Can Gas Turbines Using Nuclear Fuel Change The Energy Game?
Disclosure: Rod Adams founded the now defunct Adams Atomic Engines, Inc. His company developed nuclear gas turbines from 1993-2010.
It is a modern day truism that natural gas power plants are cheap while nuclear power plants are expensive. Natural gas plants can also be built in a variety of sizes by a larger number of suppliers who have a pretty good record of completing projects in a relatively short and predictable amount of time.
Another truth is that nuclear fuel can be very cheap on a per unit heat basis, even when compared to what is often called “cheap natural gas.” For the past three decades, the total cost of nuclear fuel has been consistently close to 60 cents per MMBTU; the price of natural gas has been somewhere between $1.80 and $4.50 for the past ten years.
When natural gas prices are near their low point, efficient, affordable, and reliably constructed natural gas combined cycle plants give that fuel the ability to dominate electricity production and sales. That is even more true in a market where gaseous waste product disposal is generously provided for free.
That’s our current situation. Nuclear plant construction costs are out of sight and schedules cannot be predicted. Fuel costs remain consistently low and emissions are virtually non existent, but gas is dominating.
It’s time to change the game by adapting the well-proven, flexible and reliable combined cycle to be able to use nuclear fuel. That will match the best available heat conversion system with the superior fuel that is more abundant, needs no pipelines, produces manageable quantities of solid waste, and can operate inside sealed buildings.
If you can’t beat them, copy them.
Simple Cycle Gas Turbines (Brayton Cycle Machines)
Modern natural gas plants include Brayton Cycle gas turbines. The cheapest ones per kilowatt of generating capacity are classified as simple cycle machines. They have no additional components designed to improve plant thermal efficiency. They consist of a compressor, a heat source, and a turbine with some inlet air ducting and filters, a fuel supply system and a stack for exhausting waste gases.
They are so cheap that they can be economically viable even if run a few hundred hours per year to meet peaks in demand. Their fuel costs are high due to low thermal efficiency. They are so simple that they can be remotely controlled and often don’t need on site operating staff. They can spin up from a cold condition to full power in 15-20 minutes.
They can burn either natural gas or distillate fuel; the main requirement is that the fuel needs to be zero or very low ash and it cannot contain sulfur or other contaminants that can attack turbine blades.
When run on natural gas, these peaking plants are dependent on “just in time” fuel delivery systems. In some regions, especially New England, natural gas delivery can be interrupted when the turbines are needed the most. That problem can be solved by installing dual fuel capability and a distillate fuel oil storage tank. Duel fuel machines with on-site storage can then burn distillate fuel oil in an emergency situation.
Unfortunately, distillate fuel often costs 4-10 times as much as natural gas on a heat content basis. This choice is made affordable by running the turbines on premium fuel only when the need is high enough to spike electricity prices.
Brayton cycle machines made their way into the electrical power business because they filled an important customer need for reliable sources of power to meet demand peaks. Since they were not expected to operate steadily, their construction costs needed to be low enough so that they could sit idle without driving the accountants crazy.
Advantages Of Brayton Cycle Machines Over Steam Turbines
These machines, derived mainly from experience in producing powerful, compact, lightweight, reliable, flexible, and cost effective engines for aircraft and ship propulsion, have numerous advantages over the venerable steam turbine power plants that have been in use on ships and in electrical power production since the First World War.
Unlike steam plants, gas turbines depend on a working fluid that doesn’t change phase over the full range of pressures and temperatures found in the operating cycle. It is a gas at the beginning, middle and end of the cycle. This attribute shrinks and simplifies the system by eliminating the need for valves, piping, and heat exchangers to separate and recover condensation. Water droplets can destroy rapidly spinning turbine blades and must be eliminated before using the high pressure, high temperature water vapor to turn a turbine.
The gases used as the working fluid also undergo a much reduced change in volume per unit mass over the range from the highest system pressure and temperature to the lowest. This cycle attribute leads to compact turbines where the size difference between stages is far less than a comparable steam turbine. Depending on the gas used, the number of stages in the turbine may also be significantly reduced compared to a steam turbine.
