Adams Engine - Goal is cheap, ultra low emission fuel coupled to cheap machinery 1

83 Comments

  1. Thanks for producing a seperate thread Rod. The discussion in the other thread was highly off-topic.

    Regarding the hybrid idea. There’s another option that is fossil-free. In that possibility, the nuclear reactor operates at full throttle almost all the time. A seperate molten salt storage system stores part of the heat for peaking, using perhaps a dedicated peaking turbine, with the rest going to a standard baseload plant and seperate turbine. Single turbine configurations could also be explored to save costs (but efficiency may suffer with this approach, depending on the type of turbine).

    Various low cost molten salts are available for heat storage. NaNO3-KNO3 eutectic for a medium temperature design, and fluoride or carbonate salts such as FLiNaK or its carbonate cousing CLiNaK for higher temperature heat storage. Graphite for fluoride salts and sand fillers for carbonates, in thermocline direct tanks, can be used to reduce the thermal storage medium and tank costs.

    Regarding nitrogen activation. Dealing with this – in the sense of engineered components, maintenance issues and procedures, etc. is likely more expensive than simply using enriched nitrogen. Large streams of nitrogen are available industrially at air seperation plants. One need only use this nitrogen in a centrifuge train to enrich it. This will be much cheaper than heavy water production. In fact heavy water production would be quite cheap in a hydrogen economy, where a similar idea can be used: co-locate the deuterium seperation unit with the hydrogen production plant. The hydrogen “tails” are perfectly fine and can be sold normally.

    Regarding the use of lower pressure nitrogen in simple cycle. As I’ve argued before, with gaseous coolants, the heat you can remove from a reactor is basically proportional to pressure, for a given material limit. A reactor is much more expensive than a simple compact combustion chamber of a natural gas combustion chamber. In other words the priorities are much different.

    Furthermore, nitrogen has in the ballpark of 5x lower heat capacity (cP) and 5x lower thermal conductivity than helium.

    To show you what I am talking about. At 1 MPa, 250 C inlet, 750 C outlet, the specific volumetric enthalphy gain for nitrogen coolant is only 0.146 MJ/m3. This compares to a PWR specific volumetric enthalphy gain of 133 MJ/m3. This is 900x better despite the tiny temperature rise of 35 degrees Celsius. A BWR is around 2000x better even, and the temp rise over the core is less than 10 degrees Celsius.

    I’m a numbers based guy. I don’t like the numbers behind nitrogen cooling.

    1. @Cyril R

      I am also a numbers based guy, but the numbers that interest me are those associated with manufacturing and production, not enthalpy.

      The whole idea is to keep the system as simple as possible and to minimize cost by using machinery that is already in volume production and using a fuel form that can be mass produced.

      In our previous conversation you pointed to the current cost of Triso based fuels – not to the fundamental cost of the fuel form once it is being produced by an automated factory in quantities of millions to hundreds of millions of balls per month.

      Nitrogen is ubiquitous in nature – 80% of our atmosphere is nitrogen gas. It is essentially inert in its natural form of N2. Of that N2, 99.6% is N-14, only 0.4% is the N-15 that you propose using.

      C-14 is not a radiological hazard, especially inside of a closed cycle system. You mentioned that it might combine with any available oxygen in the system to form a gas, but the total amount of gas that can be formed in that manner is limited by the total amount of oxygen in the system. Sure, there will be a small amount of oxygen in any initial charge of nitrogen, but after operation and production of C-14 that oxygen will be scavenged.

      With a gas cooled system, there is not as much need for charging and discharging as the reactor is heated up as there is with a water cooled reactor.

      I fully recognize that there are many suboptimal design decisions in various portions of my proposed system. My analysis, however, shows that the whole system is reasonably well optimized for my main goal – competing with large diesel engines in areas where that is the primary available option for reliable electrical or motive power.

      1. Hmm, one of the blockquotes went awry there! Try again:

        I am also a numbers based guy, but the numbers that interest me are those associated with manufacturing and production, not enthalpy.

        They are related. Enthalphy gain determines size of components and number of pebbles for example. It’s one of the reaons supercritical steam turbines have catched on so well. In the supercritical regime, the enthalphy gain beats the pants off anything else.

        In our previous conversation you pointed to the current cost of Triso based fuels – not to the fundamental cost of the fuel form once it is being produced by an automated factory in quantities of millions to hundreds of millions of balls per month.

        Actually the current cost of TRISO fuel is astronomic because it’s custom order. The base case cost of $4100/kgU assumed volume production. The more aggressive case was lower @ 1300/kgU but still much higher than PWR fuel fabrication @ $200/kgU. Here’s an excerpt from the article:

        The cost of fabricating a PBMR fuel pebbleis largely dictated by the cost of fabricating the 11,000 to15,000 TRISO particles that it contains. Literature estimatesfor TRISO particle fabrication costs differ by orders of magnitude, so there is wide variation and potential for error inestimating the PBMR fuel cycle fuel fabrication costs. TheAFCCBR (Schneider et al 2008) high estimate for fuelfabrication costs is $30,000/kgU. Subtracting out theapproximate combined cost of $5,900/kgU for the ore, SWU,and other non-fabrication front end components that go intocreating the fuel, suggests a price of $24,100/kgU for fabrication only. Based on other studies of the PBMR (Kemm1999), this estimate appears high, and so this analysisassumes high case fabrication costs of $15,000/kgU. The basecase cost estimate of $4,100 was obtained by subtracting$5,900/kgU from the AFCCBR estimate of $10,000/kgU(Schneider et al 2008). The low cost case uses John Gittus’1999 analysis of Eskom’s estimate of $1,000/kgU, or $1,300/kgU in 2009 dollars. If South Africa deploys largenumbers of PBMRs, economies of scale and learningexperience could allow low fabrication costs. However,Eskom’s estimate of $1,300/kgU appears aggressive evenwith such advantages.

        Nitrogen is ubiquitous in nature – 80% of our atmosphere is nitrogen gas. It is essentially inert in its natural form of N2. Of that N2, 99.6% is N-14, only 0.4% is the N-15 that you propose using.

        Yes but it’s far worse for heavy water (0.015%) and you also don’t get to work with the gas there. Like I said, if you can work with the gas (nitrogen or hydrogen) enrichment becomes easy. Enriching nitrogen may be cheaper than installing equipment to clean up C-14.

        C-14 is not a radiological hazard, especially inside of a closed cycle system. You mentioned that it might combine with any available oxygen in the system to form a gas, but the total amount of gas that can be formed in that manner is limited by the total amount of oxygen in the system. Sure, there will be a small amount of oxygen in any initial charge of nitrogen, but after operation and production of C-14 that oxygen will be scavenged.

        C-14 has an activity of 166,5 TBq/kg (4500 Ci/kg). So the admittedly guesstimated figure of 38000 TBq/GWe-year would amount to 228 kg of C-14 per GWe-year. It looks like you’re right, this is big enough to be a dominant oxygen sink.

        But it’s so much bigger than I thought – over a million Curies – probably big enough to cause issues of the stuff building up locally as a radiation hazard to maintenance. In fact 228 kg is so large I think you would want to inject oxygen or something else to grab it. This would require a lot of R&D to figure out (not to mention convince the regulators, including health and safety officials, who will not like millions of curies building up all over your turbine, core etc.).

        228 kg is also so large that it affects breeding noticeably; about 16 kg of neutrons will be lost to this capture. 16 kg of neutrons could fission 3700 kg of U235, for example, which is enough for 4 GWe-year.

        Perhaps some of my figures are wrong, but this means the reactor is wildly inefficient, thus burning through more and more enrichment, TRISO fuel fabrication, and uranium fuel than even I feared.

        1. Say, Cyril, would you allow me to pick your brain?

          My pseudonym, lower-case and minus the hyphen, plus 38215 at yahoo.

  2. Rod,

    I am a big fan of your design. When the fuel is really cheap, we don’t need high thermal efficiency. The simplicity of the system allows it to be installed in areas where there are few people to monitor it and / or the technical level of the people watching the system is comparatively low. The ability to load follow on a small grid is VERY important. Isolated villages and or islands will have a very small load at times but if the system can “grow with the load” without having to install a new system that would be wonderful. That is to say, you could install a system that has a 80MW total peak power output, but would be able to scale it’s output from 0MW to 80MW depending on the connected load and do the load following automatically. You now have a system that can really help an island to develop! A small highly trained crew can watch and maintain the system, keeping things in repair while the people can just plug in their new stuff with no worries. Stores and Factories (small at that level) can be up and running without worries. Less power drawn down means the fuel lasts longer. I can think of at least a thousand places to put one of those babies.

    1. TRISO fuel and high enrichment is not cheap. Using natural nitrogen as the coolant will easily increase fuel usage a factor of 2, aggravating the problem.

  3. Somebody please check my math about C-14. The above numbers of 38000 TBq/GW-yr make me think the reactor would generate one million Ci/GW-yr. Total C-14 on earth is about 300 million Ci.

    If the reactor generates energy cheaper than coal, I’m for it.

    1. Yes, about a million Curies per GWe-year. This is based on extrapolating from LWR nitrogen content. Which may err on the positive side for nitrogen cooling, as the nitrogen in PWR coolant is also in the moderator, thus suffering from extra resonance captures. In a TRISO fuelled reactor, coolant and moderator are seperate, reducing resonant absorptions. But not by much especially for smaller pebbles as the mean free path of neutrons is longer than that of a LWR, which will also tend to increase resonant absorptions…

      As a sanity check, the production of plutonium-239 in a LWR is also in the ballpark of 200 something kg per GWe-year. It’s precursor’s thermal neutron cross section (U238) is very close to that of nitrogen-14.

  4. Wouldn’t the nitrogen gas closed cycle turbine work equally well with an IFR or LFTR?
    My understaning is that they both have the ability to change power up & down quickly.

    Also in places with a shortage of cooling water I would think an open cycle turbine using air would have the advantage of not needing large & expensive heat exchangers at the low temperature end. This would require heating the air by a secondary coolant loop to keep radioactive materials from the open air, but might well be cheaper. Are there any show stoppers to this idea?

    1. It some ways it would work even better. Things like reheat become much more attractive with molten salt to gas heat exchangers than with gas-to-gas heat exchangers. MS-gas exchangers are more compact and have lower pumping cost, so the economic gain is much higher for reheat in a LFTR than in a direct cycle gas cooled reactor.

      This is also true for IFR, but much less strongly so, due to the lower achievable temperatures, and the lower heat capacity of sodium (but on the plus side, there are less issues with sodium freezing than with salts freezing which means a less restricted design space).

    1. Supercritical CO2 is probably the best of the gasses, due to it’s good density that removes much of the specific enthalphy gain problem. Not anywhere near as good as supercritical water, but orders of magnitude better than a low pressure nitrogen reactor.

      The biggest problem at the moment appears that supercritical CO2 turbine cycles are only just starting to get to tiny prototypes. That’s a big downside in technological and industrial readiness, compared to supercritical water.

