Energy From Thorium Discussion Forum

It is currently Mar 24, 2017 10:56 pm

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 52 posts ]  Go to page 1, 2, 3, 4  Next
Author Message
PostPosted: Feb 12, 2017 5:10 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
So I have been thinking about this on and off for a while, and now that I am considering postgraduate education I have decided to flesh it out a bit in case I want to use it as an outline design exercise for an MPhil.

This started off as a hybrid between an Atucha-style pressure vessel HWR and a CANDU - which was brought up by the realisation that the Atucha-IIs burnup is actually limited by the fact that it has full length fuel bundles and thus the uranium in the end bundles experiences much lower burnup than in a CANDU with an eight-bundle shift programme (Where the end sets of four bundles in the channel are in the reactor twice as long as the middle four).

Since the Atucha has no pressure differential across the channel it utilises only a 1.4mm thick channel tube and a 0.4mm thick liner tube, unlike the 4.2mm pressure tube of the CANDU.
Additionally the lack of huge mechanical strength required (as the pressure differential across the channel wall is only that introduced by the pressure drop in the channel) it is potentially possible to replace the the zircaloy with graphite - which would reduce the neutron losses to negligible levels. [As there is a connection between the moderator and HTS circuits in the pressure vessel of an Atucha design we don't need the channel tube to sag against the isolation tube to conduct heat away, as the fuel will not be uncovered in that scenario]
This would be a reactivity insertion of something like ~30mk compared to a standard CANDU fuel channel and would allow burnups of going on 11MWd/kg to be obtained with natural uranium. [Thorium SSET burnups would also be increased relative to more traditional CANDUs].

We achieve this by encasing the entire calandria-esque core structure in a prestressed cast-iron structure that will also serve as a radiation shield, and thanks to its large thickness (110 bar operating pressure will have walls at least a couple of metres thick everywhere) there is no real need for thermal insulation as would be required with a traditional vessel or with a precast concrete vessel [the cast iron can stand the full channel temperature of 300C without significant issue, and moderator temperature will be normally around 190C].
The pressure vessel would contain the core structure, the pressuriser for the primary heat transfer circuit, cavities for canned rotor pumps, compact gas heaters (explained below) and various other bits of equipment.
I am somewhat torn between using LWR style reactivity devices that are actuated from outside the vessel, or just having a vertical extension of the vessel above the core that would enable a CANDU style reactivity mechanism deck that would be under the full core pressure, with the control rods free to simply drop under gravity as there would be no pressure differential across them.

The core would contain something like 640 channels, to gain maximum economies of scale (since our vessel can be of almost any size) and also to improve neutron economy by reducing core leakage. The loading and unloading would be using a pair of standard CANDU style fueling machines pushing units through the channel (it turns out the single ended loading in the CANDU 3 concept reduces burnup since high burnup fuel can't be shunted through the centre of the core from the far end for unloading) with a typical eight bundle shift. Fuel would be as in the CANDU series.

It seems likely that you could extend the primary pressure vessels so that it enclosed both fueling machines at the ends of their reactor - this reduces the likelyhood of a LOCA due to a fueling machine failure or closure failure from escaping the pressure vessel.
Fuel could be passed two and from the machines in operation through locks designed for the purpose - only a double failure of a channel end and a lock would allow the reactor coolant to escape.
As the prestressed structure effectively prevents vessel failure the largest escape holes would be through the gas lines to the carbon dioxide turbine cycle (recuperated, multiple-reheat condensing cycle using moderator heat to reduce the need for recompression, one might be included when I do the actual calculations depending on the difference it makes).
THese lines are however pluggged by the PCHEs that would be used, which will limit the effective leak size considerably since the vessel will always retain the heat exchangers and the maximum channel sizes are quite small.

Thus a LOCA Is far less likely and far less severe if the vessle is properly designed.
Additionally the steel vessel being able to heat up considerably also means that in a blackout heat leakage through the channel and isolation tubes (the intervening space being full of stagnant moderator water) would cause the moderator tank to increase in temperature from ~190C to ~300C.
This would allow an array of Variable Conductance Thermosyphons to be used to remove ~0.5-1% of reactor heat passively in such a scenario, to air cooled radiators outside the reactor confinement. (They would start to turn on about 220C and reach full power at about 300C). They would be split up the passive cooling power and the thermosyphon is inherently double walled and could be overbuilt to withstand thefull primary loop pressure on each end to prevent them becoming a path to vent material to the atmosphere).

There would be a confinement building over the pressure vessel and turbine plant, which would have a filtered venting system as the maximum feasible LOCA is quite small as it will be a leak through a heavy water cleanup line or similar. It would not need to be hardened against major internal overpressure - and thanks to the small size of carbon dioxide plant the entire turbine plant can be inside the containment.

Using the pressure vessel as integral thermal and radiation shielding allows the gas lines to the reactor to be kept very short, keeping pipe mass and expense small and allowing for a double reheat cycle with multiple intercoolers on the carbon dioxide primary compressor - this would yield summer efficiencies (19C seawater temperature) of ~35% and winter efficiencies (5C seawater temperature) closing in on 40%.


In summary - use an Atucha/CANDU hybrid with a huge prestressed cast iron pressure vessel that has a multiple reheat supercritical carbon dioxide condensing cycle system, all inside an outer confinement building.
HIgh neutron economy, high efficiency for maximum electricity produced per unit fissile.


Top
 Profile  
 
PostPosted: Feb 13, 2017 3:21 am 
Offline

Joined: Apr 19, 2008 1:06 am
Posts: 2187
For safety of a PHWR, it may be more fruitful to change the coolant to a less volatile fluid rather than a heavy design of the vessel. The heavy water moderator is in any case cooled and kept at a low pressure.
A low vapor pressure organic fluid or a salt mixture can surely be found as coolant working at boiler temperature. Heavy pressures are best created at steam generator or equivalent and used in turbine. The primary coolant circuit in the reactor is best kept at a low pressure.


