Energy From Thorium Discussion Forum

It is currently Jun 24, 2018 5:51 am

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 30 posts ]  Go to page Previous  1, 2
Author Message
PostPosted: Jun 15, 2009 7:37 am 
Offline

Joined: Apr 24, 2008 4:54 am
Posts: 490
Location: Columbia, SC
This would seem to be easily solvable by simple engineering. I.e., put a nice big flywheel on your fuel pump such that it might take 5 minutes or more to coast down to half its initial speed on a loss of pumping power. You could also engineer in a shutdiwn signal to drop in your shutdown control rod(s) on a low pump shaft speed signal. Very simple solution to the positive reactivity effect of the delayed neutron emission in a moving fluid fuel system.


Top
 Profile  
 
PostPosted: Jun 15, 2009 2:43 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3065
I agree this particular problem will readily be solved - in fact, I don't think it is really a problem at all - we just need to anaylse the ramp down time of the pump. The natural inertia of the pump and fluid (after all the salt has some significant mass) will likely be quite sufficient without any extra steps.


Top
 Profile  
 
PostPosted: Oct 08, 2011 3:46 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5048
The paper by Eric Ottewitte provides important insights into the delayed neutron issue:

http://www.energyfromthorium.com/pdf/EIR-411.pdf

This is for fast fission of Pu239, a different beast, but still, some important issues arise.

First is the sparging rate. Most of the delayed neutrons come from decay of volatile/gaseous fission products and have very short half lives. This means if you sparge rapidly the effective delayed neutron fraction is reduced. In Ottewitte's reactor, most of these elements were lost with a 10 second sparge, but were very low at a 1000 second sparge.

Doppler is mostly useless to a chloride fast reactor - neutrons just don't go to those low speeds at all - so clearly Ottewitte was worried about keeping the delayed neutrons, and maybe it's not so bad for a LFTR. But still, sparging rate effect on delayed neutron retention should be investigated in detail for thermal fission of U233. The precursors are different and with different half lives.

Second is the decay time of these neutron precursors. Very short! Something like 1/3 to 1/2 has a half life that is shorter than the core residence time of the salt. This raises the issue of core flow rate effect on effective delayed neutron retention. For a big core like ORNL's DMSR the salt moves at a low speed of 0.5 m/s and with an 8 meter core that gives 16 second residence time. This means almost all of the I-138, I-139, Br-89 and Br-90 precursors will decay in the core, with their half-lives of 1.6-6.3 seconds. But a core with a residence time much shorter will lose most of these neutrons.


Top
 Profile  
 
PostPosted: Oct 08, 2011 11:10 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3065
Not most. To illustrate this lets take a case where the residence time is 1 second, the out of core time is 1 second, and the decay time is 10 seconds.

Picture an exponential decay curve with a 10 second half life.
First imagine a batch of nuclides born at the entry to the core. For the first second the decays will be in core - integrate the area under the decay curve from 0 to 1 as neutrons in core. For the second second the decay happens out of core - integrate the area under the decay curve from 1 to 2 as neutrons out of core. From 2 to 3 are in core, from 3 to 4 are out of core, etc.

For the nuclides born at the entrance to the reactor more decay in core than out. As the circulation rate increases the ratio of neutons out of core / total neutrons approaches the ratio out of core residence / total circulation time.

For the nuclides born at the exit to the reactor more decay happens out of core than in.

Fortunately exponential decay is very easy to integrate.

I did this and evaluated it over a broad range. The out of core residence / total circulation time is a pretty good upper bound for the neutrons lost. The precise number is somewhat better.


Top
 Profile  
 
PostPosted: Oct 08, 2011 12:36 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1947
Location: Montreal
Lars wrote:
The out of core residence / total circulation time is a pretty good upper bound for the neutrons lost.
Useful analysis -- thanks Lars.

The situation may get more complicated, depending on where we put the *outer* neutron reflector.

For example, in the French design, we all read about how the top & bottom of the cylindrical core are to be equipped with ZrC reflectors (in addition to the usual radial reflector).
For sure, once fuel flows out past the top reflector and to the HX, any delayed neutrons emitted are lost.