Gas turbine power plants do not require large tanks of tightly controlled pure water to provide system make-up to compensate for inevitable leakage and purposeful use of working fluid to clean the surfaces of heat exchanger tubes.
Simple cycle gas turbines also save space and weight by using fuels that produce purely gaseous waste products with little, if any ash particles. By using only zero or very low ash content fuels, simple Brayton cycles eliminate expensive, space-consuming heat exchangers (boilers or steam generators) that steam plants use to transfer the heat of combustion into boiling water. Instead, they spin turbines by directly expanding heated, compressed gas through the turbine’s acceleration nozzles and blade stages.
As a result of the above features, complete Brayton cycle gas turbine-based power plants are often less that 1/4th the size and mass of a comparably capable steam turbine power plant. They thus require less construction material and fewer labor hours to assemble. They are simpler to harden against seismic stresses. They need fewer, smaller buildings for the plant and also for supporting the smaller staff of people compared to steam plants.
Turning Peakers Into Longer Running Facilities
Gradually, Brayton cycle peaking plants proved their value and demonstrated their easy reliability. Customers showed they were interested in spending more for improved performance, especially in terms of specific fuel consumption.
Engineers suggested that the most dramatic efficiency improvement could come from combining Brayton cycle machines with Rankine cycles in a cascading series of heat conversions. The exhaust gases depart Brayton cycle turbines at a high enough temperature to function as the heat source for water boilers known as heat recovery steam generators (HRSG).
Maximum thermal efficiency in a simple Brayton cycle tops out at less than 40%; in a combined cycle, it is possible to achieve efficiency closer to 60%.
Of course, combining Brayton cycles with Rankine cycles reintroduces some of the cost, size and complexity of pure Rankine steam systems, but a 1000 MWe combined cycle plant includes a steam plant that is less than 1/3 the size of a 1000 MWe steam plant since about 2/3 of its output comes from the Brayton cycle turbines. It is also a more flexible power system that usually includes at least three turbines, each of which may be able to operate without the other turbines being in operation. (Many HRSG’s have the ability to be directly fed from an auxiliary boiler instead of from waste gases from the Brayton Cycle machines.)
Why Does Steam Still Dominate In Nuclear Projects?
One might logically ask if Brayton cycle machines are so obviously cheaper and simpler than Rankine cycle steam turbines, why are any steam turbines being built? Why haven’t at least a few nuclear projects using Brayton Cycle machines been commercially deployed?
Even though a few of the advanced systems being designed have Brayton cycle heat conversion systems, they are still in the minority. Even those that do are not as simple as they could be, despite the fact that capital cost reduction should be at the top of a design criteria list for any plant designer who wants to succeed in the market.
It is unfortunate, but many people trained as nuclear engineers have only a dim understanding of the massive structures, systems and components required outside of the highly refined reactor cores that are their main area of responsibility. They are experts at modeling the interactions of neutrons, a wide range of isotopes, varying coolant properties and widely differing support structures, but they don’t appear to spend much time thinking about the challenges of building, operating and maintaining enormous steam plants, pure water systems, cooling water systems, valves, pumps and some of the world’s largest saturated steam turbines.
They spend even less time thinking about the costs and schedules associated with building the massive foundations, buildings and supporting systems that those large steam plants require.
Addressing the costs of systems and structures related to converting nuclear fission heat into useful power is more important that most nuclear engineers realize. Though it’s difficult to find published studies with this information, a statement often heard in gatherings of nuclear project experts is that ~ 80% of the construction budget for fission heated steam plants is consumed outside of the “nuclear island.”
Heat Conversion Requirements Drive Nuclear Technology Selection
If a plant designer starts with the goal of adapting gas turbines to effectively operate with fission heat, it becomes almost immediately obvious that conventional light water reactor plants and their associated fuels cannot do the job.
The maximum temperatures available from such a system can’t drive a Brayton cycle machine with anything close to the desired efficiency.