Top
 Profile  
 
PostPosted: Feb 18, 2017 10:51 am 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
It appears a 700 channel ~1500MWe (38.5% basis, so winter efficiency) unit would have rather drastically lower neutron leakage than a reference 380 channel 700MWe class PT-HWR.

Including the reactivity insertion from the thinner and lower cross section channel tubes we could be looking at something approaching 49mk of total reactivity insertion [neutron leakage drops from 35mk to 14mk]. Even with adjusters in the core you could be looking at something like 13,100MWd/t - which is rather better than the 7,200MWd/t of the conventional design with the exact same fuel elements.

Depending on power peaking levels, and whether it is feasible to sacrifice xenon override capacity, it might be feasible to use Pickering style 28 fuel bundles in some peripheral channels that are lower power anyway, which would increase the lattice activity in the peripheral regions [higher uranium content and more resonance self shielding], which would provide power flattening far more efficiently than either the adjuster rods or the burnup flattening practicied in Bruce-A.
But the licenced power of a 28 element bundle is only 750kW as opposed to 900+kW for a 37 element bundle. So the level of power flattening that could be provided in such a case is limited.


Top
 Profile  
 
PostPosted: Feb 22, 2017 9:26 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Hi again Ed,

An interesting proposal you have here!

A couple questions/comments below...

1. Graphite for a water tight channel. There are probably stronger, non-swelling yet low neutron absorption options such as SiC-SiC composite. Significant advances have been made in the joining and fabrication in recent years.

2. Reactivity. No doubt about potential for increased reactivity but you have check the coefficients, especially density in relationship to the temperature coefficient.

3. Don't understand the power cycle. Is the coolant heavy water? If so how does the CO2 cycle work? Doesn't the CO2 cycle require higher temperatures?

4. Prestressed cast iron. Interesting choice. Has this ever been applied before? Sounds like a high risk concept. Cast iron is also difficult to weld so any penetrations and pipe attachments will be difficult. I would suggest to consider cast steel instead, it is superior mechanically and easier to weld. Cast steel may avoid the need for prestress.

5. Hot prestressed vessel concept. One issue with higher temperature prestress is that heating the prestress cables loosens them, which is the opposite of what you want! The cables would have to be cooled but this is sort of what you're trying to avoid presumably (ie a liner cooling system). Another issue to look at is thermal stress. Even startup and cooldown would have to be considered. In an accident with heatup you can get a hot inner surface while the outer wall is still at nominal temperature. This gets bad really quickly for sections of meter plus thickness. In terms of insulation, you are going to have to insulate this thing on the outside or cool it. Can't just have a massive hot naked cast iron block sit in a room. Maybe you can have a passive cooling system on the outside of the casting.

One option I really like is a steel sandwich filled with concrete. The sandwich can have prefabricated cooling liner and insulation in it. The cooling liner could be passively cooled in some fashion.

6. CANDU fuel. You have to consider hydrogen threats then (pressure, detonation, deflagration). Might want to consider SiC fuel rods. MIT was working on that. Haven't checked the latest status. One innovative SiC-UO2 fuel is called a micro fuel element, MFE, there was some work on it in the past. SiC would have better neutron economy.

7. Moderator passive cooling. There was some work for CANDUs to have a passive moderator heat rejection. It was just an elevated condenser with a two phase riser pipe connecting to it and a single phase downcomer going back. It could probably be passively air heat rejected using a stack of sorts. It could also be a PCHE in a pond of water if you need a compact system.

Perhaps you can play around with other interesting ideas such as natural gas fired superheat for the power cycle. (I know, fossil fuel blasphemy!)


Top
 Profile  
 
PostPosted: Feb 22, 2017 10:18 am 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
Cyril R wrote:
Hi again Ed,

Its good to see you knocking around again!
Cyril R wrote:
1. Graphite for a water tight channel. There are probably stronger, non-swelling yet low neutron absorption options such as SiC-SiC composite. Significant advances have been made in the joining and fabrication in recent years.

I have been keeping watch on those composites, but I simply used Graphite here as it was the first thing that came into my head when I thought 'very low neutron absorption stuctural material' - SiC is certainly a very interesting material, much lower absorption potentially than the zircalloy alternative.
Cyril R wrote:
2. Reactivity. No doubt about potential for increased reactivity but you have check the coefficients, especially density in relationship to the temperature coefficient.

As I understand it the CANDU is actually overmoderated because otherwise it was difficult to fit all the headers and other equipment into the lattice matrix - so the moderator temperature coefficient is actually positive in a CANDU. So increasing the moderator temperature to 180C (its pressurised) would potentially increase reactivity (or reduce it if we shift into the undermoderated regime, but the lattice pitch could always be increased).
Cyril R wrote:
3. Don't understand the power cycle. Is the coolant heavy water? If so how does the CO2 cycle work? Doesn't the CO2 cycle require higher temperatures?

One of the things that set me down this path was a paper published by a project manager at the Mohovce Nuclear Power station in Slovakia - one of his coauthors is at my university and one of my options for next year is an MPhil expanding on the work - if I can find the money obviously.
It turns out PNNL has also being doing some work on multiple-reheat recompression carbon dioxide cycles for LWR temperatures, and concluded that a condensing cycle could easily beat a conventional cycle of similar equipment complexity, although pipelines for carbon dioxide are the difficult part.
As the vessel here forms the biological shield the turbines can be potentially directly adjacent to the reactor (with PCHE gas heaters inside the vessel), the pipeline lengths are very short and we can therefore have multiple reheats without much difficulty. [The reactor inlet temperature rapidly becomes the limiting factor]
These cycles also show a substantially increased efficiency as the cold reservoir temperature decreases - and with intercooled main compressor-pumps. Which is good since winter cooling water temperatures at British stations can sometimes fall as low as 6C. And it might be worth building bigger intake structures to obtain deeper, colder seawater all year round.
Cyril R wrote:
4. Prestressed cast iron. Interesting choice. Has this ever been applied before? Sounds like a high risk concept. Cast iron is also difficult to weld so any penetrations and pipe attachments will be difficult. I would suggest to consider cast steel instead, it is superior mechanically and easier to weld. Cast steel may avoid the need for prestress.