      1. Cyril,

        Yes Supercritical CO2 is “exciting” but – expensive and hard to maintain in a remote setting. Again, the simplicity of Rod’s design and the “off the shelf” nature of the parts, ESPECIALLY the turbine make this design superior in remote settings. Remote as in currently off the grid and with poor roads / or boats going to it. A LFTR with chemical processing would NOT be suitable in these locations.

        Robert,

        Yes, this would be cheaper than coal because the whole transportation thing would not have to be created to bring the coal to the remote locations. I know of a coal burning plant in Jasper IN that is only 15MW. But the coal coming to it required a rail line for the supply. At the least it would require a pretty good road for the heavy trucks. The cost of these infrastructures would be bypassed with Rod’s design and the benefits of Electricity could grow and then attract developers who would build the roads to their factories.

        1. Expensive and hard to maintain – why? I see no reason to believe this. The fact that s-CO2 cycles are not yet commecially available is a good argument, but there is nothing inherent in s-CO2 that makes them more difficult to maintain and expensive than a nitrogen closed cycle Brayton.

          Quite the opposite should be true: they are more compact than ideal gas Braytons, and we know from chemical considerations that they will be easier to maintain the steam turbines.

          The only inherent problem I can think of is that the power density is a little TOO high. The turbine is so compact that the stress on the blades is very large. This could turn out to be a serious problem.

          1. Hi Cyril,

            Thanks for the reply, I am batting way above grade here.

            1. Expensive because as you point out they are not commercially available. Using a new system in a Nuke makes the cost factors for that system skyrocket. Replacements when they do break, or come to end of life, are going to be enormous.

            2. Hard to maintain – because of the target areas that would use an A. A. E. – remote places with no or very expensive electricity, like islands I have been to that are powered by Diesel generators that break down because the locals don’t know how to maintain them well. In these contexts if your CO2 leaks for any reason, recharging the CO2 would be far far past any technical abilities in the area and would require very specialized equipment. On the other hand, recharging with ordinary Nitrogen is not that technically difficult as illustrated by the number of tire dealers who put Nitrogen into tires to help them last longer. I am not saying that tire equipment could be used but pointing out that getting pure Nitrogen is a much easier technical process than Super Critical CO2.

            3. You may have a strong point in the overall maintenance of the equipment if the issues with C14 are as large as I think you are making them. That is way past my knowledge.

            Please understand that in many of these contexts you are not competing with coal – you are competing with Diesel fuel. Electricity prices are close to 40 to 45 cents / kwh. The reliability of the Diesel’s is fairly low – they go down often. So even if your costs for electrical production are 10 cents or even 15 cents a kwh, for the consumer 20 cents a Kwh would be a great reduction in price and for businesses the reliable electricity would be a great boon.

            If the system uses ordinary parts off the shelf and chemicals that are easily obtained the chance of using local labor to repair problems is much better. Otherwise, you will need some kind of high tech mobile repair boat / airplane that flys around the world to keep all your plants in order. If you sell hundreds or thousands of these devices – as I hope Rod does – the fleet of maintenance vehicles could be super costly.

            I guess, I watched the South African project crash mainly because of the enormous cost of helium turbines being developed. I really don’t understand the need to worry about using 2 times as much fuel – when the fuel cost is only a fraction of your costs. While at the same time pushing for an “ideal gas” that costs billions to design and manufacture a specialized turbine that will only be used a few times in the world – thus always super pricy! Even with the QA processes currently used for Nuclear Grade parts – an off the shelf version of an ordinary turbine is a series of special steps on an existing assembly line.

            Thanks for a great conversation.

        2. Remote as in currently off the grid and with poor roads / or boats going to it. A LFTR with chemical processing would NOT be suitable in these locations.

          But a DMSR might well be.  The quantities of make-up uranium are so small, they could be flown in by bush planes.

          1. @ Engineer-Poet,

            Yes, a Denatured Molten Salt Reactor could also work. I think Hyperion’s design is really good as well. I got super excited about Nuclear when I understood it could be made so simple to operate. In some places like PNG the Bush plane will be the only way for decades due to the mountains worse than West Virginia. But in those areas a micro hydro is a much better option since the elevation drop is so great and the rainfall so heavy. But in islands like the Philippines, or Indonesia boats and planes are the two options. For nuclear materials, it will be boats. I doubt you could get past the “panic” to put fuel on a plane.

            I can see from later comments that Cyril’s focus is on a different market and because of that we are talking past each other a bit. But he is one very intelligent individual. I admire Rod for trying to discuss this without giving away all his design paramaters.

  5. One more thing to think about, for those interested in a natural nitrogen cooled graphite moderated (TRISO) reactor.

    Chernobyl had a positive feedback from losing water. This is because water absorbs more neutrons than graphite. Thus losing water from the core means more efficient moderation, resulting in the runaway at Chernobyl.

    Natural nitrogen has a much larger neutron absorption than water, and no moderation to partly offset it.

    It follows that a loss of coolant in a natural nitrogen cooled, TRISO fuelled reactor would suffer from severe positive void coefficient.

    Quite possibly, it would make Chernobyl look benign.

    The more I research the topic, the more I understand the international focus on helium and CO2 for gas cooled reactors.

    1. Certainly a positive void coefficient is a real concern, and one that should not be overlooked, but I don’t think that it is as bad as you are making out.

      There’s nothing inherently unsafe about a positive void coefficient. CANDU reactors operate with a positive void coefficient, and so do the remaining RBMK’s that are still in operation today.

      The reason that the reactor at Chernobyl had such a high void coefficient was because it was running on natural uranium. To fix this problem, the Russians increased the enrichment and made a few other modifications. Problem solved.

      You overlook a serious difference between water and nitrogen as a coolant. When we talk about a “void” in water, we’re talking about a phase transition from a liquid to a steam bubble that happens rapidly and can occur very locally. Meanwhile, to get a “void” in a gas coolant that results in an equivalent decrease in neutron-absorber density requires a depressurization to near atmospheric pressures — that is, an emptying of the coolant from the entire primary circuit, which is not local and requires significant time to finish.

      Graphite-moderated gas-cooled reactors tend to have a pretty good negative temperature coefficient of reactivity. In the calculations that I have run on these reactors for a loss of coolant accident, the core shuts itself down very quickly once the temperature starts to rise, even if the control rods are not inserted.

      Don’t forget that the disaster at Chernobyl was caused by a steam explosion. The power excursion was only an indirect cause that lead to higher temperatures. In a gas-cooled reactor, there is no water in the primary circuit to cause a steam explosion.

      Although I have performed no calculations on this, my intuition tells me that any such increase in reactivity due to loss of nitrogen pressure would result in a localized temperature increase (remember that we’ve also reduced the cooling) that results in a negative feedback for reactivity through Doppler broadening and other effects.

      You’re going to have to do much better to convince me that this is a serious problem.

      1. to get a “void” in a gas coolant that results in an equivalent decrease in neutron-absorber density requires a depressurization to near atmospheric pressures — that is, an emptying of the coolant from the entire primary circuit, which is not local and requires significant time to finish.

        Unfortunately, that’s not the case. A simple heatup from eg loss of forced circulation or loss of heat sink, would greatly reduce density of the coolant. No large depressurization is needed as nitrogen-14 is very much more absorptive than water, and also has no negative void from moderator expulsion to compensate for it…

        A LOCA or ATWS or LOHS or LOFC would be unacceptable for this reactor.

        Graphite-moderated gas-cooled reactors tend to have a pretty good negative temperature coefficient of reactivity. In the calculations that I have run on these reactors for a loss of coolant accident, the core shuts itself down very quickly once the temperature starts to rise, even if the control rods are not inserted.

        Certainly Doppler is a useful and fast negative reactivity insertion, but if your void coefficient is bigger than Doppler – and it looks like this is the case – then your total reactivity is still very much positive. Even a neutral net reactivity would be a big problem in the above mentioned scenarios.

        Your assertion relates to helium cooled TRISO fuelled PBMRs. That is different because helium has essentially no void, it’s actually slightly tiny bit negative, which is good. Like I said, the international focus on helium is very understandable. You’ve basically got the game simplified to a Doppler reactivity game, simplifying safety analysis. The coolant won’t be much of a player, and any air ingress would reduce reactivity. Nitrogen-14 is a different beast. It’s absorption, much greater than water, would definately overrule Doppler, especially due to the lack of loss of moderator.

        Intuition is very dangerous Brian. I’m sure the Chernobyl operators acted on their intuition in doing their disastrous experiment, but they knew nothing of k-calculations.

        And by the way… those modified RBMKs, last I checked them, still have positive void coeffient… just not as strongly anymore due to higher enrichment.

        As for CANDUs they are not so bad because heavy water has a low absorption for neutrons, and the heavy water results in such excellent thermalized neutron spectrum that everything happens on a snails pace anyway. A PBMR has a faster spectrum than a CANDU so not much help here either.

        1. @Cyril R

          You’re still focusing on microscopic cross section for absorption, not macroscopic. You are also failing to credit the low pressure choice, which substantially limits the maximum amount of reactivity that can be inserted by a loss of coolant pressure.

          There are not enough moles of N2 in the core to have much of an effect on reactivity. The rate at which that number of moles can change is very low.

          1. Convince me. Show me a Monte Carlo or other raytrace model result that does not give large dollar reactivity insertion for partial and/or total voiding. Say your reactor operates at 1 MPa and loses coolant to 0.3 MPa. As I’ve told you before I don’t believe it is viable to operate such a poor coolant with such a low pressure, so all of this may not fly. In fact, to get a similar performance as a helium PBMR at 7 MPa you will have to operate at over 20-30 MPa.

            Bear in mind that your reactor is filled with poison (N-14), so you need extra reactivity to even start it up. Take that into account in the reactivity analysis.

            As for macroscopic cross section… N2 a molecular mass of 28 g/mole whereas H2O has a molecular mass of 18g/mole. Not a big difference.

            If there are not many moles of N2 in your core, you will not produce useful amounts of power. I don’t know in how many ways I should explain this further.

            1. @Cyril R

              What does “voiding” mean when the fluid you are talking about is already a gas? The only thing that can happen is that the pressure in the gas drops, removing a portion of the total number of moles.

              In a densely packed pebble bed core, the gas volume is 39% of the total volume of the core.

              If the core is primarily pyrolytic graphite with a density of 2.2 gm/cm3 and the coolant is N2 gas with an average temperature of 873 K (600 C) at 1 mPa, the total mass in each liter of the core is 1343 gms. Of that total, just 0.7 gms is N2 gas.

              I’ll leave it to the readers to determine just how small the reactivity change will be if there is a mass change of just 0.05% and the material that leaves the representative section of a core on a loss of pressure has an absorption cross section of just 1.8 barns.

              Your comparison of the molecular masses of N2 and H20 ignores a very important factor – one is a gas and one is a liquid.

