would change the doubling time to 23 years - without losses to protactinium or fission products. I send an email to ask them.


Top
 Profile  
 
PostPosted: Aug 15, 2009 8:35 am 
Offline

Joined: Dec 03, 2008 5:23 pm
Posts: 137
Location: Oak Ridge, TN
There are two aspects that are given them such a low doubling time: the high breeding gain (BR-1) and the very high specific power (7x that of the 2-fluid ORNL MSBR). The doubling time is inversely proportional to both of these values. When they include the fission products, U transmutation, Pa-233 captures, fissile losses in processing, they could probably at best achieve a similar result as the ORNL design. However, with their high specific power, more Pa-233 may be getting eating up. They could have tallied the Pa-233 capture rate in their model to get an estimate of this impact, even without performing a full burnup analysis.

However, if they can achieve such a high specific power, they will still have a low doubling time. The ORNL 1968 MSBR design kept the power density lower so that they could get a reasonable graphite life (8 years). There may be other thermal considerations as well. Earlier ORNL designs had a higher specific power and a lower doubling time.

Regarding the issue of the blanket salt draining and the fuel salt remaining, I need to research this more, but I would assume that the design would not allow dumping of the blanket without dumping the fuel salt. I cannot believe that this was overlooked in the MSBR design.


Top
 Profile  
 
PostPosted: Aug 15, 2009 12:21 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3068
Wouldn't the losses to Pa capture be proportional to the macro cross - sections of thorium and Pa?

So suppose we had 10 m^3 of fertile salt in circulation and had Y losses to Pa capture.
If we increase the fertile salt volume to 20 m^3 then the Pa capture losses would be Y/2.
To keep the neutronics the same the extra 10 m^3 would be added to the portion of the circulating loop that is outside the core.

I think this holds true whether the Pa is extracted or not.


Top
 Profile  
 
PostPosted: Aug 15, 2009 5:51 pm 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
Lars wrote:
Wouldn't the losses to Pa capture be proportional to the macro cross - sections of thorium and Pa?

So suppose we had 10 m^3 of fertile salt in circulation and had Y losses to Pa capture.
If we increase the fertile salt volume to 20 m^3 then the Pa capture losses would be Y/2.
To keep the neutronics the same the extra 10 m^3 would be added to the portion of the circulating loop that is outside the core.

I think this holds true whether the Pa is extracted or not.

It depends on what you take as the limiting factor for the extraction system. If it is kg salt, or kg heavy metal processed/day, then the Pa concentration is fixed. You process 10 Te salt/day, the Pa concentration is 1 day's production per 10 Te salt. Diluting the system just increases the Pa residence time.

If the extraction is limited by kg Pa extracted per day (independent of concentration) then your supposition is correct. Since we know nothing about their proposed Pa extraction system, we don't know its limits - but any chemical processing system that is trying to extract a dilute component is likely to be limited more by the bulk it has to process than by the difficulty of handling the relatively small quantity of extracted material.

Luke


Top
 Profile  
 
PostPosted: Aug 15, 2009 6:45 pm 
Offline

Joined: Jul 28, 2008 5:01 am
Posts: 462
Location: Teesside, UK
Lars wrote:
..... If the current calculations assumed the composition stayed pure u233 then just the uranium nuclide evolution would change the doubling time to 23 years - without losses to protactinium or fission products. I send an email to ask them.

What am I missing here?

U233 fission ~90% of absorptions, capture ~ 10%, average ~2.25 neutrons out/neutrons in.
To keep constant U234 inventory, it must all capture. stealing another 0.1 neutrons per U233 fission, forming U235
U235 takes another 0.1 neutrons, but ~85% of the time it fissions, on average returning 0.2 neutrons, so we're back to where we started, except that the U235 is also contributing reactivity. To keep in balance, there must be a bit less U233. Since ~10% of the fissions are of U235 rather than U233, the average neutron yield is ~ (2.25*0.9 + 2*0.1) = 2.225, or 0.025 down from pure U233
The ~15% of the U235 that doesn't fission then takes another 2 captures to become Pu238, which will go out with the fission products. This steals 2 * ~0.15 * ~0.1 = ~0.03 neutrons per U233 fission.
Altogether we are down ~0.025 + ~0.03 = ~0.055 neutrons per U233 fission. This drops their breeding ratio from 1.103 to 1.048, so the doubling time goes up by 103/48 to 5 years 8 months.

What are the other losses?

Confused


Top
 Profile  
 
PostPosted: Aug 15, 2009 7:26 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3068
Luke wrote:
Diluting the system just increases the Pa residence time. If the extraction is limited by kg Pa extracted per day (independent of concentration) then your supposition is correct.

I tend to think of no Pa extraction and in that case diluting the system does indeed reduce the Pa capture losses. If there is an extraction system and the processing speed is set by the amount of Pa extraction the same holds.