Prestressed Ductile Iron was apparently first considered by the THTR reactor people in Germany in the 70s - engineers were getting annoyed that the reactor physicists were building pressure vessels and then giant cast iron biological shields, so decided to determine what would happen if they applied the concepts of prestressed concrete to cast iron.
Here is the paper that put me onto the concept - there are numerous other ones like it.
I believe some of the GT-MHR people are considering the pressure vessels since they can be built to arbitrary sizes - one proposal I saw was to put their entire turbine plant inside a cylindrical one so that leaks won't overpressure their confinement building. Which is something I am currently struggling with.
As for welding - as I understand it the vessel blocks don't actually get welded to each other, they are keyed to stop them dislocating, but otherwise they are retained in place by the prestress wires, this also promotes a nice failure mode where an overpressure just leads to the blocks moving apart, the liner giving away locally and the core fluid vents from the gaps. When the pressure drops low enough the prestressed cables shrink again (assuming you design so they are always in the elastic regime) and the blocks reseal the gap. Albiet now it will leak slowly due to the liner break.

But no horrifying brittle failures that blow the vessel wide open.
Cyril R wrote:
5. Hot prestressed vessel concept. One issue with higher temperature prestress is that heating the prestress cables loosens them, which is the opposite of what you want! The cables would have to be cooled but this is sort of what you're trying to avoid presumably (ie a liner cooling system). Another issue to look at is thermal stress. Even startup and cooldown would have to be considered. In an accident with heatup you can get a hot inner surface while the outer wall is still at nominal temperature. This gets bad really quickly for sections of meter plus thickness. In terms of insulation, you are going to have to insulate this thing on the outside or cool it. Can't just have a massive hot naked cast iron block sit in a room. Maybe you can have a passive cooling system on the outside of the casting.

This is currently one of my major issues, but my currently proposed solution is not to insulate the vessel (as it is rather thick and apparently these vessels are often concieved of being built of cellular rather than solid blocks) and simply cool the outside, either by submerging it in a tank of water or providing cooling pipes on its surface. As the prestressed cables tend to be at, or at least near, the outer surface that would stop their temperature from increasing to a large degree. Not insulating the vessel also provides a small, but significant, heat sink in an accident situation - as the liner temperature is set by the moderator which is much cooler than the moderator will reach in the case of a blackout accident. Heat loss of 0.1-0.2% is tolerable in operation but 0.2-0.4% would be very important in an accident.
Cyril R wrote:
6. CANDU fuel. You have to consider hydrogen threats then (pressure, detonation, deflagration). Might want to consider SiC fuel rods. MIT was working on that. Haven't checked the latest status. One innovative SiC-UO2 fuel is called a micro fuel element, MFE, there was some work on it in the past. SiC would have better neutron economy.

It is very interesting work, but I was considering this thing more of a baseline that could be modified - and there is such a wealth of data on the performance and operation of CANDU fuel bundles and their fueling equipment that it would significantly increase the computational challenge to discard it.
After all as far as a CANDU fueling machien is concerned, the fuel channel it hooks up to is no different to an ordinary CANDU pressure tube, just that it seems to get superior neutron economy.
Cyril R wrote:
7. Moderator passive cooling. There was some work for CANDUs to have a passive moderator heat rejection. It was just an elevated condenser with a two phase riser pipe connecting to it and a single phase downcomer going back. It could probably be passively air heat rejected using a stack of sorts. It could also be a PCHE in a pond of water if you need a compact system.

The hard part is that I want the moderator heat in normal operation, thanks to it being in the right temperature range to help reduce the recompression fraction in the power cycle. Which improves plant efficiency.
That is why I proposed a variable conductance thermosyphon, because in a power blackout event the moderator temperature will rapidly increase to equilibrium with the coolant-side temperature, which would then activate the thermosyphons.
Cyril R wrote:
Perhaps you can play around with other interesting ideas such as natural gas fired superheat for the power cycle. (I know, fossil fuel blasphemy!)

Interestingly that's what Blair Bromley from CNL said!


Top
 Profile  
 
PostPosted: Feb 22, 2017 11:41 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Quote:
Its good to see you knocking around again!


Always good talking to a creative genius like you!

Quote:
As I understand it the CANDU is actually overmoderated because otherwise it was difficult to fit all the headers and other equipment into the lattice matrix - so the moderator temperature coefficient is actually positive in a CANDU. So increasing the moderator temperature to 180C (its pressurised) would potentially increase reactivity (or reduce it if we shift into the undermoderated regime, but the lattice pitch could always be increased).


So increasing moderator temperature in an accident increases reactivity? That doesn't sound nice? You'd need to quantify this in an accident analysis. That might be part of your thesis, it's very interesting work combining various interesting fields (materials, stress analysis, chemistry, radiochemistry, thermalhydraulics, reactor physics, transport dynamics, etc.). If you can get the tools to do this it would be a great addition to the thesis and it is fascinating work to do.

Quote:
One of the things that set me down this path was a paper published by a project manager at the Mohovce Nuclear Power station in Slovakia - one of his coauthors is at my university and one of my options for next year is an MPhil expanding on the work - if I can find the money obviously.
It turns out PNNL has also being doing some work on multiple-reheat recompression carbon dioxide cycles for LWR temperatures, and concluded that a condensing cycle could easily beat a conventional cycle of similar equipment complexity, although pipelines for carbon dioxide are the difficult part.
As the vessel here forms the biological shield the turbines can be potentially directly adjacent to the reactor (with PCHE gas heaters inside the vessel), the pipeline lengths are very short and we can therefore have multiple reheats without much difficulty. [The reactor inlet temperature rapidly becomes the limiting factor]
These cycles also show a substantially increased efficiency as the cold reservoir temperature decreases - and with intercooled main compressor-pumps. Which is good since winter cooling water temperatures at British stations can sometimes fall as low as 6C. And it might be worth building bigger intake structures to obtain deeper, colder seawater all year round.