              At 1 mPa and 873 K, a mole of N2 requires a volume of roughly 7.14 L

              A mole of H20 fits into a volume of just 20 cc or 0.02 L.

              A liter of water has a mass that is 500 times as large as a liter of N2 at 1 mPa and 873 K. (Note: I am running out of time before another appointment, so I did not make any temperature corrections for the density of water.)

              The key to having a small number of moles remove a useful quantity of heat is to move those moles through the volume rapidly with a large delta T between the inlet and the outlet. As we all should know, the power output from a heat exchanger is equal to mass flow rate x specific heat capacity x (Tout – Tin).

              Water has a better Cp, but gas can move at a higher speed and a gas cooled heat exchanger can reliably provide a much higher delta T.

              When you add the fact that a much larger core can be economical if the pressure retention requirements are substantially reduced, there are a lot of moving arrows that need to be understood before you can make a blanket determination of which is more economical.

          2. @ Cyril and Rod,

            Can you two experts give a lay person some idea of the total size of the reactor? What size is a BWR and what size is an Adams Atomic Engine? All in with the whole kitten kaboodle? I mean when you keep talking about a physical difference in size is the A.T.E. 10 times larger? 2 times larger? What are we looking at?

            1. The core power density of the pebble bed reactor that is part of the Adams Engine concept is about 3-5 MW / cubic meter.

              An 5-15 MWe system, including everything needed to produce electrical power output, could fit into a cube that is 10 M on each side.

          3. Convince me. Show me a Monte Carlo or other raytrace model result that does not give large dollar reactivity insertion for partial and/or total voiding.

            Well, I’ve been wondering this myself, so I set up a model in MCNP5 to explore this. It’s a very simple model (i.e., suitable for a comment on a blog, perhaps not more, and I can send you the input decks if you like), but it should serve to illustrate the first-order magnitude of the factors involved here.

            My results confirm my intuition that the effect of removing the nitrogen on the eigenvalue is small compared to the effect of raising the temperature. Keep in mind that the margin for temperatures in a TRISO gas-cooled reactor is hundreds of degrees C. Thus, any increase in reactivity will result in temperature increases that eventually swamp the effect of the removal of nitrogen, before the temperatures reach a point where fuel failure is a concern.

            To give more details, I considered a very simple model using Rod’s parameters of a 1 cm diameter pebble in nitrogen coolant at 1 MPa of pressure. To determine the effect of nitrogen, I compared a calculation with nitrogen coolant at the specified temperature and pressure in the spaces between pebbles with a calculation that uses no nitrogen at all — i.e., total voiding, which is the bounding condition. I examined several enrichments from a typical LWR value of 4 wt% to near the upper limit of LEU of 19%. Although there was a substantial increase in the effect of the nitrogen as the enrichment was lowered, at enrichments above 8 wt% (the value used for HTR-Modul, and a low value for gas-cooled reactors), the effect of temperature was clearly more significant.

            Of course, I might have made some mistakes in setting up this model, so the usual caveat emptor applies. I don’t claim to be perfect. I only claim to have run a model, which is more than anyone else has bothered to do before commenting here.

            I agree, Cyril, that intuition is very dangerous, especially when one’s intuition is coming from water-cooled-reactor land. Welcome to the world of gas-cooled reactors! Things are a little different here. Please take your time to get used to the place, but I think that you’ll find it easier to get along if you refrain from insulting the intelligence of the locals. 😉

          4. My results confirm my intuition that the effect of removing the nitrogen on the eigenvalue is small compared to the effect of raising the temperature. Keep in mind that the margin for temperatures in a TRISO gas-cooled reactor is hundreds of degrees C. Thus, any increase in reactivity will result in temperature increases that eventually swamp the effect of the removal of nitrogen, before the temperatures reach a point where fuel failure is a concern.

            That’s not really true. The limiting transients (ATWS under LOFC) result in peak particle temperatures of around 1500 to 1600 degrees Celsius (slightly higher was confirmed by some experiments using refractory metal strips with melting point near 1600 Celsius that had melted after the test).

            This is with helium, which has a significant negative reactivity effect upon voiding. It’s a quite good moderator, much better than graphite atom per atom, and would be more widely used if it had a higher density. Helium also has a higher thermal conductivity. Bear in mind also that the convective heat transfer upon loss of flow with 1 MPa nitrogen is extremely poor. That’s a moment when the high flow velocities, made possible by high flow velocities and radial vessel design, are not available.

            I’m not sure about your model because I can’t see any of it. How many pcm did the model give upon full voiding? Did you take into account the absorption of nitrogen in terms of the excess reactivity needed to offset that, compared to helium (which reduces the reactivity needed compared to vacuum)? What dimensions and segments did you use for the core in the model?

            I agree, Cyril, that intuition is very dangerous, especially when one’s intuition is coming from water-cooled-reactor land. Welcome to the world of gas-cooled reactors! Things are a little different here. Please take your time to get used to the place, but I think that you’ll find it easier to get along if you refrain from insulting the intelligence of the locals.

            I have no intention of insulting anyone. I have an interest in figuring out the truth. I have, however, little patience for the sort of pedantism that you displayed earlier; talking about insulting people – dissing an article – from the gas cooled reactor world, if you must know- over a misplaced comma (!) is not appropriate.

            I am not a water cooled reactor guy. More of a molten salt guy, but that’s largely because molten salt promises to be a better coolant than water. Coal plants are water cooled, so we have to come up with something at least as good to compete on even grounds. I like water cooled reactors mostly because we’ve got them now and can start building more now, and because they’re good at making plutonium for future Gen IV startup (including gas cooled reactors if economical). I hope the gas cooled reactors end up economical, this would be another ace in the hand against fossil fuels.

          5. That’s not really true. The limiting transients (ATWS under LOFC) result in peak particle temperatures of around 1500 to 1600 degrees Celsius

            The limiting transients that result in fuel temperatures in that range are depressurized LOFC transients, or D-LOFC. It’s what I call Depressurized Conduction Cooldown (DCC).

            Helium also has a higher thermal conductivity. Bear in mind also that the convective heat transfer upon loss of flow with 1 MPa nitrogen is extremely poor. That’s a moment when the high flow velocities, made possible by high flow velocities and radial vessel design, are not available.

            The conductivity of helium is negligible in a DCC event, as is its convective heat transfer. In fact, in the event that the vessel remains pressurized, nitrogen’s poor heat transfer characteristics will be beneficial to the vessel, because that is the component that takes the biggest hit from a pressurized LOFC transient, P-LOFC (which I call Pressurized Conduction Cooldown, PCC).

            I’m not sure about your model because I can’t see any of it. How many pcm did the model give upon full voiding? Did you take into account the absorption of nitrogen in terms of the excess reactivity needed to offset that, compared to helium (which reduces the reactivity needed compared to vacuum)? What dimensions and segments did you use for the core in the model?

            This is a simple k-infinity calculation, the goal of which was to determine the relative effects. The parameters were chosen to produce conservatively high effects from the nitrogen gas. The geometry is a 1-cm-diameter pebble, and I assumed that this was a standard PBMR pebble that was shrunk to a 1-cm size. That is, I used the typical mass and uranium content that is commonly cited for a PBMR pebble to determine the material densities.

            The pebbles are arranged in a simple cubic structure, which is not only easy to model, but results in a packing factor of 0.52, which means more helium surrounding the pebble than typical, more dense structures. This would tend to overstate the reactivity effect of the nitrogen.

            In examining the thermal effects, I considered only changes in cross-sections. I made no attempt to consider the effect of thermal expansion on the densities of the fuel and the moderator.

          6. The limiting transients that result in fuel temperatures in that range are depressurized LOFC transients, or D-LOFC. It’s what I call Depressurized Conduction Cooldown (DCC).

            The difference between the pressurized or depressurized ATWS was very small, both were around 2100 degrees Celsius.

            The conductivity of helium is negligible in a DCC event, as is its convective heat transfer. In fact, in the event that the vessel remains pressurized, nitrogen’s poor heat transfer characteristics will be beneficial to the vessel, because that is the component that takes the biggest hit from a pressurized LOFC transient, P-LOFC (which I call Pressurized Conduction Cooldown, PCC).

            Yes, gasses are poor coolants, hence the near 2100 degrees celsius figure for both pressurized and depressrized conditions. However, during ATWS the vessel will almost certainly fail… 700 degrees Celsius on ferritic (and even austentic stainless) steel is way overasking the limits of this material. So the appropriate end state for PBMR ATWS is the depressurized condition with failed primary loop.

            This is a simple k-infinity calculation, the goal of which was to determine the relative effects.

            Brain, please keep in mind that small changes in k-inf have very large changes in Rho. And you need to consider leakage too. For example, consider the reactor is operating at k-eff of 1. Then the nitrogen voids and k-eff becomes 1.01. This may not sound like much, but it adds 900 percent milliRho!!! To offset this you only have the alpha of fuel, if it is -6 pcm/K then your temperature will rise an additional 150 degrees Celsius in the ATWS transient. You’d be pushing 2300 degrees Celsius in stead of ~2150 (well the exact temperature will have to be calculated iteratively but you get the point).

            With a k-eff change to 1.02 you’ll have to offset 1960 pcm. 326 degrees Celsius rise over the already too high transient temperature with helium PBMR!!

  6. Continuing. With a molecular weight of N2 of 28 and H2O at 18. We have a factor of 1.56 reduction of number of moles for N2.

    The thermal neutron capture cross section for water being 0.34. For nitrogen being 1.91. An increase of 5.62 for nitrogen.

    Thus, 5.62/1.56 = 3.6 times the absorption of water.

    Gasses have lower density, but that doesn’t help much; you just need a lot more volume of gas in your core. In fact the volume is huge even with a factor of 20 increase in temperature rise over a PWR.

    Water is a powerful moderator – the most powerful in fact in terms of macroscopic slowing down power. Nitrogen does almost nothing. So you don’t get a partially compensating negative reactivity effect from reduction of moderation during voiding…

  7. Proof of claims. There has not been much work on nitrogen cooled reactors, but there has been much work on nitride fuels especially for fast reactors.

    http://www.neutron.kth.se/publications/library/NitrideSafety.pdf

    Notice the high void worth of over +1000 pcm with natural nitrogen – far greater than the Doppler term is negative.

    This is for nitride fuel, which has only nitrogen in the fuel, not in the much larger mass of coolant.

    It is also for a fast spectrum, where losses through n,p are lower, even considering the contribution of the fast resonance area:

    http://www.nndc.bnl.gov/sigma/getPlot.jsp?evalid=14972&mf=3&mt=103&nsub=10

    The report’s conclusion: nitride fuels can be used safely, if enriched nitrogen-15 is used….

  8. Your comparison of the molecular masses of N2 and H20 ignores a very important factor – one is a gas and one is a liquid.

    As I tried to tell you, I ignored the gas density because it is compensated by the larger volume.