Sodium cooled reactors come a bit closer, but the boiling point of sodium at atmospheric pressure is still too low for reasonable Brayton cycle efficiency. Many engineers have good reason to be wary of sodium coolant at low pressure; they would have even more cause for concern if asked to consider high pressure sodium as a primary coolant.
Even molten salt reactors, which were initially developed with the idea of using them as the heat source for Brayton cycle jet engines or turbo-props, have heat exchanger material limitations that make it unlikely they can produce gas turbine inlet temperatures above about 800 ℃. A complex Brayton cycle that includes large, expensive components like recuperators and intercoolers can effectively use that temperature to produce reasonable cycle efficiency, but there isn’t much headroom for improvements without a material breakthrough.
The added complexity of the system negates some of the previously described Brayton cycle advantages.
High temperature gas cooled reactors have been under almost continuous development and refinement since the 1940s. They have demonstrated the capability to deliver gas at temperatures at temperatures as high as 1200 ℃ using technologies available in the mid 1960s. Though most high temperature reactor designs today aim for a more modest introductory temperature of 750-800 ℃, there is headroom for future improvements.
One of the keys that may unlock high temperature gas reactors and free them for wide use in an almost unlimited number of applications is the Triso coated fuel particle. This innovation, developed almost 50 years ago, coats a tiny particle of fission fuel in multiple layers of material that combine to seal fission products inside the particle.
As long as the coatings are properly applied and remain intact, radioactive materials remain securely contained. Even if there are minor releases from improperly coated or damaged particles, the public is protected by several additional barriers. Through a lengthy, well-controlled, well-designed and consistently managed development program, the U.S. DOE and its partners have developed credible, repeatable processes that result in reliable fuel capable of long core exposures.
As long as high temperature reactors have “Triso Inside™” and are designed to ensure that their fuel temperatures remain within a broad band, currently reaching as high as 1800 ℃, they should be able obtain construction and operating permission on the basis of providing adequate protection.
Fuels with Triso Inside will be more expensive than conventional nuclear fuels when they are first introduced, but there is a well trodden path towards mass production cost improvements. Triso fuel suppliers should act like traditional fuel suppliers to encourage widespread use of their product, including creating identity advertising, assisting with licensing, training system designers, and potentially offering financing to developers who are creating innovative ways to use their product.
Perhaps some well capitalized fuel suppliers with exceptional project management and marketing skills could leverage their chemical engineering expertise into a lucrative new line of work.
Cooled By Nitro™
The other decision that can throw the door wide open on developing nuclear heated Brayton cycles for almost any conceivable energy application is selecting a working fluid that can take advantage of existing Brayton cycle machinery.
A small niche of nuclear plant designers has been interested in using hot gases passed directly though high temperature reactors to spin gas turbine machines since the earliest days of nuclear power development. Unfortunately, nearly all of them assume that helium is the only gas that can be used to directly cool a nuclear reactor.
That assumption – which is false – leads in two possible direction. One is the path that General Atomics and PBMR took on paper, which was to conceive of the use of a helium turbo machine. That path turned into an effective dead end in the real world; the challenges associated with developing helium turbo machinery are far greater than imagined. PBMR ran out of money before approaching a useable machine; GA took the safer path of not even trying to exit development studies on their own dime.
The other alternative is to use helium cooling for the reactor, but to add a gas to gas heat exchanger to move the heat into air or nitrogen so that conventional turbo machinery designed for combustion turbines can be used. This path adds a large, costly component and reduces system efficiency because of the inevitable drop in turbine inlet temperature caused by the heat transfer process. Adding the heat exchanger is particularly problematic when looking into the future; there are well identified paths for increasing the temperature capability of Triso particle-based fuels, but material breakthroughs will be required to enable cost effective heat exchangers at temperatures higher than 800-850 ℃.
There have been three nuclear heated Brayton cycle machines operated in the US. Two – HTRE-1 and HTRE-2 – used atmospheric air and put the turbine gases directly into the desert air in the location where they were tested. The other was the ML-1, which was designed to produce 300 KWe and to be mounted on a trailer that could be pulled to a remote communications site.
The ML-1 was “cooled by Nitro™” which allowed it to use conventional turbo machinery. After all, nitrogen gas makes up 80% of atmospheric air and has almost identical thermodynamic characteristics.
