Energy From Thorium Discussion Forum

It is currently Dec 15, 2017 8:58 am

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 63 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next
Author Message
PostPosted: Jul 12, 2015 4:57 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
Carbon first shrinks, but then swells with more irradiation, so I can't imagine this would work well unless for very low burnup fuel (maybe good for NU fuel but nothing like the high burnup TRISO fuels tested). It would first prestress and then tear itself loose. I'd still prefer all SiC here. Though we may be overly worried about a small problem. I imagine the reaction rate of C to CO would be very low at these CO2 coolant temperatures.


Top
 Profile  
 
PostPosted: Jul 12, 2015 5:31 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
It was a problem with the AGR but only over the years long life of the reactor (it is considered the most likely thing to finally close the reactor if the cracks in the moderator blocks don't do it first).
If methane suppression is able to keep the levels down to that rate we should be fine for fuel.

The main problem I see with TRISO is that the heavy metal content of typical graphite encapsulated (44% particle fill fraction) material is only 0.65g/cc.
Even if we assume the space occupied by graphite is empty (can we just sinter TRISO particles into a block? That should give them a bit more stability) then we are going to struggle to fit enough material in the fuel channel.

The fissile content is just too low (UO2 manages 9.2g/cc)

I don't think we are going to do much better than some sort of ceramic fuel really.
If we go with liquid metal bonding we can push the fuel rod diameters a bit higher which will reduce the effective amount of cladding metal in the fuel, if we accept 316SS as the fuel cladding.

EDIT: We could probably get away with TRISO if we allowed on load refuelling because it would permit fuel to be exhausted in a relatively short time - but it would have the problem we could produce a large volume of HLW.


Top
 Profile  
 
PostPosted: Jul 12, 2015 6:34 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
Direct cooled TRISO is a little different than the graphite matrix stuff.

For one thing you get basically close random packing (vibratory compacted random sphere bed), packing fraction 0.62-0.63 range.

For another you have direct cooling so enormous surface area, in stead of just some graphite channels... so you can get away with big increases in the kernel size. All this was investigated by Tsiklauri and others.

So turns out the optimal is 4-8x bigger kernels (2-4 mm range, say), this gets you quite a high UO2 fraction per sphere (0.5 or better). So you can actually get close to a 3 g/cc HM in the fuel zone with UO2 kernels.

Granted not as good as cylindrical fuel, but its pretty hard to get beyond 5 g/cc with that even though UO2 has 9 g/cc. You need space for coolant you know! The more so with gas reactor. In fact you may end up needing cooling fins to make up for poor gas heat transfer, which further cuts into neutron economy and space needs... on the whole the two concepts shouldn't be too far apart. With regular cladding you definately get higher HM, but you get that cladding as neutron penalty... with TRISO you're a bit lower in HM but don't have the neutron penalty of all that metal.

In any case if you have a HEC then this is going to mean a lot of neutron captures in the insulation (even if you use MgO rather than YSZ).


Top
 Profile  
 
PostPosted: Jul 12, 2015 6:52 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
So for CANDU,

http://www.iaea.org/inis/collection/NCL ... 006133.pdf

483 mm x 102 mm x (Pi/4) = 4 liters, 20 kgU, 5 gU/cc, ~5.7 gUO2/cc.

That's with good old water coolant and still enormous pump power... I quite imagine having to tune things down for gaseous coolant, even with enormous dT. You'd want larger rod pitch and probably fins of some sort between the rods.


Top
 Profile  
 
PostPosted: Jul 12, 2015 7:15 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
Yttrium stabilised Zirconia comes out quite good on the neutron absorption stakes simply because its 70% porous - in other words its mostly coolant, which in a gas cooled reactor is essentially empty space.

I don't think you can make Porous MgO practically. Whatever it is it going to be nothing compared to steel cladding....


Do you have anything on direct cooling TRISO? I can't seem to find much material.