    You unfairly compare liters of core, when the gas cooled reactor at 1 MPa has very roughly 200x more volume than the PWR core. 0.5 kWt/l for nitrogen gas @ 1 MPa yields roughly the same thermal limits as a helium gas @ 7.5 MPa & 5 kWt/l. The PWR core also has less than 39% coolant fraction to add to that.

    A loss of coolant accident results in a pressure equalization between the containment and the reactor. If the containment has a design pressure of 1/3 the normal operating pressure of the gas coolant, it follows that you get around 2/3 “voiding” (agree it is a bit strange wording for a gas cooled reactor).

    Yes, a liter of core voided isn’t much nitrogen, but there are so many more liters of core in your reactor that it basically negates itself. At LWR operating temps the density is around 700-800 kg/m3. This is around the same factor of 200-250 or so in density difference.

    I gave you a reference that looked at nitrogen voiding for a nitride fuelled fast reactor. This reactor has far fewer kg of nitrogen in the fuel than your reactor has in the coolant, and the coolant is easier lost than the nitrogen in the fuel. Yet the study found an unacceptably large void worth of >1000 pcm for nitrogen-14. It also found that you could get the void to be almost zero for nitrogen-15.

  9. The key to having a small number of moles remove a useful quantity of heat is to move those moles through the volume rapidly with a large delta T between the inlet and the outlet. As we all should know, the power output from a heat exchanger is equal to mass flow rate x specific heat capacity x (Tout – Tin).

    It doesn’t help enough. Helium reactors already do this (500 C temp rise). A BWR with a 5 C temp rise still beats the pants off it. Most people don’t realise how good a coolant boiling water is. I used the helium case as a base case where to compare the nitrogen coolant under the same thermal limits, so I already accounted for a 500 C rise across the core.

    Higher temp rise results is severe thermal stresses especially during transients and accidents, where the core average temperature can equilibrate to the entire core, resulting in serious thermal stresses of all sorts of core internals, piping, nozzles, control rod guide tubing, heat exchangers etc.

    1. @Cyril R

      Don’t forget that I provided the core configuration information already showing how a radial inflow core works. Many of the characteristics of that configuration mitigate the thermal stress issues that you mention.

      1. What happens if the flow is lost? If your vessel is the decay heat removal path as in PBMR, then the hot helium has to flow backwards to reach that vessel. The vessel will then see transient gas spikes touching the vessel wall of >700 Celsius more than the normal operating temperature.

        1. @Cyril R

          Thermal stresses in a pressure vessel and associated ducting are determined by additional variables; not just the magnitude of the temperature changes on the surface.

          Yes, if the forced flow stops in an operating reactor core with radial inflow of moderate pressure nitrogen, the temperature of the core will generally tend to equalize, though the boundaries will still be at a lower temperature than the center. However, the pressure vessel for a system designed for a maximum pressure of say 2 MPa is not the same thickness as one designed for 7-10 MPa. In this core, the coolant flow pattern is optimized to reduce flow resistance. It has a large annulus in the center where the temperatures will be the highest.

          Gas will continue to flow through the core, even if the compressor stops running and the core pressure drops to the same pressure as the rest of the system. Even if all coolant pressure is lost, the system will be full of atmospheric pressure gas that has essentially the same specific heat transfer capacity as the normal coolant gas. The inlet temperature will actually fall compared to operating inlet temperature since the compressor has stopped adding heat to the coolant.

          The heat removal capability will be reduced from the operating heat removal capability, but fission would have stopped on the loss of the compressor so the only heat that needs to be removed is decay heat.

          1. My question has to do with the specific nature of a radial flow design, specifically that it cools the vessel by the colder inlet gas. If flow coasts down, the vessel temperature will be exposed to much higher, near outlet temperatures.

            If you are familiar with ASME code limits you’ll know there are some very stringent demands on thermal stresses. A sudden increase of several hundred degrees Celsius in vessel temperature at or above operating pressure would be unacceptable. Depressurization is typically used in water cooled reactors to address this, but if you do this with a gas cooled reactor the heat transfer will be cut one to two orders of magnitude so you end up with higher temperatures.

            Typically PBMRs use the vessel as a heat removal path for decay heat post shutdown. If the vessel is thin enough, you might be able to prevent excessive heatup at the gross material thickness, but you still have to deal with quite high thermal gradients in that vessel for this solution… cold air on the one side, gas hotter than lava on the other…

            It’s just a difficult design consideration for this radial flow arrangement.

          2. My question has to do with the specific nature of a radial flow design, specifically that it cools the vessel by the colder inlet gas.

            The idea that the inlet gas is somehow “cooling” the vessel is a common misconception about modular HTGR’s.

            In a typical modern HTGR, pebble or prismatic, the “cold” inlet flow runs between the metallic core barrel and the pressure vessel. However, rather than cooling the vessel, this coolant flow actually serves to heat it during normal operation. Why do I say this? Because once the forced coolant flow is removed, the vessel immediately begins to cool.

            What is often overlooked in these designs is that the plant is set up to constantly remove heat from the vessel. This is an important passive safety feature of the reactor. For example, during normal operation of a 600 MWt HGTR, over 1 MW of heat is constantly being removed from the vessel by the Reactor Cavity Cooling System (RCCS). Pretty much all of this heat comes from the “cold” inlet coolant flowing up along the inside of the vessel, which warms the vessel.

            The direction of flow for a radial design is quite different from a traditional downward-flowing HTGR, but I expect that the inlet flow will still have this warming effect. Whether it actually does or not depends on the details of the design, which I don’t know.

          3. The idea that the inlet gas is somehow “cooling” the vessel is a common misconception about modular HTGR’s.

            I’m being pedantic, but, no it’s not a common misconception. The colder inlet flow helps remove gamma and neutron heating of the vessel, which, for a gaseous coolant vessel is considerable. This heat source alone is certainly greater than the 1 MWt the RVACS removes, for a 600 MWt design (for reference, CANDU-6 shields around the reactor remove 3-6 MWt of neutron and gamma heating). That’s cooling in the strict sense (removing heat). In terms of protecting against high temperatures (cooling in a loose sense) the colder inlet flow helps prevent excessive temperatures, and importantly, thermal gradients, compared to a simple straight axial flow. In such an arrangement, the top of the vessel would be over 500 degrees Celsius hotter than the bottom, which would violate various ASME code limits for thermal stresses. The inlet flow effectively insulates against the high outlet temperatures and temperature gradients…

            PWRs and BWRs also apply this trick, but it’s not so severe for these reactors with such low temperature rise, so the top part of the vessel actually operates at outlet temperatures.

            But let’s not fuss over definitions. The more important question is what this design will do to the vessel temperature during a transient, which you haven’t looked into. If the RVACS cools the vessel, then you will get a natural circulation path going on with hot gas moving back towards the cooled vessel wall. Then the vessel wall will suddenly “see” gasses at over 1000 degrees Celsius, 750 degrees or more above the temperature it normally “sees”. So you have a possible thermal shock, differential gradients across the vessel wall that induce differential stresses through the wall. A detailed transient temperature and resulting stress analysis would have to be done, to see if it is acceptable for this design…

          4. Wow! You’re so off base that it is difficult to know where to begin.

            Are you talking about a radial inflow core that Rod mentions, or are you talking about a traditional modular gas-cooled reactor that uses axial flow?

            I was referring to the traditional design. Your use of the acronym RVACS is somewhat telling and hints at your inexperience with these designs. I guess I shouldn’t have been surprised when you referred to events as LOCA’s.

          5. Are you talking about a radial inflow core that Rod mentions, or are you talking about a traditional modular gas-cooled reactor that uses axial flow?

            The radial flow one.

            I was referring to the traditional design. Your use of the acronym RVACS is somewhat telling and hints at your inexperience with these designs. I guess I shouldn’t have been surprised when you referred to events as LOCA’s.

            RVACS is reactor vessel cooling. PBMR uses a variant of this.

            LOCA is an appropriate term for any reactor, as even with 1 MPa gas pressure, a break of piping would result in losing most of the coolant. Gasses are coolants just like liquids are. Loca for gasses can be much more severe because you can’t benefit from boiloff cooling or from low pressure liquid coolant design (that uses pools of liquid to prevent total loca).

            You are being pedantic and even arrogant again. Please don’t do that. I am allergic to such behaviour, especially when I know where you are wrong and have not addressed my arguments (rather you chose to insult me and focus on definitions that distract from the main arguments I am making.

            1. Gasses are coolants just like liquids are. Loca for gasses can be much more severe because you can’t benefit from boiloff cooling or from low pressure liquid coolant design (that uses pools of liquid to prevent total loca).

              You are again talking about high pressure gas cooling, not the low pressure gas cooling that I advocate. (In a very lonely fashion, I might add.)

              When normal heat extraction from a fission reactor is nitrogen gas at 1 MPa, there is a readily available pool of coolant available to handle decay heat in the case of a loss of pressure accident. Air, which will always be available, is a gas with nearly exactly the same specific heat capacity as the original coolant.

              As Brian pointed out, there is a large margin between normal operating temperature and the temperature at which any fuel damage starts to happen. When forced flow stops, the reactor will heat up and shut down. As long as action is not taken to shut the flow path, there will still be some coolant passing through the core. That flow will be sufficient to remove enough decay heat to prevent any fuel failures. It will also be sufficient to keep the core pressure boundaries from overheating or experiencing a heat up that causes thermal stresses that exceed its material limitations.

              In this configuration, there really is no way to have a complete loss of coolant accident (a LOCA) since that would require a physically impossible to create vacuum. (I am not proposing that Adams Engines would be well suited for space applications, by the way.)

          6. When normal heat extraction from a fission reactor is nitrogen gas at 1 MPa, there is a readily available pool of coolant available to handle decay heat in the case of a loss of pressure accident. Air, which will always be available, is a gas with nearly exactly the same specific heat capacity as the original coolant.

            As Brian pointed out, there is a large margin between normal operating temperature and the temperature at which any fuel damage starts to happen. When forced flow stops, the reactor will heat up and shut down. As long as action is not taken to shut the flow path, there will still be some coolant passing through the core. That flow will be sufficient to remove enough decay heat to prevent any fuel failures. It will also be sufficient to keep the core pressure boundaries from overheating or experiencing a heat up that causes thermal stresses that exceed its material limitations.

            Asserting that what remains to be demonstrated for this reactor configuration. LOCA means you have to shut down the pumps. Then you lose the extremely fast flow rate that you are proposing to remove the needed heat at helium cooled PBMR power densities. The higher the normal flow rate, the bigger the issue with loss of forced circulation because the natural circulation flow speed will be so much smaller than normal forced flow. This is the case also when there is no LOCA, for example an ATWS with LOFC. This event will also vent much of the coolant out (as per the Ideal Gas Law) through safety valves to prevent excessive pressurization challenging the vessel and piping. If the coolant has positive void worth, then that adds fission power.