Cool! Much cooler than I thought, if you get my meaning. I'd still look at putting in a gas fired superheater (make it biogas or green gas, adding another interesting dimension to the thesis). Possibly you can simplify the CO2 cycle if you have access to an external superheat source. Add a CO2 scrubber or membrane so you can be self sufficient in your makeup CO2 needs!

Quote:
Prestressed Ductile Iron was apparently first considered by the THTR reactor people in Germany in the 70s - engineers were getting annoyed that the reactor physicists were building pressure vessels and then giant cast iron biological shields, so decided to determine what would happen if they applied the concepts of prestressed concrete to cast iron.
Here is the paper that put me onto the concept - there are numerous other ones like it.
I believe some of the GT-MHR people are considering the pressure vessels since they can be built to arbitrary sizes - one proposal I saw was to put their entire turbine plant inside a cylindrical one so that leaks won't overpressure their confinement building. Which is something I am currently struggling with.
As for welding - as I understand it the vessel blocks don't actually get welded to each other, they are keyed to stop them dislocating, but otherwise they are retained in place by the prestress wires, this also promotes a nice failure mode where an overpressure just leads to the blocks moving apart, the liner giving away locally and the core fluid vents from the gaps. When the pressure drops low enough the prestressed cables shrink again (assuming you design so they are always in the elastic regime) and the blocks reseal the gap. Albiet now it will leak slowly due to the liner break.


Interesting stuff. For welding I was thinking more along the lines of any pipework attached to this thing. Even if it is integral it is going to need many penetrations, tendon sleeves, prestressing galleries, pipe attachments, maybe some entrance hatch.

Quote:
But no horrifying brittle failures that blow the vessel wide open.


I should hope that that is a requirement! But it's not a given. Cast iron is definately not as ductile or tough as cast steel and may suffer some in service embrittlement. Inclusions and imperfections are manufacturing issues with thick walled, large castings, especially if the casting is so big that it must be cast on-site. This is nuclear we're talking about... many requirements on materials, placements, inspection, testing.

Quote:
This is currently one of my major issues, but my currently proposed solution is not to insulate the vessel (as it is rather thick and apparently these vessels are often concieved of being built of cellular rather than solid blocks) and simply cool the outside, either by submerging it in a tank of water or providing cooling pipes on its surface. As the prestressed cables tend to be at, or at least near, the outer surface that would stop their temperature from increasing to a large degree.


You'd have to run some numbers on thermal stress. 200-300 degrees over a couple meters of metal is going to be a large stress. You can get cracks if the stress exceeds the prestress substantially. Young's Modulus is really big... couple hundred billion Pascals for most metals.

Quote:
Not insulating the vessel also provides a small, but significant, heat sink in an accident situation - as the liner temperature is set by the moderator which is much cooler than the moderator will reach in the case of a blackout accident. Heat loss of 0.1-0.2% is tolerable in operation but 0.2-0.4% would be very important in an accident.


In AGRs they do insulate it, on the inside, but they also have a liner cooling system, it removes a significant fraction of heat. Something like 0.5% in AGRs IIRC. You could also consider that as an option. Probably need a liner anyway for corrosion resistance and maintaining coolant purity. Cooling jacket could be either tubes or PCHE style plates behind the liner. Meltable insulation is of course a cool option too, but a bit hard to model because of phase change.

Quote:
It is very interesting work, but I was considering this thing more of a baseline that could be modified - and there is such a wealth of data on the performance and operation of CANDU fuel bundles and their fueling equipment that it would significantly increase the computational challenge to discard it.
After all as far as a CANDU fueling machien is concerned, the fuel channel it hooks up to is no different to an ordinary CANDU pressure tube, just that it seems to get superior neutron economy.


Ok, if you want to keep the CANDU bundle geometry that could be done with the SiC-SiC stringer/bundle type fuel. Would be worthwhile to consult an expert in this field for the latest developments.

Quote:
The hard part is that I want the moderator heat in normal operation, thanks to it being in the right temperature range to help reduce the recompression fraction in the power cycle. Which improves plant efficiency.
That is why I proposed a variable conductance thermosyphon, because in a power blackout event the moderator temperature will rapidly increase to equilibrium with the coolant-side temperature, which would then activate the thermosyphons.


That might work, you'd need a lot of these thermosyphons. I guess you can just run a normal cooling system for power heat recovery and then have any passive concept that rapidly loses heat with increasing temperature. Heat pipes can be configured to be like that too, as they work on phase change so are going to have a fairly steep cut-off point, especially wick-less ones.

Quote:
Interestingly that's what Blair Bromley from CNL said!


Really, that's interesting. It's not a new idea though. I actually ran numbers on this many years ago (where goes the time). For the AP1000. Turns out that (based on steam tables differences in specific enthalpy comparison), superheating that cold steam would to modern 600+ deg C superheat steam conditions adds some 50% enthalpy, and pushes up efficiency by at least 30% and maybe more with multi reheat stages. So installing the superheater is like having a twin unit AP1000 for the price of one plus a gas heater. Oversimplifying of course but that's pretty amazing and rather tempting. Reductions in turbine cost by going to full speed superheat turbine are substantial, not to mention eliminating the complicated moisture separator-reheater, the steam separator, and steam dryer (thus making containment smaller too since the nuclear SGs become much smaller). Almost too good to be true so it makes you wonder why it hasn't been done yet.

By the way if you are planning on having the turbine in the containment, you'd have to look at safety and operational aspects of that. Can it be maintained, what about turbing missiles, CO2 LOCAs, electrical/hydrogen generator fires etc. (hopefully you can use a helium cooled generator with your setup).