EDIT:
I have found his patent for BWRs and his original articles.
I assume we will not bother with attempting to use them as a pebble bed?
We could just use an outer pyrocarbon layer to sinter them into a porous block - then just remove them like conventional fuel elements. It would allow the outer sleeve to be less tightly fitted to the element block - turning into a debris strainer. After all the coefficient of expansion of the elements and the sleeve will be different anyway.
And fuel elements vibrating in the sleeve is probably not good.

The lack of a fission gas plenum allows for potential axial shuffling which could be useful.

I suppose the advantage with this concept is that the thermal shutdown transient is essentially irrelevant - the containment layers on the kernels can handle silly high temperatures.


Top
 Profile  
 
PostPosted: Jul 12, 2015 9:42 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
It may be 70% porous, but its also thick, ballpark of 10x the cladding thickness, or more. Its insulation - unlike cladding, it needs quite a bit of thickness. We have to be careful with too high a porosity since the CO2 will flow through it and that loses your insulative properties. The SC-CANDU work just assumed the coolant to be stagnant but with 70% porosity and crazy high coolant velocities I wouldn't be too sure of that. Plus you need a protective liner sheet for the insulation. It would still be a lot of material in neutron flux, so this reactor won't win awards on neutron economy. You won't beat CANDUs anyway and unless you want to run on natural uranium fuel I don't really see the point in optimizing for neutron economy. Just run slightly higher enrichments...

The Russians did some cool tests with supercritical water and TRISO fuel with SiC coating.


Top
 Profile  
 
PostPosted: Jul 13, 2015 2:51 am 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
CANDU has 4mm of zircalloy pressure tube and about 0.5mm of calandria tube.
Our pressure tube is only about 2mm thick as its at low temperature. We have about 0.5mm of liner tube (its full of holes so its effective thickness is going to be smaller I imagine)
Similarly the two fuel annulus tubes are going to be about 0.5mm thick, but one is near the tube centre so the amount of material is smaller so say about 0.75mm overall.
That takes us to 3.25mm and then we consider the insulation. If we can use this annular scheme the temperature at the liner is only 230C or so, so our insulation can be quite thin so I think an equal zirconium mass to a CANDU tube is not out of the question.
I am not sure lateral flow through the insulation would be much trouble really - the gas has to flow sideways through the liner tube, then flow laterally through the labyrinthine insulation material and then back through the tube again, that is going to have a huge relative pressure drop.
With no steel cladding neutonics should be quite good.

As to higher enrichments that only works till you hit a hard enrichment limits

EDIT:
Would you make the inner and outer zone be simple nested cylinders for ease of manufacturing or would you make it a very gradually tapered cone to prevent stagnant coolant regions/wasted volume in the core. Albeit the wasted volume is essentially empty so I don't think it really hurts the reactor much.


Top
 Profile  
 
PostPosted: Jul 13, 2015 10:02 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
The Tsiklauri SCWR work was tapered. Too bad the link is dead, it had some interesting "micro fuel elements" fuel channel design. (it was actually a monolithic pressure vessel but with internal calandria, so the vessel could be cooled with feedwater all around). Non tapered should work fine though.

Would you consider a triplex SiC pressure tube? It will save you some neutrons, and you never get a fuel channel overheating now (even with insulation failure accidents). Should be quite doable to seal these off with double high temp braze to stainless headers or some such.

We're good on enrichment. CANDUs make do with 0.7%, despite all the Zr cladding. 3x lower HM I think just means going to 2 ish % enrichment, basically AGR levels and less than LWR fresh core levels. Even reload enrichments should be fine with the 5% limit. Newer enrichment plans typically have higher licensed enrichments because they see the potential market for Gen IV (enrichment plants last a long time so you gotta be strategic) and most have higher enrichment need.


Top
 Profile  
 
PostPosted: Jul 13, 2015 12:27 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
If we assume a 200mm pressure tube internal diameter, and a 110mm 'hot tube' diameter, we could then have a ~170mm outer sieve diameter.
That gives us a cross section in the fuel 'hopper' of 126 square centimetres, if the tube is 800cm long that gives us a 100 litre fuel bed.