            But keep in mind that the limiting transients for PBMRs are due to lack of decay heat removal from either pump failure or loss of heat sink. It is decay heat, combined with the limited capability of the RVACS to remove heat, that pushes the maximum temperature. It actually takes about 2 days to get to this temperature. It is about 1500-1600 Celsius at that time, then the decay heat removal through the vessel matches the decay heat rate so temp goes down. 1600 Celsius is near the point where fuel failure starts to occur.

            Lower pressure coolant means worse natural circulation during loss of flow. The radial reactor configuration will only be of limited help there as you lack the high flow speeds available with forced circulation to make up for the lower coolant heat removal. Inherently, radial flow configurations also have less natural circulation capabilities (the flow first has to reverse, and the heat source and heat sink will not enjoy large thermal leg advantage).

            It’s also not clear that fission will stop after loss of forced circulation, with positive void N-14 being vented either through a pipe break or safety relief valves. If the void is stronger than the doppler negative term, and from the reference I provided this appears likely, fission power will increase. If the void is as strong as the doppler negative term, fission power will stay at the same level. In either case the fuel kernels will rapidly damage and melt, releasing volatile fission products.

            With helium, void is decidedly negative, so there’s a double whammy with a switch to a poisonous coolant. Brain appears to have access to RELAP so if the assumptions are all modelled correctly it might be possible to figure this out.

            1. @Cyril R

              It’s also not clear that fission will stop after loss of forced circulation, with positive void N-14 being vented either through a pipe break or safety relief valves. If the void is stronger than the doppler negative term, and from the reference I provided this appears likely, fission power will increase. If the void is as strong as the doppler negative term, fission power will stay at the same level. In either case the fuel kernels will rapidly damage and melt, releasing volatile fission products.

              I don’t know if you are still following this thread, but I have received a partial answer to your question about the reactivity effects of pressure transients in a pebble bed core cooled by N2 gas at approximately 10 bar.

              I didn’t burn this core as it is most reactive at BOC. I added a little burnable poison in the fuel and reflector but it would need more to hold down the BOC reactivity. The nominal core multiplication factor at full power BOC is 1.147. This is high but within range of a control rod bank.

              I then perturbed the core by dropping the N2 pressure down to atmospheric to simulate a full depressurization. keff bumped up to 1.155, an increase of about 84 cents. Definitely enough to initiate a power transient but hardly a DBE. Thermal feedback and automatic control systems would easily overcome this. Modeling such a transient would take a different set of tools and some time but I think you can dispel the rumors of catastropic nitrogen reactivity insertion.

          7. The radial flow one.

            Well, I was referring to the conventional, downward-flowing modular high-temperature gas-cooled reactor, which should have been clear from my comment. For these reactors, the inlet flow does not cool the vessel. It warms it. During normal operation, the peak temperature of the vessel is lower than the inlet temperature.

            But if you were talking about the radial flow one, then why did you refer to the 1 MTt being removed by the RCCS of a 600 MWt design?

            (In case you’re wondering, I’m referring to your statement, “This heat source alone is certainly greater than the 1 MWt the RVACS removes, for a 600 MWt design.”)

            Can you please point me to some details about the 600 MWt radial-flow gas-cooled reactor design? Since I occasionally work on these types of designs — in fact, I currently have a couple of papers being reviewed for publication — I’m really curious to see any open-literature descriptions of this design, which I’ve never heard of. I’d hate to think that I’ve missed something. Thanks!

            You are being pedantic and even arrogant again.

            I am not the one trying to lecture Rod on his reactor design, and you’re the one who has claimed to be pedantic.

          8. But keep in mind that the limiting transients for PBMRs are due to lack of decay heat removal from either pump failure or loss of heat sink.

            The limiting transient depends on component that is being challenged. If you are worried about the fuel, then a complete depressurization of the primary system is the limiting transient, not a loss of forced cooling.

            It is decay heat, combined with the limited capability of the RVACS to remove heat, that pushes the maximum temperature.

            No. An operating Reactor Cavity Cooling System (RCCS) has no trouble removing the heat, since it is conservatively designed for its limiting transient, which in this case is a variation of the pressurized loss of force cooling scenario, which results in lower peak fuel temperatures than the depressurized case. The real problem is getting the heat to the vessel.

            It actually takes about 2 days to get to this temperature. It is about 1500-1600 Celsius at that time, then the decay heat removal through the vessel matches the decay heat rate so temp goes down.

            That is not so. Even the graphs in the Syd Ball paper that you have cited shows that this is not the case.

            The rate at which heat is removed from the vessel by the RCCS is almost completely decoupled from the decay heat because of the large thermal mass and thermal resistance of graphite between the active core and the vessel. Even the most basic analysis can demonstrate this.

            The temperatures finally peak when the thermal gradient through the outer reflectors finally reaches a point (through the gradual heating of these reflectors) that the heat can be driven (relatively) rapidly through this region. At that point, the heat can finally begin to be transferred out of the central part of the core, where, in an annular core design, the center graphite reflectors have been soaking up this heat and keeping the fuel temperatures from becoming too high.

            Inherently, radial flow configurations also have less natural circulation capabilities (the flow first has to reverse, and the heat source and heat sink will not enjoy large thermal leg advantage).

            In the limiting transient for the fuel in a conventional modular gas-cooled reactor, i.e., complete depressurization, the effect of convection by the remaining gas is so weak that it is completely neglected in the analysis. In the pressurized scenarios, which are more challenging for other components, the consequences will depend heavily on the geometry, which still hasn’t been well defined for this radial-flow, nitrogen-cooled reactor that Rod mentions.

          9. But if you were talking about the radial flow one, then why did you refer to the 1 MTt being removed by the RCCS of a 600 MWt design?

            (In case you’re wondering, I’m referring to your statement, “This heat source alone is certainly greater than the 1 MWt the RVACS removes, for a 600 MWt design.”)

            Radial flow would also have the inlet flow going all around the vessel wall. So it is also “heated” by the inlet gas. But the inlet gas is colder than the outlet gas, so the vessel operates much cooler than with a simple once-through in/out design. However the >1 MWt for 600 MWt reactor core refers to the neutron and gamma heat load deposited in the vessel and reactor cavity for RCCS type cooling systems. I suspect it will be more than 1 MWt from just gamma and neutron heating, but it depends on the design specifics (reflector thickness etc.)

            Can you please point me to some details about the 600 MWt radial-flow gas-cooled reactor design? Since I occasionally work on these types of designs — in fact, I currently have a couple of papers being reviewed for publication — I’m really curious to see any open-literature descriptions of this design, which I’ve never heard of. I’d hate to think that I’ve missed something. Thanks!

            Rod provided a link for a chemical radial flow reactor design that he wants to use for his reactor.

            http://www.johnsonscreens.com/sites/default/files/6/705/Internals%20for%20Radial%20Flow%20Reactors.pdf

            One thing I wonder about with this design for a nuclear application is that it seems very poor in natural circulation removal of afterheat, compared to an axial flow design where you get a cooled cold leg (vessel area) and a heated hot leg that promotes natural circulation.

            I am not the one trying to lecture Rod on his reactor design

            I am not lecturing rod. Those are your words. It is my job to point out areas in industrial design that need attention. This is not lecturing, it is constuctive criticism and alleviating concerns. Rod has clearly given them a lot of thought, and most of my concerns have been alleviated, though specific areas of concern remain. nitrogen-14 activation (specifically radioactive dust in the reactor system and positive void) and very high flow speeds (power consumption) in the piping, nitrogen circulator/blower, and afterheat removal in a radial flow reactor design that gets poor natural circulation. None of these are technically showstoppers, but they may reduce the economics of the design (through high parasitic consumption, poor fuel utilisation, poor power density, use of enriched nitrogen etc.). If Rod wants to compete with diesels then there’s big margin for economics there.

          10. The limiting transient depends on component that is being challenged. If you are worried about the fuel, then a complete depressurization of the primary system is the limiting transient, not a loss of forced cooling.

            If you had bothered to look at the ATWS graphs from Syd Ball, you would see that the difference in end state fuel damage between pressurized and depressurized is only a few percent, 57% and 59% respectively. Since ATWS is the limiting transient, and since it will fail both fuel and vessel (due to overheating) to a high degree, your statement is clearly proven wrong by Syd Ball.

            I wrote that before, but you ignore most of the arguments I make. It feels like sittng with a child in a merry go round. Every time you make the same argument because your arrogence causes you to have problems reading. This is sad. I think we can have a more honest discussion Brian.

            No. An operating Reactor Cavity Cooling System (RCCS) has no trouble removing the heat, since it is conservatively designed for its limiting transient, which in this case is a variation of the pressurized loss of force cooling scenario, which results in lower peak fuel temperatures than the depressurized case. The real problem is getting the heat to the vessel.

            It clearly cannot remove the heat from the ATWS event, because much of it is fission heat, which cannot be adequately removed with a RCCS system. Only decay heat, with no fission power, can de dealt with by the RCCS alone.

            That is not so. Even the graphs in the Syd Ball paper that you have cited shows that this is not the case.

            The rate at which heat is removed from the vessel by the RCCS is almost completely decoupled from the decay heat because of the large thermal mass and thermal resistance of graphite between the active core and the vessel. Even the most basic analysis can demonstrate this.

            Uhm I’m sorry Brain but you’re wrong again. In the LOFC transient provided by Syd Ball and also by many others, the peak fuel temperature rises to around 1500 degrees Celsius or so, then drops because the heat removed by the RCCS is increased to the point of matching decay heat generation. Even the most basic analysis can demonstrate this.

            In the limiting transient for the fuel in a conventional modular gas-cooled reactor, i.e., complete depressurization, the effect of convection by the remaining gas is so weak that it is completely neglected in the analysis. In the pressurized scenarios, which are more challenging for other components, the consequences will depend heavily on the geometry, which still hasn’t been well defined for this radial-flow, nitrogen-cooled reactor that Rod mentions.

            Which means nitrogen coolant is more likely to fail the vessel (by overheating and overpressurizing) than helium coolant. This is going to restrict the design. We don’t want the vessel to fail.

          11. I wrote that before, but you ignore most of the arguments I make. It feels like sittng with a child in a merry go round. Every time you make the same argument because your arrogence causes you to have problems reading. This is sad. I think we can have a more honest discussion Brian.

            No, we cannot, because I keep trying to explain these things to you and you refuse to listen. In particular, you refuse to listen when I point out that your “arguments” are simply wrong — very wrong. The reward for my effort is to be insulted by being called “childish” and “arrogant.”

            Hell, you’ve even got the chutzpah to say that Syd Ball, who wrote the one paper that you have bothered to reference to support your “arguments,” doesn’t know what he’s talking about in the conclusions of his own study!

            Frankly, I give up. It is pointless to argue over the internet with someone who is so stubborn that they cannot bother to learn anything, because they believe that they already know it all.