Top
 Profile  
 
PostPosted: Feb 22, 2017 5:10 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
Cyril R wrote:
So increasing moderator temperature in an accident increases reactivity? That doesn't sound nice? You'd need to quantify this in an accident analysis. That might be part of your thesis, it's very interesting work combining various interesting fields (materials, stress analysis, chemistry, radiochemistry, thermalhydraulics, reactor physics, transport dynamics, etc.). If you can get the tools to do this it would be a great addition to the thesis and it is fascinating work to do.

I think we could stay out of the overmoderated regio giving the lower density of the moderator at 180C, but that does lead to wierd reactivity insertions when the reactor is cooling down during an ordinary shutdown that will require additional control absorbers to keep it under control.
Can't use CANDU style liquid control absorbers though due to wanting to keep gas out of the pressure vessel, last thing we need is a LOCA causing the pressure in the vessel to drop and thus creating a giant bubble that might stop coolant getting to fuel elements. Will have to be mechanical absorbers - which were planned for the ACR so its probably not a big issue.
Cyril R wrote:
Interesting stuff. For welding I was thinking more along the lines of any pipework attached to this thing. Even if it is integral it is going to need many penetrations, tendon sleeves, prestressing galleries, pipe attachments, maybe some entrance hatch.

I think it will come to many hundreds, before we even include the 1400 fuel channel closures and all that [~700 channels with one at each end].
It might be feasible to use relatively small blocks for the areas around pipe attachments and similar and then do the welding in a factory before shipping it to the site. We can move blocks up to a hundred tonnes relatively feasibly.
The access hatches are more troublesome but it appears you need forged rings or similar to carry the stressing load around the opening.
Cyril R wrote:
You'd have to run some numbers on thermal stress. 200-300 degrees over a couple meters of metal is going to be a large stress. You can get cracks if the stress exceeds the prestress substantially. Young's Modulus is really big... couple hundred billion Pascals for most metals.

One interesting idea is if the liner can bridge small gaps without support, it might be possible to taper the blocks at the inner face so that differential expansion doesn't try and break the blocks apart - they will just change shape to close the gaps.
Just an idea though, but I iamgine it would be hard to design a vessel that did that without it being much weaker.
Cyril R wrote:
That might work, you'd need a lot of these thermosyphons. I guess you can just run a normal cooling system for power heat recovery and then have any passive concept that rapidly loses heat with increasing temperature. Heat pipes can be configured to be like that too, as they work on phase change so are going to have a fairly steep cut-off point, especially wick-less ones.

I think probably a few hundred of them, but they would be small and probably easily manufacturable in a factor (only ~10cm in diameter or something). That way a leak in any one of them is a negligible issue.
But a lot of the top of the pressure vessel would likely be covered in them, would have to fit it around the reactivity mechanisms in the centre of the core. Especially since the core is only 6m long.
Cyril R wrote:
Really, that's interesting. It's not a new idea though. I actually ran numbers on this many years ago (where goes the time). For the AP1000. Turns out that (based on steam tables differences in specific enthalpy comparison), superheating that cold steam would to modern 600+ deg C superheat steam conditions adds some 50% enthalpy, and pushes up efficiency by at least 30% and maybe more with multi reheat stages. So installing the superheater is like having a twin unit AP1000 for the price of one plus a gas heater. Oversimplifying of course but that's pretty amazing and rather tempting. Reductions in turbine cost by going to full speed superheat turbine are substantial, not to mention eliminating the complicated moisture separator-reheater, the steam separator, and steam dryer (thus making containment smaller too since the nuclear SGs become much smaller). Almost too good to be true so it makes you wonder why it hasn't been done yet.

So early PWRs and BWRs did that, especially in the US and Sweden - but it seems to have fallen out of favour in the late 60s, I imagine when moisture seperators became a big thing.
Also it appears most of that work was with direct oil firing - and that fell out of favour after the oil crisis and improvements in refining technology put the price of residual fuels thorugh the roof.

Additionally it is a bit of an operational headache if your turbine plant can't function if you have no oil or gas available for whatever reason. Although if you only want 600+C superheat, you might even able to use an OCGT as the superheater, which would increase your efficiency even more.

Or even a FIRES block, although that would be interesting.
Cyril R wrote:
By the way if you are planning on having the turbine in the containment, you'd have to look at safety and operational aspects of that. Can it be maintained, what about turbing missiles, CO2 LOCAs, electrical/hydrogen generator fires etc. (hopefully you can use a helium cooled generator with your setup).


The turbines are relatively lightweight (they weigh <100t tonnes rather than 5000t or whatever for a normal turbine), although we would be dealing with full speed machines so the energy per unit turbine weight is significantly higher.
The Carbon dioxide LOCA is the big one that I have identified, as the carbon dioxide in the turbines might be pressurised all the way up to 350 bar, and there will be significant holdup in the recuperators, precoolers, intercoolers and in the gas heaters. It could easily hugely overpressure the reactor building.

The idea of a second pressure vessel around the turbine plant starts to look significantly better as it would provide protection against both those threats and would also stop generator fires from affecting the reactor building as the pressure vessel would be vented to the outside of the confinement building through suitable ducts.
Another alternative would be a CANDU style vacuum building containment, but instead of having a pure water douse it would douse with sodium hydroxide solution. That would suck up carbon dioxide with ease if anything happened.
You could even line the confinement buildings with blocks of porous activated carbon which would absorb large quantities of carbon dioxide as the building pressurised - but that sounds a little crazy and ofcourse activated carbon is quite flammable so it would be a potential fire risk.

Splitting the turbine plant in two is an obvious solution - and you would not necessarily have to have two generators or such as the shaft is much shorter than in an LWR so you could have the turbomachines all be coaxial.
Given that our recuperators and gas coolers are made of small modules thanks to the manufacturing constraints on PCHEs and H2X heat exchangers and you might not loose much plant efficiency there.