5mm fuel elements contain a uranium dioxide kernel 4.8mm in diameter, which will contain 0.53 grammes of uranium.
A 60% packing factor means that there will be ~9170 elements per litre - which is 486kg of uranium.
Which is more than an AGR channel has in it in a similar diameter. (If you include the graphite gas gab)


Top
 Profile  
 
PostPosted: Jul 13, 2015 3:30 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
The inner tube could probably be a bit smaller.

CO2, 60 bar, 700C has a density of 32 kg/m3, isobaric heat capacity of 1.24 kJ/kg/K. With a dT of 400K you'd be running about 500 kJ/kg, so with a channel power of 5 MJ/s giving only 10 kg/s flow rate, kinematic viscosity 1.26*10e-6. Using the handy online pressure drop calculators, that's only about a bar of pressure drop for an 80 mm inner pipe. Also I doubt you need more than a 10 mm cold inlet annulus between the fuel annulus and the pressure tube insulation liner, just based on pressure drop.

We may want to run the thing a bit hotter. SiC is like Formula 1 stuff: it works best when its hot. It actually dislikes running cold when there's neutrons flying about. PWR fuel claddding may actually be "too chilly" for SiC.

How big can we make the channels? Presumably with just CO2 around, this ends up being undermoderated too much pretty quickly. Good reason to have gas/void in the central parts of the coolant channel though!


Top
 Profile  
 
PostPosted: Jul 13, 2015 3:56 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
I am not sure we really want to let the pressure drop be too large - after all that is one of the thing that leads to the enormous pumping power of gas cooled reactors.

If we keep the 170mm outer sieve diameter, a 20mm annulus gives us a 190mm insulation tube diameter, which gives us a 190.5mm insulation inner diameter, thanks to our low inlet temperature ~200C we can get way with about 3mm of insulation, giving us 193.5mm inner pressure tube diameter.

Excel is about a 700MPa UTS and we are allowed to load it to a third of that.
A handy hoop stress calculator gives the thickness required for that at 2.7mm (the bigger tube has hurt our performance but that is to be expected).

I would be worried about running much hotter than a 590C exhaust, because that might stop us using zircaloy for the hot-side seive tube. On the plus side it does provide access to the crazy USC turbines available.
I would consider SiC but I am not sure it is proven enough for things like that.


Top
 Profile  
 
PostPosted: Jul 13, 2015 7:18 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
Well, high pump power is just what you're going to get with gasses. Stuff is compressible, an annoying property for a reactor coolant. In stead of pushing it about you're always just pushing it into itself, which is a lossy process. Most of the energy is recovereable but still big circulators is what you're going to get with gasses.

If the annulus is 20 mm then you get 170 + 20 + 20 = 210 mm, but that's nitpicking. You're right its pretty thin.

Design to 1/3 rupture is reasonable.

For an internal with no joint or pressure retaining function, multiplex SiC is a good and proven enough option I'd say. For a big pressure vessel that's a whole different ballgame. But small pressure tubes, sure. As long as its the multiplexed (fiber reinforced) stuff, not the ceramic... its pretty good.

Re the inlet temp, won't you want to use the 250C feedwater that most of these modern cycles use? If so you want the inlet gas well above 250C, otherwise you're just cooling the hard won feedwater heating. And anyway, the SiC really loves to operate hot, so increasing the inlet gas temperature is fine. Its only when you get above 450C or so that creep starts to limit design stresses, I'd be fine with a 400 C in and if we already insulate the inside of the fuel channels, its not a stretch to insulate all the rest of the ducting and pipework (where internal pressure space isn't so limited).


Top
 Profile  
 
PostPosted: Jul 13, 2015 7:30 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
Problem is that if we start increasing the inlet temperature we increase our pump power even further, and it starts to get more than a little ridiculous - AGR had huge circulators even with a 300 celsius delta-T.