            If I had a blackboard available and plenty of time, perhaps I could adequately explain some of the physics involved in these designs with some diagrams and equations and give you some idea of how they are typically modeled. Trading comments on a blog just isn’t going to cut it.

            Sorry, but you’re just some random loon on the Internet, and I do these calculations for a living. Publish a paper on this topic, and perhaps we can talk at the conference at which it is presented. Until then, I’m afraid that my pro bono lessons to you on how gas-cooled reactors really work just isn’t worth my time and patience.

            Goodbye armchair reactor designer.

  10. Rod,
    Without getting into the neutronics of the core, have you thought about PCHE (Printed Circuit Heat Exchangers) They are compact with over 700 W/m^2-K and efficiencies of 98%.

    http://www.heatric.com/hres/Heatric%20standard%20brochure

    I am doing some RELAP modeling of these puppies coupling PRISM to a Thermal Energy Storage system. A commercially available technology that would shrink your heat exchangers by about 85% and increase the power to weight of the engine.

    1. @Cal

      I am always willing to entertain technological improvements. There is only one place in the system for a heat exchanger, but it is admittedly a large component that is probably going to be a significant cost contributor.

    2. PCHE’s are amazing. They are not only much more compact and reduce pressure drop and pump power, but also eliminate more than 99% of the leakage events in conventional heat exchangers, namely through heat affected weld zones, seals, gaskets, brazing… it does this by not needing any of these techniques.

      For nuclear applications, there is an interesting licensing discussion going on with these heat exchangers. They are very difficult to inspect in service, a requirement for licensing. It seems that in service inspection wouldn’t be necessary as all the major leakage/failure modes of conventional heat exchangers are eliminated with PCHEs…

  11. What size is a BWR and what size is an Adams Atomic Engine? All in with the whole kitten kaboodle? I mean when you keep talking about a physical difference in size is the A.T.E. 10 times larger? 2 times larger? What are we looking at?

    The core of Rod’s design would probably be around 0.5 kWt/liter. A BWR is around 50, so 100x as large. PWRs are around 100 kWt/liter, 200x as large.

    A 125 MWe helium cooled PBMR design has a core power density of around 5 kWt/liter. It is 20 meters tall and 6 meters wide. It operates at 2 MPa higher pressure (9MPa) than a BWR yet the BWR produces around 1000 MWe for that size vessel. PWR produces around 1200 MWe for such vessel size (but needs big steam generator vessels).

    The ESBWR is 27 meters tall and 7 meters wide. Not much larger than the PBMR yet produces 1550 MWe versus 125 MWe. The ESBWR containment building is actually shorter than the PBMR building.

    Rod’s poorer coolant (nitrogen) combined with lower operating pressure means around 10x lower core power density than PBMR. So a PBMR sized version of Rod’s design would produce around 13 MWe. Actually the efficiency of the Brayton is lower, so 10 MWe is a more reasonable guesstimate for the power output.

    So here’s the choice, a same size vessel that produces 10 MWe or 1000 MWe. Which would you like.

    On the plus side, Rod’s turbine is much more compact than the low pressure Rankine cycle of the BWR. Rod may also increase the power density by a factor of up to 4 by reducing the pebble size, but it’s not clear to me how far this could be stretched (hundreds of brilliant scientists have settled on 6 cm pebbles, there must be good reason for this).

    1. @Cyril R

      The core power density of a pebble bed reactor is not limited by the ability of the coolant gas to remove heat from the surface of the pebbles. Other factors govern.

      The power density of the reactor in the Adams Engine concept is the same as other gas cooled pebble beds.

      I say again – take a good hard look at the drawings of nuclear power plants that you can find at the below link:

      http://econtent.unm.edu/cdm4/browse.php?CISOROOT=/nuceng

      You should notice that the size of the reactor core is a very tiny portion of the complete power plant for a light water reactor with a steam plant secondary and all of the accompanying systems that add the many layers of protection that make those systems safe.

      Bigger CORES with lower power density and high temperature fuel do not necessarily imply poor economics or low system power density.

      Nuclear energy systems do not necessarily compete against other nuclear energy systems; the Adams Engine is designed to be a cost effective, environmentally superior alternative to large diesel engines. In order to compete in that market, it has to have many of the same characteristics of a diesel engine in terms of simplicity of construction and ease of operation.

      As I stated in the post that started this conversation, I am not currently pursuing the concept and may never get around to pursuing it. I remain convinced, however, that it is a pretty decent system with some real advantages.

      BTW, do you know much about the air cooled reactors that have been operated for extensive periods of time with excellent safety records. Those machines use a coolant that is 80% naturally occurring nitrogen, which is 99.6% N-14 with no isotope enrichment.

  12. The core power density of the pebble bed reactor that is part of the Adams Engine concept is about 3-5 MW / cubic meter.

    So you have a 10x worse coolant heat removal that uses the same power density core.

    What does that do to the TRISO peak temperature? It pushes it over the thermal limits specified (you need to operate well below 1600 Celsius peak particle temperature in normal operation to have some margin to transients).

    5 MWt out of the pebbles is based on 8 to 9 MPa helium.

    The core power density of a pebble bed reactor is not limited by the ability of the coolant gas to remove heat from the surface of the pebbles.

    I don’t mean to be pedantic, but, yes it is limited by that. The heat removal determines the peak operating temperature. Better heat removal means more power density for the same peak operating temperature.

    The heat removal given that limit basically scales with pressure, for gaseous coolants. If 8 MPa helium gets 5 MWt/m3 then 2 MPa helium gets 1 MWt/m3 with about the same peak particle temperature. Nitrogen is a worse coolant than helium at a given pressure, because it’s thermal conductivity is 4 to 5 times lower than helium.

    I think your smaller pebble idea would help a bit, but there’s a limit to the size based on breakage from pebble bed weight, and you need an unfuelled sleeve to protect the particles against erosion from thermal cracking, swelling of graphite and the like. PBMR uses a 5 mm sleeve so that leaves 0 mm left for fuel for a 10 mm sphere. But perhaps the sleeve can be thinner if the pebbles don’t have to be moved about by refuelling systems?

    You should notice that the size of the reactor core is a very tiny portion of the complete power plant for a light water reactor with a steam plant secondary and all of the accompanying systems that add the many layers of protection that make those systems safe.

    Bigger CORES with lower power density and high temperature fuel do not necessarily imply poor economics or low system power density.

    That is what I am seeing. The PBMR building is taller than the ESBWR building, even though the former produces an order of magnitude less electricity. See the Per Peterson study on materials and building volume. GT-MHR does reasonably well, about similar to the ESBWR in specific materials usage, but a larger nuclear input and larger building size per MWe.

    Nuclear energy systems do not necessarily compete against other nuclear energy systems; the Adams Engine is designed to be a cost effective, environmentally superior alternative to large diesel engines. In order to compete in that market, it has to have many of the same characteristics of a diesel engine in terms of simplicity of construction and ease of operation.

    I have to admit that my own interest lies in replacing coal with nuclear. Coal plants are big, cheap, complicated and use an economical coolant, supercritical light water. I want nuclear plants to compete here, which means going cheap and going for more economical coolants, not worse ones. Simplicity is good but it won’t help in this application if it results in order of magnitude increase in all equipment, building, pumps, and piping sizes.

    I honestly think that your idea has merit, if certain changes are made to the design, specifically nitrogen enrichment and higher operating pressure. Diesel is expensive so your economics case will the easier. At the same time, history has shown that reactors and also gas turbines have big scaling factors – economy of scale – so it’s interesting to see if the cost is still lower than diesel.

    BTW, do you know much about the air cooled reactors that have been operated for extensive periods of time with excellent safety records. Those machines use a coolant that is 80% naturally occurring nitrogen, which is 99.6% N-14 with no isotope enrichment.

    I know them not. Please elaborate.

    1. The heat removal determines the peak operating temperature. Better heat removal means more power density for the same peak operating temperature.

      I do not disagree with the above statement; I do disagree with your characterization of moderate pressure nitrogen (~1 MPa) flowing in a radial inflow reactor consisting of elements as I have described them (1 cm vice 6 cm diameter) as having worse heat removal than high pressure helium flowing from one end to the other of a tall core.

      What I said

      The core power density of a pebble bed reactor is not limited by the ability of the coolant gas to remove heat from the surface of the pebbles.

      was not phrased as I should have phrased it – I meant that the peak operating temperature is not determined by the specific heat transfer coefficient of the coolant gas. The heat removal equation has additional variables that can be adjusted.

      1. The radial flow concept is good for reducing pressure drop, where you could get an order of magnitude improvement, but it’s not such a large factor improvement in heat removal per unit volume of core. The amount of gas is a big factor there, and the amount of gas scales almost proportionally to pressure.

        I think the biggest remaining factor would be the size of the pebbles – yes this could get you the order of magnitude improvement over my estimate, but again I don’t know if 1 cm pebbles are feasible from a mechanical loading perspective. The PBMR pebbles are 60 mm and have 5 mm unfuelled sleeve to protect against erosion, thermal cracking and other mechanisms that could dislocate TRISO particles out of the discrete pebble configuration. That is obviously something to be avoided. Can’t have loose TRISO particles falling in bulk to the bottom of your reactor where you can’t remove them and they will increase power there (actually the TRISO particles would likely remain intact if the number is limited).

        The point being that with the same 5 mm unfuelled graphite layer, a 10 mm particle would have not fuel particles in it!

        The good news is that much of that 5 mm is needed to protect the pebbles against milling into each other with online refuelling, a problem that your design doesn’t have…

        1. The amount of gas is a big factor there, and the amount of gas scales almost proportionally to pressure.

          Once again, the amount of heat that can be removed from a heat exchanger is a rather simple and well understood equation:

          Power = mass flow rate x specific heat capacity x (Tout – Tin)

          (It is also equal to U x A x (delta T))

          Shorter flow paths and a broader flow area enables substantially higher velocities for the same pressure drop. Higher velocity enables the same mass flow rate with a lower pressure.

          (BTW, though the specific heat capacity of He is 5 times as high as N2, the mass per mole of N2 is 7 times as large. The volumetric heat capacity for N2 is thus 7/5 or 1.4 times as high as helium for the same pressure in the gas.)

          1. Aw, now I get it. You’re not going for a lower pressure drop, you’re using a higher flow rate and keeping pressure drop similar to axial flow PBMRs. This clearly works for the core, yet you’d have some pretty serious flow velocities going on in the piping and blowers… you’d have to especially check if blower power doesn’t get excessive.

            (BTW, though the specific heat capacity of He is 5 times as high as N2, the mass per mole of N2 is 7 times as large. The volumetric heat capacity for N2 is thus 7/5 or 1.4 times as high as helium for the same pressure in the gas.)

            Agree, I forgot to compensate for molar mass there. The lower thermal conductivity (4-5x lower) is still there though, compared to helium. Perhaps it more or less cancels out with the higher volumetric heat capacity.