You are right, if the Pa extraction system is limited by either the thorium or salt volumes then the gain is smaller. In the limit for fast processing, if your processing rate is fast enough to pull out all the Pa before any decays or captures then it makes no difference at all. In the limit for slow processing it is equivalent to no processing and the loses are cut in half. For realistic systems, even those chemical processes limited by the thorium content adding an outside holding cell will reduce Pa losses.


Top
 Profile  
 
PostPosted: Aug 15, 2009 7:42 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3068
Luke wrote:
U233 fission ~90% of absorptions, capture ~ 10%, average ~2.25 neutrons out/neutrons in.
To keep constant U234 inventory, it must all capture. stealing another 0.1 neutrons per U233 fission, forming U235
U235 takes another 0.1 neutrons, but ~85% of the time it fissions, on average returning 0.2 neutrons, so we're back to where we started, except that the U235 is also contributing reactivity. To keep in balance, there must be a bit less U233. Since ~10% of the fissions are of U235 rather than U233, the average neutron yield is ~ (2.25*0.9 + 2*0.1) = 2.225, or 0.025 down from pure U233
The ~15% of the U235 that doesn't fission then takes another 2 captures to become Pu238, which will go out with the fission products. This steals 2 * ~0.15 * ~0.1 = ~0.03 neutrons per U233 fission.
Altogether we are down ~0.025 + ~0.03 = ~0.055 neutrons per U233 fission. This drops their breeding ratio from 1.103 to 1.048, so the doubling time goes up by 103/48 to 5 years 8 months.

What are the other losses?

Sorry, I miscounted and converted u234 to u235 twice.
But the other losses would be: 1) Pa capture and 2) fission product capture.

By the way, I think we could extract the 237Np instead of the 238Pu and recover 0.015 of the neutrons. This would assume that UF6 and NpF6 behave sufficiently differently that we could separate them by fractional distillation or similar method when we have them both as a gas in the fluorinator.


Top
 Profile  
 
PostPosted: Aug 17, 2009 11:23 am 
Offline

Joined: Mar 07, 2007 11:02 am
Posts: 914
Location: Ottawa
Thanks so much for posting the paper.

I've only skimmed it so far and the subsequent posts, I'll try to give more details later. They are certainly making a lot of assumptions to simplify things so I think they are giving a bit too much weighting to their "breeding factor". The approximations are a fine way to start but they probably shouldn't then start using this factor for estimating doubling times.

The simplifications will of course pretty dramatically lower the real breeding ratio. First, they assume no losses to fission products, the truth is, no matter how fast you can remove them you will still lose significant neutrons to them mainly because some of the rare earth FPs have huge cross sections so even if you can remove them every day or two they still steal lots of neutrons. They make the same assumption about no Pa losses which is OK I suppose since the cross section is much lower. Second as Lars points out, they assume that it is always just pure U233 where in reality the steady state usually comes about 2/3 U233 and the rest higher isotopes of U of which U236 is a neutron parasite with 0.4% of neutron absorptions in the ORNL Two Fluid (i.e. not fertile like U234 or fissile like U235).

I am also a bit confused about the low graphite lifetime and starting inventory. For the starting inventory they assume about the same total salt volume (about 20 m3) and fissile concentration as does ORNL but seem to get half the fissile inventory (342.5 kg versus about 700kg for ORNL). I wonder if someone just made a mistake and only quoted the fissile inventory in the core (i.e. half fissile is in the core and the other half is in the pipes and heat exchanger). It is technically correct to say that the critical inventory is 342.5kg but if you are calculating a doubling time you have to include the other 342.5kg outside the core in the heat exchanger.

The very short graphite lifetime is also still a bit of a puzzle to me, they seem to be keeping everything very close to ORNL's design but somehow get a much lower graphite lifetime. I wonder if it is something about flux flattening or the lack thereof. I've only skimmed things so far though, maybe I'm just missing something.

I also agree with other posts and the paper in rightly worrying about the safety issues of losing blanket salt (something ORNL didn't really mention much). That is one of the huge benefits of switching from the old idea of interlacing fuel and blankets salts within the core and the new concept of only having the blanket salt outside a small diameter core (and going to a modestly long cylinder or tube if you want to get a healthy total core power). With the blanket only outside the core, it is acting as a very weak reflector of neutrons back into the core, thus if it drains away or gets less dense by heating up then you actually lower reactivity because the core is now losing some of these reflected neutrons. As a point of clarification, this assumes that you don't put any sort of good reflecting material within or outside the blanket zone because you could then also end up with a positive void/temp coefficient for the blanket salt, i.e. you can have graphite in the core but avoid using much in the blanket or at the outer vessel wall.

All in all though it is very exciting to see modern modeling of the Two Fluid design, I'll try contacting them shortly (something I should have done long ago...)

David LeBlanc


Last edited by David on Aug 18, 2009 8:47 am, edited 1 time in total.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 41 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group