We might have an advantage in that if we go for the ~260 celsius inlet temperature (good catch about that by the way) then the fact that the fuel beads will run hotter than the coolant temperature might allow the outer beads to run hot enough that it won't matter about creep since although they might be bathed in cool gas the beads themselves will be hot enough.
There will be a small amount of heating through the cool-side sleeve since its made of a relatively conductive material (its a metal after all, even if zircaloy is not great in the metal stakes) and is in contact with the outer part of the fuel bed.
I doubt strain will be a problem, especially with the relatively low burnups at ~5% enrichment.

And gah, I thought I had made a mistake (confusing radius for diameter :\ ) - gives us a 213.5mm diameter, which is a 3mm pressure tube, which is still thinner than the CANDU tube!
With ~3mm or so of insulation at 70% porosity and our ~0.5mm liner tube our effective thickness is still only 4.31mm (Zirconia has 90% of the zirconium density of the metal).

So it comes out just slightly better than the conventional CANDU pressure tube/calandria tube combination.

I will dig out some stress figures for SiC multiplex used in pressure tubes and see what the thickness' come out like.
If its decent we could potentially use it on the hot side sieve tube, at which point we could push to the AGRs outlet gas temperature.....


Top
 Profile  
 
PostPosted: Jul 14, 2015 6:01 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1494
I just realised we have a somewhat unusual accident scenario to consider.

Failure of either the inner or outer sieve tube with the collapse of the bed column into the bottom half of the fuel channel.
The inner sieve is far more likely to fail due to greater corrosion (if its zircaloy its by far the hotter of the two), and it is one where the gas pressure will attempt to disassemble the bed, failure of the cold side sieve might be okay until reactor shutdown because the gas pressure drop will push the bed against the hot sieve tube anyway.

Failure would be detectable by a decrease in the pressure drop as the bead are expelled from the bed, which I assume would trigger a reactor scram.
The coolant velocities are sufficiently low the beads should still fall through it (I make it ~25m/s at the top of the tube and falling as you move downwards), so they will pile up in the bottom of the centre tube (which is drawn off from the top).
I would recommend that the fuel assembly be made to sit on a stainless steel 'seat' which has a hole that corresponds to the hot side tube drilled into it, forming a sump that the first few beads will fall into as they cascade through the tube.
If its a metre tall it would be able to take ~10L or 10% of the channel beads, which means the rest would pile up above it.
It would however delay the possible reactivity insertion, giving time for control rods to come in and shut down the core.

It might also be worth having BWR style rods coming into the vessel from below as the fuel will pile up there in an accident.
It would mean surrendering gravity insertion though.

Also a bit worried about pulling the assembly out if the bed is only supported by the sieve tubes, so I recommend the fuelling machine/rig lower a rod through the hot tube and engage with the base plate of the assembly which can be heavy duty as its not in the neutron flux. That loads the assembly in compression rather than tension.

EDIT:
Estimate is the surface area of the 917,000 elements in the channel is roughly 72 square metres.
That is ~12x the surface area in the AGR tube - so the fuel should run far cooler.


Top
 Profile  
 
PostPosted: Jul 14, 2015 7:26 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5056
Absolutely, fuel failure scenarios are a bit odd with this configuration. Easy to fix though, with a SiC inner tube, and a backup strainer at the end of the outlet fuel channel, to catch any particles if they move about. I'm thinking the inner tube is part of the fuel assembly and gets replaced during the offline refuelling you proposed. Any particles that are dislodged would still be retained by the end strainer, that would also be part of the fuel assembly...

Also, dislodging particles doesn't fail them. If anything they are now in a place with more coolant and no adjacent hot particles. If the amount of dislodged particles is very large the fuel channel would choke a bit, overcrowding the end strainer, but it still wouldn't fail the particles.

Quote:
Estimate is the surface area of the 917,000 elements in the channel is roughly 72 square metres.
That is ~12x the surface area in the AGR tube - so the fuel should run far cooler.


You're starting to get the great advantage in surface area of this concept! we could in fact increase power density a lot.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 63 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 0 guests


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