But suppose that the top (and bottom) reflector is moved axially away from the core, so as to allow some space for the recirc plumbing ?
....and suppose we also put a sufficiently large gap between the core tank and the radial reflector, to accommodate the HX units ? (...in more thermalised reactors than the French fast fluoride MSR, that would mean a gap between the moderator cylinder and the outer reflector cylinder)

In such a configuration, the delayed neutrons emitted in the HX & recirc plumbing end up, after a few bounces, getting reflected back into the open top (and bottom) of the reactor.
That also goes for the delayed neutrons emitted in the radial space anywhere along the height of the core cylinder - regardless of whether there is a moderator or not - since the space communicates with the one above & below the reactor core.
In the radial moderator case, even neutrons that come in radially inward and are subsequently bounced back out, end up getting shunted to the open core ends by the outer reflector sending them axially up or down the gap between the concentric cylinders.

Of course this works both ways: the open top & bottom of the core will be very leaky, so the overall picture is one where lots of neutrons will be circulating in the gap - top & bottom, as well as radial - surrounding the reactor core.

The only important part is that the outer reflector and the HX & piping equipment installed in the gap have very low neutron absorbtion x-section, so that the vast majority of neutrons bouncing around in the gap eventually find their way to the highly-absorbing fuel salt exposed at the top & bottom.

An interesting aspect of this geometry is that, in principle, it doesn't matter how big the gap is: Any size is OK, as long as the outer reflector is effective (both in terms of material and any neutron leak paths) -- and as long as its overall size is "reasonable", so that it doesn't impact the overall hardware cost too much.

However, one supposes that a bigger gap is only needed to accommodate bigger HX & piping equipment.
That would likely have a negative effect, since a greater number of delayed neutrons would tend to get absorbed in the fuel salt flowing through the HX, rather than making their way back to the core.
Depending on the neutronic coupling of the HX's to the core (greater for unmoderated tank than for core surrounded by moderator), the impact of the HX absorbtion may be more-or-less significant (the extreme case is of course one where the core *is* the HX -- with a stagnant fuel salt calandria having channels for secondary-side coolant....).

In either case, there is undoubtedly some geometric optimisation to be done.... For example, a high-aspect ratio core (i.e. tall & narrow) will of course be less leaky at the ends - while still acting "like a magnet" for neutrons emitted in the gap.
There are limits to this approach though -- both because of fuel channel volumetric power and delta-T across the length of the core, as well as possible core neutron flux instability at the more extreme aspect ratios.


Top
 Profile  
 
PostPosted: Oct 08, 2011 2:23 pm 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3065
The extreme value of the delayed neutrons is not to convert fertile to fissile (although that is good). Their extreme value is in making control of the reactor easier as we can operate the reactor slightly below prompt critical and the reactor time constant increases substantially.

Those neutrons are only extremely valuable if they cause fission in the core. I don't imagine the fissions being caused inside the HX would be extremely valuable for control purposes.

But this idea isn't completely foreign. It is pretty close to putting the HX on the outer edge of a blanket and a bit less extreme than the ideas for putting the HX on the edges of the core itself.

For a molten salt reactor the accident scenario where the pumps stop and we get the sudden reactivity increase of having all the delayed neutrons remain in the core is the one we will have to pay the most attention to. I hope it can be reasonably solved by the inertia of the pumps and flowing fuel.


Top
 Profile  
 
PostPosted: Oct 08, 2011 3:01 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1947
Location: Montreal
Lars wrote:
But this idea isn't completely foreign. It is pretty close to putting the HX on the outer edge of a blanket and a bit less extreme than the ideas for putting the HX on the edges of the core itself.

I was thinking single-fluid MSRs.

For two fluids, with a blanket, I presume that the blanket material would be too dense to allow any significant HX delayed neutron coupling to the core -- except of course unless there was another reflector and a gap to shunt the neutrons back to some exposed part of the core, as described above.


Top
 Profile  
 
PostPosted: Oct 08, 2011 3:04 pm 
Offline

Joined: Apr 24, 2008 4:54 am
Posts: 490
Location: Columbia, SC
A nice heavy flywheel should be no problem, and give you a coast down of a couple of minutes easily.

From the AP1000 description documents:
Quote:
A flywheel, consisting of two separate assemblies, provides rotating inertia that increases the
coastdown time for the pump. Each flywheel assembly is a composite of a uranium alloy flywheel
casting or forging contained within a welded nickel-chromium-iron alloy enclosure. The upper
flywheel assembly is located between the motor and pump impeller. The lower assembly is located
within the canned motor below the thrust bearing. Surrounding the flywheel assemblies are the
heavy walls of the motor end closure, casing, thermal barrier flange, stator shell, or main flange.


Also look at page 103 of http://www.studentpipeline.org/afci/ms/theses/petroski.pdf


Top
 Profile  
 
PostPosted: Oct 08, 2011 5:29 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5048
Lars, just to make some things clear here.