Either way, many interesting ideas.
We could also put a combined heavy water upgrader/tritium removal plant inside the confinement building which would reduce the risk of tritium leaks and would also provide more volume for a carbon dioxide overpressure to expand into.
Perhaps even water cooled buffer stores for spent fuel.


Top
 Profile  
 
PostPosted: Feb 22, 2017 5:33 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Quote:
One interesting idea is if the liner can bridge small gaps without support, it might be possible to taper the blocks at the inner face so that differential expansion doesn't try and break the blocks apart - they will just change shape to close the gaps.
Just an idea though, but I iamgine it would be hard to design a vessel that did that without it being much weaker.


Thermal stress is a big issue because most solutions that are strong and stiff are not good for thermal stress, and vice versa. It takes some ingenuity to addres both strength/rigidity and thermal stress simultaneously.

Quote:
The Carbon dioxide LOCA is the big one that I have identified, as the carbon dioxide in the turbines might be pressurised all the way up to 350 bar, and there will be significant holdup in the recuperators, precoolers, intercoolers and in the gas heaters. It could easily hugely overpressure the reactor building.


One interesting feature of CO2 is that it is scrubbable. For example a caustic (KOH/NaOH/other) scrubber can efficiently remove CO2. This is another interesting study field for you perhaps.

A passive pressure suppression containment would work pretty well for a CO2 containment. Scrub the CO2 while quenching any steam simultaneously. Just need a caustic tank connecting the drywell (reactor vessel space) with the wetwell (caustic tank).

If you have accident resistant ceramic fuel then maybe you can get away without a containment, otherwise things may not be so well. Unless your filter in the filtered vent is extremely good (many 9's efficiency).

Tritium management is going to be interesting too if you have no containment. Especially if you have a LOCA.


Top
 Profile  
 
PostPosted: Feb 22, 2017 6:36 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
Cyril R wrote:
Thermal stress is a big issue because most solutions that are strong and stiff are not good for thermal stress, and vice versa. It takes some ingenuity to addres both strength/rigidity and thermal stress simultaneously.


That appears to be the big thing here. The problem I have with a liner cooling system is what happens in an accident if it fails.
We then have the difference appearing from nowhere across a vessel that might not be designed to deal with it.
Cyril R wrote:
One interesting feature of CO2 is that it is scrubbable. For example a caustic (KOH/NaOH/other) scrubber can efficiently remove CO2. This is another interesting study field for you perhaps.

A passive pressure suppression containment would work pretty well for a CO2 containment. Scrub the CO2 while quenching any steam simultaneously. Just need a caustic tank connecting the drywell (reactor vessel space) with the wetwell (caustic tank).


I am slightly nervous over having a rain of caustic soda falling on all the equipment in the turbine space during a LOCA - that could be a plant loss if it ruins equipment. All the equipment would have to be designed to deal with that. That is why I supposed we might have a seperate vacuum space, after all you would just not put anything in that area that might be damaged by the caustic soda.
Cyril R wrote:
If you have accident resistant ceramic fuel then maybe you can get away without a containment, otherwise things may not be so well. Unless your filter in the filtered vent is extremely good (many 9's efficiency).

Tritium management is going to be interesting too if you have no containment. Especially if you have a LOCA.

Since our largest plausible line break is going to be qiute small (smaller than a coolant channel thanks to the refueling machines being inside a 'hood') then loss of material could be quite small.
We could probably have a filtered containment system that could prevent overpressurising, or just douse it under lots of water.

The hard part is stopping all the tritium escaping, an oversized tritium seperation plant could be helpful here to remove as much tritium from the reactor as possible, so that the total tritium inventory is minimised.
But I think we need some sort of secondary confinement building, even if it not built for massive overpressure and has filtered vents.
We are also in the unusual situation where the turbine plant is at higher pressure than the reactor, so the turbine plant is unlikely to have any radioactive material in it under operation, except perhaps due to a leak in the last reheat stage (that might be lower pressure than the reactor).

Another thing I might consider is a dry fuel store based on the one at Wylfa - 100mm diameter fuel elements a metre long were stored in a naturally convecting carbon dioxide atmosphere with a peak heat output of nearly 1 kilowatt.
That translates to a heat output of something like 500W for the CANDU fuel element, and whilst these elements don't have fins they have large scale subdivision not present in a Magnox fuel element, and the Zircalloy (or SiC) can cope with a much higher temperature.

Even a 13.1GWd/t CANDU bundle will get down to 500W quite quickly, whihc minimises the actively cooled fuel that has to be inside the confinement building. (The magnox elements were/are stored 12 to a tube, in tubes 15m high).


Top
 Profile  
 
PostPosted: Feb 22, 2017 7:24 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Quote:
The problem I have with a liner cooling system is what happens in an accident if it fails.
We then have the difference appearing from nowhere across a vessel that might not be designed to deal with it.


Presumably the liner cooling system could be designed with sufficient water inventory that it could absorb the heat by boil-off until decay heat is much lower or cooling is restored.

Perhaps you can consider a pool type "reactor" but with a pool around the reactor vessel. The liner system could then simply be connected to the pool and passively thermosyphon water into and out of the pool.

If you use cast steel the casting should have sufficient tensile strength to be used at temperature without any prestressing, so that's another option to consider.

Quote:
I am slightly nervous over having a rain of caustic soda falling on all the equipment in the turbine space during a LOCA - that could be a plant loss if it ruins equipment.


It's odd stuff, it is mainly a biological risk to anything that's alive, but is surprisingly non-corrosive to most materials at ambient conditions. Reason why it makes for good sink declogger!

In the low concentrations required for scrubbing though it is not a major hazard (household chemical level).

Quote:
That is why I supposed we might have a seperate vacuum space, after all you would just not put anything in that area that might be damaged by the caustic soda.


Sounds similar to a wetwell-drywel concept of BWR. The wetwell is separated from other stuff.

Quote:
The hard part is stopping all the tritium escaping, an oversized tritium seperation plant could be helpful here to remove as much tritium from the reactor as possible, so that the total tritium inventory is minimised.