Unlike helium, nitrogen is available in almost unlimited quantities in an unlimited number of geographic locations. It also works well in the same machines that have been developed and refined to use atmospheric air and combustion products. Using nitrogen as the reactor coolant and the working fluid for the turbo machines eliminates the need for a separate heat exchanger.
High temperature nitrogen cooled reactors will be large, low power density components, but the “reactor” performs the functions of heating the turbine gas, storing the fuel, and providing effective decay heat mitigation. When comparing sizes to other potential power sources, the important metric is the size, cost and schedule for complete systems instead of focusing on certain components.
During early stages of development and deployment, system designers should keep it simple and not worry much about thermal efficiency. They can build roadmaps for future improvements, however, that build on the advances that have already been proven to work well in combustion Brayton cycles, including developing combined cycles, co-generation for process heat applications, co-generation for district heating, and inter cooling and regeneration.
One improvement path that can be carefully developed is uniquely available to closed Brayton cycle systems. Unlike air breathing systems, it is possible to increase the pressure in closed Brayton cycles so that the same volumetric gas flow produces greater amounts of heat transfer and improved power output. Designers should be cautioned to start slowly in this area, there are safety and simplicity benefits associated with using system pressures that are the same as those used in combustion gas turbines.
Though all of this may seem logical and almost obvious when laid out in this manner, there are some valid reasons why this path has not yet been taken. Diving into those reasons is beyond the scope of this document; suffice it to say that most of the barriers have been overcome except the one at the starting line. There are impressive returns available to those that recognize that it’s time to open the gate and let the contestants begin their race.
Good introductory article for engineering students. I’ll send this nice tutorial to an engineering prof friend who teaches power conversion.
GE claims up to 44.1% (LHV) efficiency for the LMS100. I recall an efficiency curve which peaked out at 46%, but that appears to have been retracted.
The closed-cycle nitrogen turbine doesn’t get rid of the heat-exchanger problem; it just moves it to the cold end of the heat engine, where heat must be rejected to the environment.
As for compact systems, supercritical CO2 appears to have it all over nitrogen in that department. NetPower (probably under water right now in La Porte, TX) is working on an Allam cycle combustion plant which uses supercritical CO2 for the bulk of the working fluid; there’s a bit of water formed in combustion which is condensed and removed at the end of the cooler. A CO2 system designed to run at 1000 C could easily be re-engineered to operate at 700 C instead, with some loss of efficiency (but with fuel costing less than $1/mmBTU, that doesn’t matter much). It appears that the turbomachinery for CO2 will be available OTS fairly soon.
I’m particularly struck by the cold-end temperatures of those cycles. What I recall is that the cooler receives the working fluid at around 170 C. This would lend itself to a very compact natural-draft cooling tower, though it would probably need to be “wet” rather than “dry” under summer conditions in many climates.
GE has a different definition for simple cycle than I was taught in thermodynamics. To wit, GE describes their machine as “highest simple cycle efficiency gas turbine in the world. Its intercooled gas turbine system”
Intercoolers are not generally considered to be part of a simple cycle machine.
I agree that a closed nitrogen system does not eliminate the heat exchanger problem. It limits it to a low pressure, reasonable temperature device made from commonly available, moderately priced industrial materials. If I had my “druthers”, I’d evaluate the alternative of an open cycle machine that is more closely related to HTRE-1&2 than ML-1
It’s a simple cycle because it’s not a combined cycle.
The multi-spool aeroderivative design lends itself to intercooling because there’s a natural separation between the high and low pressure compressors and they already run at different speeds. That’s probably why GE did it.
I was taught that there were multiple classifications of gas turbine cycles including:
Combined with reheat
co-generation (which uses a heat recovery steam generator but not a steam turbine)
and several additional.
Sure, there is a natural place in which to put an intercooler, but that is another heat exchanger with an auxiliary water system that probably includes pumps, valves, pipes, pressure sensors, temperature sensors, and some provision to protect against effects of leaks. There may be additional subsystems to support the pumps with lubrication and cooling water and to provide remote or local controls for the valves and pumps. It means either more wires to route sensor information to the central display system.