  13. To David’s reply – agree on all points except for maintenance. CO2 can be delivered via standard gas bottles, as used in welding and such. It takes up less space than bottles of nitrogen. Nitrogen can be liquified for crygenic delivery, but that is not simple technology on the level that you’re considering.

    If areas don’t have the skill to maintain a diesel generator, they will certainly not have the skill to maintain even the simplest of nuclear reactors and turbines. An old diesel generator – not a common rail injection computerized machine of modern days – is about as simple as you can get in terms of maintenance. Combustion gas turbines are potentially lower maintenance, but it takes skill to work with high RPM turbines, gears, aligning axles correctly with submillimeter precision, etc.

    A CO2 turbine’s compactness would allow all major compontents to be shipped by truck or even by pickup truck (usually lots of pickup trucks around in remote places). You could install most components by hand, without any cranes. A 10 MWe turbine fits in a pickup truck, and all loose components can be lifted by 1 or 2 men.

    Here’s a few pictures of the PBMR:

    http://holbert.faculty.asu.edu/eee460/dfg/pbmr_factory.jpg

    http://www.schillerinstitute.org/graphics/conferences/070915_Kiedrich/ferreira/four_pack_layout.jpg

    http://www.neimagazine.com/Pictures/web/b/a/x/PBMR6a.jpg

    Doesn’t look much simpler than a BWR to me. In fact it looks very similar, with only minor exceptions such as the lack of moisture seperators and steam dryers.

  14. Rod, have you done any criticality and enrichment calculations on this small TRISO core?

    The 300 MWt helium cooled PBMR uses 8% enrichment.

    Your design would be under 50 MWt. That means a lot more neutron leakage. Plus you have the captures on nitrogen-14.

    Are you sure you can make this reactor go critical on <20% enrichment with TRISO fuel?

  15. There’s a great presentation from MIT on the pebble bed reactor:

    http://web.mit.edu/pebble-bed/Presentation/HTGR.pdf

    Over 2.7 million manhours of engineering to date with 450 equivalent full-time staff (including major subcontractors) working at this time

    •Over 12,000 documents, including detailed P&IDs and an integrated 3D plant model

    •Detailed Bill of Materials with over 20,000 line items and vendor quotes on all key engineered equipment

    12000 documents and 20000 items. So much for pebble beds being simple things. Notice the engineering drawing on page 40: looks very complicated!

    I also found a concept that used helium cooling with an indirect nitrogen Brayton cycle. “ACACIA-INDIRECT: A SMALL SCALE NUCLEAR POWER PLANT FOR NEW MARKETS”

    http://www.ecolo.org/documents/documents_in_english/HTR-Aliki-Baytron.pdf

    This would solve the issues with natural nitrogen coolant in a different, way, by employing a heat exchanger – nice opportunity for a PCHE HX application. This is much closer to the experience of Peach Bottom, Fort St. Vrain and the THTRs, that means less R&D needed (except for the HX but it seems commercially available from Heatric).

    Rod may face some competition.

    1. @Cyril R

      Thanks for the link. I’ll add it to my extensive library of resources detailing the best ways to unsuccessfully develop high temp gas cooled reactors. Not that I have had any more success with Adams Engines, but our total investment so far is less than 500K (my labor is not priced).

      1. Perhaps it would help if you started off with an even smaller power rating than 10 MWe? If the project is smaller it will be easier to get results with a limited project funding from DOE/military outfits.

        Most remote diesels are under 5 MWe, if memory serves. Even the ones at military bases tend to be quite small in unit size.

  16. Here’s an interesting issue with smaller pebble sizes that I mentioned various times but we haven’t discussed further.

    The PBMR pebble is 60 mm in diameter and has an unfuelled protective coating (just graphite without fuel particles) of 5 mm. This coating is needed because graphite is not very strong and we don’t want cracks or other damage to dislocate fuel particles to unknown areas in the reactor. This gives it a fuelled matrix volume of 57.9%.

    A 10 mm pebble with the same 5 mm unfuelled coating would have a fuelled matrix volume of 0%. Clearly that won’t work.

    Now, Rod has mentioned not refuelling online, this helps to reduce friction wear on the pebbles. Suppose we can use a 2 mm thinner coating because of this.

    The coating then becomes 3 mm thick, resulting in a fuelled matrix of 4 mm in diameter. The fuelled matrix volume is still only 6.4% of the volume. Thus the 10 mm static pebble contains 9x less fuel per unit volume than the 60 mm moving pebble. This is clearly not enough.

    Perhaps the coating can be even thinner at just 2 mm, where the fuelled pebble matrix becomes 6 mm. The fuelled volume is still only 21.6% of the pebble volume, 2.68x less fuel per unit volume than the 60 mm pebble. Still not enough…

    Perhaps this is part of the reason for the at first glance seemingly large pebble size used in the PBMR.

      1. Agree, silicon carbide is a much better (harder, stronger) coating. It also protects against moisture ingress; it forms SiO2 which is highly protective up to 1600 degrees Celsius… eliminating an important chemical problem in TRISO fuel.

        However, this means some R&D on the fuel is still needed… one of the advantages of the standard TRISO fuel pebbles is you can use qualified fuel, no need for further work.

        Fortunately, the NRC does not require ASME standards on core internals… though I’ve always wondered why silicon carbide micro pressure vessels are supposed to comply with the standards.

  17. This document describes some thermohydraulics of pebble bed flow:

    http://maxwellsci.com/print/rjaset/v4-1659-1666.pdf

    From equasion (1) it is clear that the pressure drop is inversely proportional to pebble diameter. Thus, the 1 cm pebble will have 6x higher pressure drop than the 60 mm pebble. The density (p) is has a similar effect (x2 actually) so the slightly lower density of nitrogen @ 1 MPa (compared to helium @ 7.5 MPa) is also a disadvantage. Mass flow (m) is also disadvantaged with nitrogen being heavier than helium. Advantageous factors are the larger flow area from a radial design and the shorter flow path. Looks like the larger flow area would more or less be cancelled out by the higher mass flow needed, and shorter path length would be more or less cancelled out by the smaller pebble size. The pressure drop coefficient would be worse for nitrogen than helium based on Reynolds numbers. But using the equasions modified for nitrogen, the absolute pressure drop is probably under 200 kPa which is perfectly acceptable.

    The convective heat transfer coefficient from equasion (8) shows that the heat transfer coefficient is also inversely proportional to pebble size. However, the larger pebble bed surface area per volume scales much faster than that to compensate, so you end up with more heat transferred despite the poorer heat transfer coefficient.

    Smaller pebble size and shorter bed flow path improve thermal conductivity which is important during accidents. Looks like Rod’s design can deal with afterheat despite the poorer coolant natural circulation due to smaller pebbles, especially because Rod’s design is small (high vessel area per unit decay heat).

    Good news for Rod.

  18. According to Sensitivity Studies of Modular High-Temperature Gas-Cooled
    Reactor (MHTGR) Postulated Accidents by Syd Ball from Oak Ridge National Laboratory,

    4.4 P-LOFC with ATWS: In this PBMR design, recriticality occurs at about 28 hours, and T(fuel)- max reaches 2127°C at 103 hr. Maximum vessel temperatures are also higher, 711°C at 145 hr. Fuel failure after 7 days was 57%. Variations in this accident are sensitive to fuel and moderator temperature-reactivity feedback coefficients. As with the GT-MHR, if after recriticality the SCS is started (with still no scram), peak fuel temperatures would exceed limits even more due to the selective undercooling.

    4.5 D-LOFC with ATWS: Recriticality occurs at 31 hr. In this case, T(fuel)-max is 2166°C at 137 hr, and the maximum vessel temperature (496°C at time = 168 hr) was still rising slowly after a week. Fuel failure at the end of the week was 59%.

    Both ATWS cases, the first one with intact vessel at pressure and the second in depressurized condition, very large (more than half) fuel failure occurs. The first one will likely fail the vessel anyway as carbon steel has no strength left at 710 degrees Celsius, so depressurized end state is appropriate even with no initial pipe breaks.

    This is not inherent safety. Failure to scram means large fuel failure and vessel failure are almost guaranteed. Fuel failure means fission product in coolant, vessel failure means fission product in containment… a big mess a la Fukushima.

    The extreme temperatures of >2100 Celsius are for helium with about -2 pcm/K coolant coefficient. Imagine what it looks like for nitrogen with +5 or even +10 pcm/K coolant coefficient…

    1. While Syd Ball is a highly respected name in gas-cooled reactor circles, you’re not doing him any favors by misinterpreting his study.

      Notice that this is a sensitivity study. It’s purpose was not to make realistic predictions. Observe what Syd wrote in his conclusions:

      The differences in the predicted absolute values of peak temperatures (for both fuel and vessel) for the two concepts for given accident scenarios should not be taken as definitive, since their finalized design features, such as vessel insulation strategies, have not been factored into the simulations. Other aspects of the predictions, such as assumed irradiated core thermal conductivities, temperature-reactivity feedback functions, and heat-sink related emissivities, are also dependent on many factors that should be considered in detail for specific design and operating conditions.

      The value of sensitivity studies at this point (i.e., early) in the design and analysis is to provide estimates of the uncertainties in the predictions, and to guide further efforts in improving the design as well as the accuracy of the predictions.

      It’s difficult to make sense of the results without knowing what assumptions, parameters, and material properties Dr. Ball was using. Was the graphite fully irradiated or not (which has a huge impact on the accident scenarios)? What emissivities were used?

      Often, when running these calculations, the analyst simply assumes the worst value for every parameter. The result of these assumptions is specularly high (and very unrealistic) temperatures. I know, I’ve done these calculations myself. For a sensitivity study, it doesn’t matter, because the purpose is to determine how sensitive the system is to changes in the parameters. The actual values don’t really count, except as sideshow curiosities.

      I don’t know what parameters were used by Dr. Ball in this study, but the suspiciously low values that he used for the bypass of the GT-MHR lead me to guess that he was running his calculations will all or mostly conservative assumptions. Once again, this is not unusual, but it means that his results do not indicate what would be expected to actually happen during an accident.

      Anyone who would claim from this paper that “this is not inherent safety. … a big mess a la Fukushima” clearly hasn’t read the paper very carefully, particularly not the abstract:

      Both of the MHTGR designs studied – the 400-MW(t) Pebble Bed Modular Reactor (pebble) and the 600-MW(t) Gas-Turbine Modular Helium Reactor (prismatic) – show excellent accident prevention and mitigation capabilities because of their inherent passive safety features.

      And that is the take-away message folks.

      1. I’m terribly sorry Brian but Syd Ball’s own figures don’t support his conclusion. 59% fuel failure with ATWS is huge.

        In stead a better conclusion would be that PBMRs are very safe if the control rods work, and if the fuel burnup is modest. PBMRs are reasonably safe with high burnup fuel if the control rods work.