Makes sense. This thing is going to generate some hydrogen from radiolysis, so you can just integrate the tritium stripper into the hydrogen recombiner.

Quote:
But I think we need some sort of secondary confinement building, even if it not built for massive overpressure and has filtered vents


Tritium does not care much about confinement. Even containments have issues keeping the stuff in. It's an escape artist. If the tritium is present as heavy water then all you can do is dry and recirculate the confinement air. If the tritium is present as hydrogen you have more options.

You might want to reconsider the CO2 power cycle. If all it offers you is a percent point increase in efficiency it is hard to justify the risk of a new power cycle. This was a major factor in the demise of the PBMR. Developing a new power cycle is serious business not to be undertaken lighly even if you have a billion in funding. There has to be serious advantage like a large fraction increase in efficiency for this to be worth the risk.


Top
 Profile  
 
PostPosted: Feb 22, 2017 8:52 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
Cyril R wrote:
Presumably the liner cooling system could be designed with sufficient water inventory that it could absorb the heat by boil-off until decay heat is much lower or cooling is restored.

Perhaps you can consider a pool type "reactor" but with a pool around the reactor vessel. The liner system could then simply be connected to the pool and passively thermosyphon water into and out of the pool.

This is an interesting concept, but we would have to ensure that the tubes containing the water at ambient pressure were able to withstand the pressure inside the vessel without failure in a wide variety of scenarios, including a blockage that stops water flow through them, after all a failure would create a LOCA into the coolant pool.
Cyril R wrote:
It's odd stuff, it is mainly a biological risk to anything that's alive, but is surprisingly non-corrosive to most materials at ambient conditions. Reason why it makes for good sink declogger!

In the low concentrations required for scrubbing though it is not a major hazard (household chemical level).

I suppose since it will be dousing from the ceiling some equipment might be protected by shelters over it, if there is anything that might be damaged by it.
In a non blackout accident we could also provide a clean water 'rinse' that would simply serve to wash off the caustic from anything under it, reducing the time its sitting on components to a few minutes at most.
Cyril R wrote:
Makes sense. This thing is going to generate some hydrogen from radiolysis, so you can just integrate the tritium stripper into the hydrogen recombiner.

There has been a lot of work with similar systems for retrofit to CANDUs, and they have shrunk the equipment drastically - but using radiolysis hydrogen is an interesting idea that could reduce the equipment size further, although the processing rate might be a bit high to rely on it exclusively.

Cyril R wrote:
You might want to reconsider the CO2 power cycle. If all it offers you is a percent point increase in efficiency it is hard to justify the risk of a new power cycle. This was a major factor in the demise of the PBMR. Developing a new power cycle is serious business not to be undertaken lighly even if you have a billion in funding. There has to be serious advantage like a large fraction increase in efficiency for this to be worth the risk.

I just ran a simulation using the 170-180C moderator temperature and assumed no recompressor to see what would happen.
Efficiency comes out at 35.1% with 19C cooling water, and 36.2% with 5C cooling water.
This assumed 310C reactor outlet temperature, with a 305C peak gas temperature. If temperature could be marginally increased this would increase significantly.

However 25% of the heat input was consumed at below the moderator temperature (moderator cooling will be roughly 10% based on Atucha 1), which indicates that a recompressor could significantly raise efficiency, but that calculation is significantly more involved so I will attempt it tommorow.

But 40% winter efficiency would seem to be within reach. However the primary gain here is the ludicrous compactness of the plant, and with carbon dioxide turbines finally moving somewhere for natural gas use I think the turbine plant is less risky than it would have been say, five years ago.
I think the risk is worth it, after all the turbine plant on modern reactors is absolutely enormous.

EDIT:
If we assume ~320C core outlet temperature as in the Advanced CANDU, that implies a ~315C gas heater peak temperature.
Which increases efficiency to something like 36.3/37.4% Without a recompressor.

Considering ACRs use basically the same technology, such improvements might be possible with this reactor.
At which point with a recompressor we are beating down the door of 40% (gross) efficiency, and our only major house loads are the heat transport/moderator pumps and the cooling water pumps. We have no feedwater pumps such are defined as house loads in steam cycles.


Top
 Profile  
 
PostPosted: Feb 23, 2017 9:38 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Quote:
This is an interesting concept, but we would have to ensure that the tubes containing the water at ambient pressure were able to withstand the pressure inside the vessel without failure in a wide variety of scenarios, including a blockage that stops water flow through them, after all a failure would create a LOCA into the coolant pool.


This seems fairly easy, e.g. if each individual tube connects at both ends to the pool, you'll have massive redundancy.

tubes are pretty good at dealing with pressure... a 2 mm thick stainless steel tube of 50 mm OD will have a burst pressure of well over 300 bar at reactor temperatures. More likely though you will leave holes in the casting near the vessel ID so you can just slide the tubes in there and they would be protected by the casting metal.Then you also have a smooth face for any liner and insulation.

You might want to reach out to these guys for your casting:

http://www.siempelkamp.com/fileadmin/me ... _09_EN.pdf

Quote:
I suppose since it will be dousing from the ceiling some equipment might be protected by shelters over it, if there is anything that might be damaged by it.
In a non blackout accident we could also provide a clean water 'rinse' that would simply serve to wash off the caustic from anything under it, reducing the time its sitting on components to a few minutes at most.


Probably no need to douse rooms with equipment. Just connect the room to a tank via submerged sparger. The tank cover gas then vents to a filter to e.g. a stack. The room needs to be designed as a containment with low design pressure.

Quote:
I just ran a simulation using the 170-180C moderator temperature and assumed no recompressor to see what would happen.
Efficiency comes out at 35.1% with 19C cooling water, and 36.2% with 5C cooling water.
This assumed 310C reactor outlet temperature, with a 305C peak gas temperature. If temperature could be marginally increased this would increase significantly.