IOW, it’s not a “simple” cycle with a compressor, heat exchanger, turbine and heat sink.
CO2 cycles are very interesting and promising. Though for TRISO fuel they do have an issue – CO2 is not compatible with graphite at elevated temperatures. There’s some carbothermic reduction going on – CO2 + C = 2CO type reactions.
CO2 should be fine with silicon carbide, however. It’s been proposed as a coolant for SiC clad fuel fast reactors.
Pfft. You put an SiC outer skin on your pebbles, assuming you want to run them that hot. ΔG for the C+CO2->2CO reaction doesn’t go negative until about 700°C. If you keep things in the 550°C range proposed by Dostal a graphite outer shell should be fine.
“ΔG for the C+CO2->2CO reaction doesn’t go negative until about 700°C. If you keep things in the 550°C range proposed by Dostal a graphite outer shell should be fine.”
Yes, at those low temperatures there’s much less of an issue. However, we would have to know that this is true also for supercritical CO2 in a high neutron and gamma flux level found inside the core of a nuclear reactor. The experience with supercritical water in reactors has been rather poor, for example, compared to supercritical water in coal plants. Unless you can make the core internals, control rods, guide tubes, etc. all out of SiC, this would also apply to any metallic materials in this environment.
The other thing to consider is this is a nuclear reactor, which can have transients that heat up the core and coolant to much higher than normal temperatures. What does the emergency cooling system/decay heat removal system for this reactor look like?
“You put an SiC outer skin on your pebbles, assuming you want to run them that hot.”
Yes this is sort of what I was hinting at. It would probably be a good idea anyway; graphite pebbles create friction from rubbing and moving through the core, creating dusting and possibly breakage. SiC outer layer would help with it’s much higher toughness and hardness.
Thanks a lot for this well thought-out and excellent narrative on your “nitro” reactors (man that sounds good), as well as many other good atomic insights.
Tech question – have you calculated the buildup of C-14 over the years? For closed cycle and 60 years of operation this should be pretty large in terms of Bq inventory. Since the Brayton requires periodic opening for maintenance you’d have to consider this term.
Off topic but the South Texas Project points out that their nuclear power plant kept generating despite Hurricane Harvey while of course the wind turbines in the vicinity had to shut down.
like the Holos reactor
Could you elaborate on the type of residual heat removal/emergency cooling systems you had in mind for this “NiTriso” reactor system?
My initial reaction to this is my standard view that excessive regulations and QA standards, as opposed to technology (such as the coolant or thermodynamic cycle used) are the main reasons for nuclear’s high costs. And I will devote a couple sentences to that here. If 80% of the cost is outside the nuclear island, one needs to ask whether it is really necessary for NQA-1 to apply outside the nuclear island. Perhaps a reactor design that can go indefinitely w/o active cooling would strengthen arguments to change that.
OK, that said, I think there is a lot truth to what Rod is saying here. I understand the point that, since nuclear fuel is so cheap, it makes sense to sacrifice some thermodynamic efficiency to reduce the cost and complexity of the system.
The use of gaseous coolant has other significant advantages. Let me give an example. The water in the coolant loops of LWRs needs to be demineralized (to prevent corrision and other undesirable chemical issues). The problem is that, in nuclear’s case, the minerals in the water are highly radioactive. Thus, high radiation levels exist around the water pipes/tanks, and even higher doses occur around the concentrated minerals extracted from the water (e.g., resin beads, etc..). One of the largest buildings in the AP-1000 plant (the Aux building) is largely devoted to demineralization. Most of that building consists of thick steel reinforced concrete walls, in order to shield the radiation coming from the tanks and pipes in each room. And, of course, max nuclear grade QA is applied to all of this. Indeed, I believe some of the components that CB&I had trouble delivering (because they couldn’t deal with the uniquely strict nuclear QA requirements) were Aux building structure.
The use of a gaseous coolant could largely eliminate (or at least greatly reduce) such issues. No need for demineralization, the associated concentration of highly radioactive material, and the associated need for heavy shielding.