First the sparging rate issue, is what makes it worse. The delayed neutron precursors with longer precursor half lives suffer most here in terms of delayed neutron loss - simply because they tend to be removed rather than decay to give up the neutron in the salt. For example with a 10 second sparge, more than 80% of Br-87 and I-137, and 70% of Br-88, are lost.

Check out figure 5.5 on page 40:

http://www.energyfromthorium.com/pdf/EIR-411.pdf

Second is the issue that makes it better: the delayed neutron emitters with shorter precursor half life tend to give up their neutron in the core. Yes its true that neutrons are also made in the top of the reactor, but you can take the mean residence time to solve that and if your delayed neutron precursors have half lives much shorter than that then they will mostly decay in the core.

It should be considerably better for a long residence time core flow, perhaps by as much as 1/4 improved retention of delayed neutrons. The residence time in the heat exchanger is MUCH shorter than that of the core because its flow speed of the former is much higher. Considering your numbers, this makes the situation much better than just the volume makes it appear.

So an interesting question becomes, if you do gentle sparging in a long fuel residence time core (big core) how much bigger can you make the out of core inventory? Or, with the same out of core inventory, how much better does the reactor behave?

If we normally assume that 1/3 can be out of core neutrons, then with a flow speed 3x higher it means we can make the HX as big as the core, rather than just 1/3 the size! (of course we still want to minimize fissile inventory but the delayed neutron losses seem much less of a problem now).

Bob, thanks, would be great if we can use a flywheel and also hopefully the canned motor design from the AP1000. We don't need such high pumping power, several times less than a PWR per electrical megawatt generated, so it is likely that even one such an AP1000 pump is way too much pumping power. So we can do with a smaller flywheel as well. On the down side, we don't yet have fluoride lubricated bearings. I think these have to be made of dissimilar contacting materials to avoid 'contact welding' that someone here pointed out recently. Possibly SiC ball bearings in a Hastelloy N casing.


Top
 Profile  
 
PostPosted: Oct 08, 2011 6:30 pm 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1947
Location: Montreal
Cyril R wrote:
So an interesting question becomes, if you do gentle sparging in a long fuel residence time core.....

Another question might be: How many loops of the core-HX circuit does a typical delayed neutron precursor do, before it actually exits the fuel salt and is captured outside ?

I'm thinking specifically of the case where fuel salt is NOT bled to a separate sparging equipment (where the delayed neutron precursors would obviously be out of the reactor-HX loop, guaranteeing 100% loss of those neutrons), but rather the case where the off-gases are captured by a fume hood above the core, after the precursor has looped around a number of times and hopefully decayed, to yield its delayed neutron....


Top
 Profile  
 
PostPosted: Oct 09, 2011 12:07 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3065
Cyril R wrote:
Lars, just to make some things clear here.

First the sparging rate issue, is what makes it worse. The delayed neutron precursors with longer precursor half lives suffer most here in terms of delayed neutron loss - simply because they tend to be removed rather than decay to give up the neutron in the salt. For example with a 10 second sparge, more than 80% of Br-87 and I-137, and 70% of Br-88, are lost.

Check out figure 5.5 on page 40:

http://www.energyfromthorium.com/pdf/EIR-411.pdf

I read earlier that by sparging at 60 seconds few of the delayed neutrons were lost to the sparge. From the figure 5.5 it looks like we might want it to be a bit longer - perhaps 2 or 3 minutes. A faster sparge means quicker removal of tritium and spares some neutrons Xe capture so there has to be a balance. But I'd guess that keeping the delayed neutrons should have the higher priority here. If we have delayed neutrons being lost to sparging then we have to operate closer to prompt critical and in an accident situation those delayed neutrons don't disappear so we have a reactivity increase.
Quote:
Second is the issue that makes it better: the delayed neutron emitters with shorter precursor half life tend to give up their neutron in the core. Yes its true that neutrons are also made in the top of the reactor, but you can take the mean residence time to solve that and if your delayed neutron precursors have half lives much shorter than that then they will mostly decay in the core.

True enough - one of several good things that come with a larger core - all balanced by the cost of the fissile.
Quote:
It should be considerably better for a long residence time core flow, perhaps by as much as 1/4 improved retention of delayed neutrons. The residence time in the heat exchanger is MUCH shorter than that of the core because its flow speed of the former is much higher. Considering your numbers, this makes the situation much better than just the volume makes it appear.