Cool. Can you try to see if the temperature goes up to say 600 C with a gas superheater? Just curious what efficiency you can get then (probably recompression would not be worth it anymore with that temperature?).

In terms of the Rankine competition. Isn't the latest batch of LWRs pushing 38% winter efficiency already? AREVAs product line is claiming 37% net efficiency.


Top
 Profile  
 
PostPosted: Feb 23, 2017 9:54 am 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Perhaps something like this arrangement?


Attachments:
sketch.png
sketch.png [ 22.92 KiB | Viewed 289 times ]
Top
 Profile  
 
PostPosted: Feb 23, 2017 6:00 pm 
Offline

Joined: Jun 19, 2013 11:49 am
Posts: 1419
Cyril R wrote:
This seems fairly easy, e.g. if each individual tube connects at both ends to the pool, you'll have massive redundancy.

The nice thing about these vessels is you tend to get massive redundancy all over the shop. The more I learn the more I come to the conclusion that they are inherently superior to standard steel pressure vessels.
It is certainly a neat idea, it all depends on how good liquid submerged insulation at 110bar is these days.
Anything with closed cells is out as it would be crushed at that pressure, so it is basically down to gas filed insulation (which I am not a fan of for reasons already given) or some sort of solid ceramic which might not be very good.
Cyril R wrote:
tubes are pretty good at dealing with pressure... a 2 mm thick stainless steel tube of 50 mm OD will have a burst pressure of well over 300 bar at reactor temperatures. More likely though you will leave holes in the casting near the vessel ID so you can just slide the tubes in there and they would be protected by the casting metal.Then you also have a smooth face for any liner and insulation.

Casting some cavities for pipework near the inner-face of the blocks seems like a good idea, it would certainly reduce the heat stress issues associated with the hot vessel.
Cyril R wrote:
You might want to reach out to these guys for your casting:

http://www.siempelkamp.com/fileadmin/me ... _09_EN.pdf

That has certainly improved the feasibility of this idea in my mind I think, if we can really cast 250t blocks of ductile iron then the number of components in the vessel is much smaller than I foresaw.
Cyril R wrote:
Probably no need to douse rooms with equipment. Just connect the room to a tank via submerged sparger. The tank cover gas then vents to a filter to e.g. a stack. The room needs to be designed as a containment with low design pressure.

You could adopt an Ontario Hydro type scheme and potentially have two or even four units served by the same Sparger.
Very large line breaks would be quite rare in operation I think..
Cyril R wrote:
Cool. Can you try to see if the temperature goes up to say 600 C with a gas superheater? Just curious what efficiency you can get then (probably recompression would not be worth it anymore with that temperature?).

These cycles do not do very well with superheaters - since you would end up with a recuperator exit temperature higher than that of the nuclear reactor. And the ratio of 7.5MPa/35MPa specific heaters at the LWR temperature range is only about 1.2.
So it would turn into a recuperated gas fired cycle with nuclear auxiliary heat, more than anything.
Cyril R wrote:
In terms of the Rankine competition. Isn't the latest batch of LWRs pushing 38% winter efficiency already? AREVAs product line is claiming 37% net efficiency.

I believe the Mitsubishi APWR is pushing 39% net efficiency in a very specific set of circumstances. But these things require absolutely enormous steam plants and I am not convinced it is worth it given the much smaller size of the turbine and gas heater plant in this system. Pressure vessel volume not used for huge conventional steam generators (PCSGs are quite hairy from what I can tell) can be used for more fuel channels!
The carbon dioxide system is an issue but if we put everything inside the confinement/containment building we can dilute the carbon dioxide presssure spike to a significant degree.

I recently found out that the Wylfa reactors of ~450MWe net each had over 6000 fuel channels, with 8 1m long elements per fuel channel - which gives you an idea of the difficulties of building high power density gas cooled reactors, but also makes me feel better about the feasibility of 1400 channel closures.

The only problem is that our core requires very expensive heavy water so we can't just build it to arbitrarily large sizes to achieve lower power densities.


Top
 Profile  
 
PostPosted: Feb 23, 2017 6:28 pm 
Offline

Joined: Jul 14, 2008 3:12 pm
Posts: 5006
Quote:
The nice thing about these vessels is you tend to get massive redundancy all over the shop. The more I learn the more I come to the conclusion that they are inherently superior to standard steel pressure vessels.


They're complicated though. Tendon collection galleries, cable galleries, prestressing equipment and inspection abilities. Compared to just a metal membrane that holds pressure.

They are also easy to screw up. Look up Crystal River NPP. The utility made a mess of that (I guess that is total stupidity and that isn't much of an argument against prestressed structure).

Quote:
It is certainly a neat idea, it all depends on how good liquid submerged insulation at 110bar is these days.
Anything with closed cells is out as it would be crushed at that pressure, so it is basically down to gas filed insulation (which I am not a fan of for reasons already given) or some sort of solid ceramic which might not be very good.


Well, there's many ways to skin this cat. Interestingly stagnant water (heavy or light) is a good enough insulator so all you have to do is trap water so it can't flow. That's common practice in process engineering, it is typically called thermal sleeving and is widely employed in piping, especially in nozzle to vessel connections... all you have to do is weld some stainless sheets with their ends blocked as a "liner" and then heavy water gets in, and stagnates to insulate. Simply increase the number of sheets to increase the insulation. The insulation doesn't have to be very good, just decent enough that it doesnt become an annoying parasitic heat loss for the liner. Unless you have some interesting low temperature heat demand somewhere... I would say anything under 1% reactor heat loss is fine.

What size vessel are you looking at? Even a quite large cast steel vessel won't need to be meters thick even to hold 200 bar and without prestress. A 10 meter casting of 0.5 meters thick would deal with your pressures easily without prestress cables.

Prestress is required for concrete high pressure vessels because the material has basically nil tensile strength. Metals are very different...


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 52 posts ]  Go to page 1, 2, 3, 4  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:  
cron
Powered by phpBB® Forum Software © phpBB Group