A couple follow on questions.
In the past I’ve asked whether, for an HTGR, the (TRISO) fuel kernels could be the only nuclear-grade components in the entire plant. The reason being that the size, geometry and materials of the core ensures that the kernel temperature limits cannot be exceeded, via passive thermal conduction alone. Even any (plausible) reconfiguration of the core geometry would not result in a release.
People on this site responded by pointing out that moisture ingress could be a significant problem. OK, so perhaps just the reactor vessel and primary coolant loop piping would be nuclear grade. Or will regulators pull on the string and imagine that (implausible) failures of other components could cause a breach in the primary coolant loop, causing moisture ingress. How unfortunate…
If there is no steam cycle plant nearby, couldn’t we argue that there is no potential source of a significant amount of water/moisture. Or is the moisture level in the (normal) air enough to be a problem? Also, has anyone estimated the release that could occur as a result of such moisture ingress?
A related question. If we wanted to add a steam cycle, to increase thermo efficiency, could we locate the steam plant some distance from the reactor, and pipe the hot gas over? How much loss would that entail? Also, not to beat a dead horse, but if the steam cycle plant were far away from the reactors, couldn’t that whole plant be built to normal industrial standards?
Technically, a TRISO topology is that inner SiC layer is the fuel clad, and outer SiC is the containment, so yes you can claim that the fuel kernel is the only nuclear component, given a demonstrable inherent core safety and passive cooling. Likely though there will be some heat sink structure and that will have to be safety grade and seismically qualified. Otherwise it will depend on fuel alone basically. You have to prove very low fuel failure rates and defects etc.
In terms of steam, HP steam can be transported a long distance. LP steam is much worse. So if you use a modern HP superheated or supercritical turbine it will work well.
No Cyril R!! There is only ONE layer of SiC on a TRISO particle. There are 2 pyrolitic carbon layers, inner and outer layers, called IPyC and OPyC that are used to protect the single SiC layer and hold it in compression during irradiation.
great article, whilst a simple and efficient land based alternative to fossil fuels would be ideal, more conventional reactors like Starcore is going down that route as are other SMR’s a lighter gas turbine alternative can do nothing but good.
I wholeheartedly agree with you the real pollution and environmental damage is shipping, a smaller than 100MW station installed in a bulk carrier would reduce pollution and greenhouse gases by a enormous amount whilst allowing increased speed and shorter journeys.
“It is unfortunate, but many people trained as nuclear engineers have only a dim understanding of the massive structures, systems and components required outside of the highly refined reactor cores that are their main area of responsibility.”
I am fan of Khan Academy, high quality, free, easily accessible, well organized simple tutorials, that provide a basic overview on a wide range of academic subjects. Perhaps there is need for an “Adams Academy” – high quality educational tutorials that explain all the intricate components that go into converting heat energy into electrical energy.
What about using one or more Echogen’s EPS100 (i.e., specs – goo.gl/qMREbV ) sized to match a NuScale individual reactor? Waste Heat Input kW 33,300 MMBTU/h 114, Waste Heat Supply Temperature °C532, °F990, Heat Flow Rate kg/s 68 lb/s150? It uses the super critical CO2 cycle?
another activation related question: has anyone calculated the dose rate in the turbine from N-16? Some N-16 would be produced from capture in the nitrogen-15. Small cross section I know but N-16 is a very high gamma emitter. It would be useful to know the dose and shielding requirements for the Brayton and other equipment.
N-16 has a short half life. N-16 is also produced in BWR’s due to radiolysis of water, then the O-16 turning into N-16. Some of the piping from the reactor to the turbine needs to be shielded, yet by the time the flow exits the turbine, the N-16 is back to O-16. A side item is that a gas recombiner is necessary to react the oxygen and hydrogen, making water. This is exothermic.
Here’s a basic maybe dumb question, but I’ve been told that the only stupid question is the one that isn’t asked and have come to believe it.