Yes the out of core / total fuel inventory is an upper bound.
Quote:
So an interesting question becomes, if you do gentle sparging in a long fuel residence time core (big core) how much bigger can you make the out of core inventory? Or, with the same out of core inventory, how much better does the reactor behave?

You can probably make the out of core inventory as big as you care to pay for the fissile there. The fissile represents a pretty big portion of the initial cost of the reactor so you don't want to go overboard. The reactor probably behave very well already. Yes it will behave better with more delayed neutrons in the core during normal operations but likely it is excellent already.


Top
 Profile  
 
PostPosted: Oct 09, 2011 12:15 am 
Offline

Joined: Jul 28, 2008 10:44 pm
Posts: 3065
USPWR_SRO wrote:
A nice heavy flywheel should be no problem, and give you a coast down of a couple of minutes easily.

From the AP1000 description documents:
Quote:
A flywheel, consisting of two separate assemblies, provides rotating inertia that increases the
coastdown time for the pump. Each flywheel assembly is a composite of a uranium alloy flywheel
casting or forging contained within a welded nickel-chromium-iron alloy enclosure. The upper
flywheel assembly is located between the motor and pump impeller. The lower assembly is located
within the canned motor below the thrust bearing. Surrounding the flywheel assemblies are the
heavy walls of the motor end closure, casing, thermal barrier flange, stator shell, or main flange.


Also look at page 103 of http://www.studentpipeline.org/afci/ms/theses/petroski.pdf

If I recall correctly, the French analysis showed there was a problem if the coast down time was less than 100mSec and no problem if it was greater than 1 second. I'm thinking it is likely the pump and fluid themselves provide sufficient inertia but certainly a flywheel could solve this if they pump and fluid don't have sufficient inertia by themselves.


Top
 Profile  
 
PostPosted: Oct 09, 2011 2:50 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5048
100 milliseconds! I guess we have no issues then, as long as we don't do anything stupid such as put massive fast closing isolation/shutoff valves in the primary circuit.

But flywheels are great to have. Helps with starting up the reactor from zero power, too.

Ottewitte calculated a 7% loss of delayed neutrons for a sparging time of 100 seconds. And 25% lost in the external HX and piping.

Per Peterson wanted to use a low temperature drop compact heat exchanger, having a power density of 50 MW thermal per cubic meter. For 2250 MW thermal core power, that means 45 cubic meters. I previously thought that very large to apply to a MSR. But thinking about it, almost half the volume is occupied by secondary salt, and there is the HX metal, so only about 20 cubic meters would be fuel salt. Then with, say, twice the flow speed in the HX compared to the reactor, that's equivalent to 10 cubic meters core equivalent in out of core fraction. So this HX should work even for a quite small core of 20 cubic meters.

Jaro's HW-MSR does need to work harder, with that small fuel salt inventory. It will help if he puts a carbon filter trap for Br and I between the siphons and the top reflector. I've talked to our active carbon supplier for the PSA units our clients have, and he says that's a tough environment for a carbon filter bed, primarily due to the high temperature. He said ionizing radiation wasn't a big problem but was not too confident about neutrons.


Top
 Profile  
 
PostPosted: Oct 09, 2011 5:48 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5048
By the way, how do the n,2n neutrons fit into the delayed neutron/reactor control equasion? These aren't lost in the external piping so that makes it better, right? As in, fewer prompt fission neutrons are needed to sustain the chain reaction?


Top
 Profile  
 
PostPosted: Oct 09, 2011 8:56 am 
Offline
User avatar

Joined: Nov 30, 2006 9:18 pm
Posts: 1947
Location: Montreal
Cyril R wrote:
HW-MSR does need to work harder, with that small fuel salt inventory.

If we operate with NU - which is dirt cheap - we could easily afford to have a LARGE fuel salt inventory in the HX's: Having a small fuel salt inventory in the core is only desirable for making the whole core very small -- less HW, fewer fuel channels to replace periodically, etc.

If we make sure that the delayed neutrons emitted from the HX's are not lost -- i.e. an outer reflector keeps them all in -- then we could have nice low-delta-P HX's and minimum pumping power.....

I'm guessing that a good configuration might be a single HX per reactor, with a fairly high-aspect ratio rectangular x-section, and with the long axis of the x-section oriented radially to the core cylinder (for minimal reactivity).
Sort of like the letter "i" - with a big dot (call it the "iReactor" :lol: )
That way, the delayed neutrons would have a good chance of getting out of the HX, and into the gap separating the core & HX from the outer reflector.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 30 posts ]  Go to page Previous  1, 2

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