        That was a surprising result for me. I’d previously assumed that PBMRs do well in ATWS events. I thought they were walk away safe. This assumption came from me first reading the PB-AHTR response to ATWS. It was very benign with the worst case ATWS being under 1500 degrees Celsius peak particle temperature.

        My takeaway message here is you need a good coolant. Since nitrogen is far worse in natural circulation than helium, it doesn’t look good for nitrogen ATWS events.

        1. That was a surprising result for me. I’d previously assumed that PBMRs do well in ATWS events. I thought they were walk away safe. This assumption came from me first reading the PB-AHTR response to ATWS. It was very benign with the worst case ATWS being under 1500 degrees Celsius peak particle temperature.

          Cyril R

          As Brian pointed out, Dr. Ball was not performing a safety analysis on a finalized design. He was performing a sensitivity analysis study on a preliminary configuration, which often involves using a model with every input being set at the most conservative value. By doing that, it is easier to see what parameters make the most difference by making repeated runs of the same model and adjusting one parameter at a time.

          There is one very important parameter that is different between the 600 MWth reactor and the ones I am talking about – I was envisioning reactors with a power output of 3-200 MWth, with most of my focus on reactors in the 50 – 75 MWth range. I was also envisioning core power densities on the very low end of those proposed for gas cooled reactors. Those two parameters have a rather large influence on peak fuel temperatures after an ATWS event.

          One more thing – fuel damage in well made TRISO particles does not start until temperatures exceed 1600 C for an extended period of time. As long as you keep temperature below that level, you have proven passive safety. Keeping them to a maximum of 975 C does not buy you any additional safety.

          1. As Brian pointed out, Dr. Ball was not performing a safety analysis on a finalized design. He was performing a sensitivity analysis study on a preliminary configuration, which often involves using a model with every input being set at the most conservative value. By doing that, it is easier to see what parameters make the most difference by making repeated runs of the same model and adjusting one parameter at a time.

            Most of the sensitivity analysis showed very small temperature differences with varying the variables. This shows the analysis is fairly robust. So I don’t really see the strength of your argument here. It’s true I can imagine some design changes that would help a lot, such as stainless steel thermal shunts coupling the barrel with the vessel, and I don’t quite understand why such simple upgrades haven’t been used.

            One interesting result is that the GT-MHR does much better in ATWS transients, with only around 1700 degrees C fuel temp, compared to 2100 in PBMR. The PBMRs reduced contact area within the pebble bed is probably an important factor in its poor ATWS behaviour.

            There is one very important parameter that is different between the 600 MWth reactor and the ones I am talking about – I was envisioning reactors with a power output of 3-200 MWth, with most of my focus on reactors in the 50 – 75 MWth range. I was also envisioning core power densities on the very low end of those proposed for gas cooled reactors. Those two parameters have a rather large influence on peak fuel temperatures after an ATWS event.

            Agree, there will not be an issue for such small reactors. But I want nuclear to compete with large coal plants, not just remote diesel generators, as interesting an early market the latter could be, I really want to get rid of coal ASAP.

            One more thing – fuel damage in well made TRISO particles does not start until temperatures exceed 1600 C for an extended period of time. As long as you keep temperature below that level, you have proven passive safety. Keeping them to a maximum of 975 C does not buy you any additional safety.

            Well, the interesting thing is that this isn’t really true. The retention is actually complicated. The particles leave certain fission products through, such as radiosilver, palladium, etc. to a high degree. It helps to keep temperature lower because all of these diffusion mechanisms are temperature dependent. Cesium and krypton are volatile and will pass through particles that happen to have a thinner coating (these things happen with fluidized bed coating processes). Also, the internal fission gas release and also gas pressure itself scales with temperature. So a lower temperature desing relaxes stresses in the particle coatings. This may allow ultra deep burn (>15% FIMA) which is currently impossible with gas reactor TRISO.

            I actually expect that deep burn gas cooled TRISO will not perform well at 1600 degrees Celsius in a transient. Only mild burn fuel has been tested for retention at this temperature. There is very limited experience with deeply burned TRISO fuel in heatup transients.

            There are various other mechanisms such as kernel migration that are mitigated by lower temperatures. Silicon carbide and graphite do much better at 900 celsius than at 1200 in normal operation. It has to do with swelling and lattice repair mechanisms, all very boring and complicated and I won’t go into it here, but the point is that lower temperature really helps with deep burn fuel performance, especially in a transient.

            Also keep in mind… the passive safety systems are all engineered components, which have real though unlikely failure modes (natural draft chimneys can collapse, shearing can cause flow blockage, passive circulation tubes can still break/block flow, etc etc….). All very beyond design basis but it’s nice to claim walk away safety in severe events. In a nuclear world there will be between 10000 and 100000 reactors so we should plan on things happening from time to time. If you have more margin to failure you can have more of your chimneys be damaged, for example. But a more important point is the behaviour during transients, which is so much better for FLiBe cooled pebbles than helium cooled pebbles.

            1. Well, the interesting thing is that this isn’t really true. The retention is actually complicated. The particles leave certain fission products through, such as radiosilver, palladium, etc. to a high degree. It helps to keep temperature lower because all of these diffusion mechanisms are temperature dependent. Cesium and krypton are volatile and will pass through particles that happen to have a thinner coating (these things happen with fluidized bed coating processes). Also, the internal fission gas release and also gas pressure itself scales with temperature. So a lower temperature desing relaxes stresses in the particle coatings. This may allow ultra deep burn (>15% FIMA) which is currently impossible with gas reactor TRISO.

              @Cyril R

              Are you aware of the success of the current fuel development program for the NGNP program? Here is a link to a 2010 presentation http://www.nrc.gov/public-involve/conference-symposia/ric/past/2010/slides/th37pettidpv.pdf

              Since that presentation was developed, the program has successfully tested fuel with a burnup of 19% FIMA. I know the leaders of that program pretty well; they are excited by the results they are getting. I also happen to work for the people who developed the manufacturing processes. Again – pretty exciting stuff.

              One of the reasons I put my former company to sleep and then completely folded it up was that the DOE program timeline way much longer than our “runway” of funding. We found out in 2007 that the NGNP fuel development program was going to make great strides – by 2021. That seemed like a very long time to wait, but it is now not so far away and it is still on a success path.

              With regard to my aim at the diesel market as an early adopter, I guess we just have different views of how to successfully introduce new technology to the market. I like the model that has been proven to be successful for products as diverse as automobiles, refrigerators, computers, mobil phones and big screen TVs. Find customers who really need the product you can produce when the technology is young and who can pay the prices required to make a profit at that stage of development. Use the profits to more fully develop and deploy the product, seeking ever larger markets. Rinse, spin and repeat.

          2. Are you aware of the success of the current fuel development program for the NGNP program? Here is a link to a 2010 presentation http://www.nrc.gov/public-involve/conference-symposia/ric/past/2010/slides/th37pettidpv.pdf

            Since that presentation was developed, the program has successfully tested fuel with a burnup of 19% FIMA. I know the leaders of that program pretty well; they are excited by the results they are getting. I also happen to work for the people who developed the manufacturing processes. Again – pretty exciting stuff.

            The 19% FIMA burnup was for T <1150 Celsius. There's a big difference between that and a transient temperature of 1700-2100 Celsius. The real question is how this 19% burned fuel behaves in such transients. I highly doubt they will survive, from what I've read about fission gas pressure/release, combined with deteriorating mechanical (and even chemical) properties of SiC and PyC at these temperatures….

            Also, bear in mind these results only discuss fission gas retention, measured as krypton release. They do not talk about the specific fission products that diffuse through intact fuel particles. The really interesting things are not gross fuel failure but diffusion of radiosilver, radiopalladium, radiocesium, CO pressure and kernel migration… just to name a few.

            For a specific excellent read, see this document:

            http://www.inl.gov/technicalpublications/documents/3169885.pdf

            Note that all of these are highly temperature dependent. So it helps to keep temperature down in normal operation if you're going for deep burn. In terms of internal pressure it helps to keep temperature down in transients.

            With regard to my aim at the diesel market as an early adopter, I guess we just have different views of how to successfully introduce new technology to the market. I like the model that has been proven to be successful for products as diverse as automobiles, refrigerators, computers, mobil phones and big screen TVs. Find customers who really need the product you can produce when the technology is young and who can pay the prices required to make a profit at that stage of development. Use the profits to more fully develop and deploy the product, seeking ever larger markets. Rinse, spin and repeat.

            Actually we agree here. Your market introduction plan is a very good one. Where we disagree is what the next step will be. Once you’ve got a lot of little nitrogen cooled reactors that compete with 50 cents per kWh diesel fuel, that doesn’t mean you have an easy inroad to reduce the cost to 5 cents per kWh in order to compete with coal. Yours is an interesting plan, but risks becoming a niche high cost market player, with no easy path towards competing with coal on large grids. In my book that is a dead end. But perhaps I’m pessimistic.

  19. Ug. Something went wrong with the link to the radio story I tried to include. I meant to put it under the second paragraph, with the title, “Are Mini-Reactors The Future Of Nuclear Power?”. Not as a link with last paragraph as the title.

    1. For comparison, the fluoride salt cooled TRISO pebble bed reactor PB-AHTR gets a peak fuel temperature of 975 degrees Celsius for the ATWS event.

      http://pb-ahtr.nuc.berkeley.edu/papers/07-001_PB-AHTR_LOFC_Griveau.pdf

      The difference with PBMR is amazing. 975 degrees Celsius means absolutely zero fuel failure. The 850 C coolant outlet temperature is also a modest 150 degrees C higher than normal outlet, which would not fail the primary loop. In the ATWS, where the PBMR goes Fukushima (in terms of core damage that is – at F-D the control rods worked), the PB-AHTR doesn’t flinch.

      If the PBMR goes south down the ATWS path, you’re screwed. You can’t inject water to cool things down because the reactor would go Chernobyl. Even with borated water the reactor would explode from H2O-graphite reaction which is explosive at elevated temperatures… all you can do is try to add negative reactivity somehow and prey.

  20. Here’s a better reference for C-14 production from nitrogen in gas cooled reactors:

    http://www.osti.gov/bridge/servlets/purl/7114972-2rlK4h/7114972.pdf

    30 ppm nitrogen in the fuel blocks graphite makes 164 Curies/GWe-year.

    164/0.00003 = 5.47 million Curies/GWe-year for a nitrogen cooled reactor.

    The volume of fuel element graphite would be within a factor of 2 of the volume of coolant. Still, 2.5- 5 million Curie seems a very high estimate. They mention ppm not wt% so it should be accurate.

  21. Rod,

    Not too sure what turndown time are like on nukes, but Ameren uses Callaway’s excess electricity to operate the Taum Sauk pumped storage station.

Comments are closed.

Recent Comments from our Readers

  1. Avatar
  2. Avatar
  3. Avatar
  4. Avatar
  5. Avatar

Similar Posts