The combustion turbine I used to work with had “baskets.” These baskets were at the front end of the turbine after the compressor. Fuel and compressed air entered the basket and were ignited. Great expansion of the gas took place as the cool compressed air expanded as it became exhaust gas. The expansion caused pressure to move the turbine blades.
It seems to me that to be able to replicate this action, you would need a fluid heated by the reactor to heat this compressed air. This is the part I have trouble with. The combustion baskets were quite small. They were made of some type of high temperature alloy. (hastelloy?) Could a heat exchanger be made that small? How would you retrofit the combustion section of the turbine to heat the compressed air. Could you surround it with a large heat exchanger that wrapped that section of the turbine?
Most reactors are water cooled. Since water will be available, this would gain you a bit of efficiency in being able to use evaporative cooling at the air intake to the unit.
I like this idea. There is a large proliferation of gas peaking units in the country. If they could be replaced by small nuclear units, natural gas could be used for better purposes. Since they would become commonplace, nuclear fear would dissipate.
Gas turbine burner cans are usually Inconel, IIUC.
Burning fuel is just a way to add heat to the gas. Anything will do if it gets to a high enough temperature. Someone had the idea of using molten silicon (freezing point 1414 C) as thermal storage for a gas turbine. That would actually have greater thermal efficiency than when burning fuel, because the ratio of specific heats of air is greater than that of combustion products and expansion will cool the gas more.
Hey, there’s another use for excess generation (from unreliables or otherwise): use it to heat up thermal storage at combined-cycle power plants. When you need the CC plant you’ve got a supply of energy available which doesn’t require any fuel to be burned.
Molten silicon is pretty cool, weird high heat capacity going on.
What materials would be compatible with molten silicon? Don’t you need heat exchanger tubes or plates, and some storage vessel? Vessel could be internally insulated with some refractory but a heat exchanger seems tricky. I imagine molten silicon being quite reactive with air, especially very hot air. Can’t imagine semiconductor industry using molten silicon heat exchangers.
Silicon carbide tubes?
They obviously exist, or there would be no way to crystallize it. I suspect SiC would corrode by exchanging carbon with the melt (though maybe not). I’d look to things like alumina which can run white-hot.
Obviously. It turns out that NaK eutectic has a boiling point just above the silicon melting point, while pure potassium boils well below it. It looks like you could make a heat pipe with an NaK mixture with its BP tuned to be somewhat below the silicon melting point, say 1350 C. This would let you separate your heat storage and its network of heat-extraction things and the high pressure air supply of the GT.
Turning it into silicon dioxide is counterproductive; one or more walls between the melt and the air supply are definitely in order. Heat pipes have very low temperature drop due to the phase change, so that appears to be a promising avenue of research.
I crunched some numbers and found that you’d store about 1.1 kWh(th) per liter of silicon (solid; the liquid density is higher but you have to size the vessel for the solid). This comes to 660 kWh(e)/m³ to a 60%-efficient gas turbine. You’d have some volume occupied by heat-pipe tubes and whatnot, but the idea of a 10-meter cube of heat-storage stuff that can store on the order of 600 MWh(e) sounds like the answer to a prayer. The storage capacity of Ludington is only about 40 GWh; you could do that with 66,000 cubic meters of silicon storage, or 16.5 unit cubes of 20 meters on a side. This is positively tiny; most petroleum tank farms are far bigger.
Alibaba lists metallurgical silicon prices as low as $300 per ton (metric I assume). At $500, the cost of the storage medium comes to about $1.90/kWh. This is a spectacular price. The technical challenges are considerable, but that kind of payoff ought to be attracting heaps of interest.
Hmm ok, those are impressive figures!
I get the following:
Heat of fusion: 1926 kJ/kg
Liquid density @ mp: 2.57
= 4950 kJ/l
= 1.375 kWh/l
Plus some superheating/subcooling you’d be looking @ 1.4+ kWh/l
I guess you can line the storage vessel with alumina or some such ceramic, but making HX tubes out of it is going to be tricky. You’d want something that has good leak tightness, good tensile strength, good creep resistance. Probably triplex SiC tubes would work. Complex SiC parts can be made using liquid silicon capillary intrusion (ie reaction bonding). You’d have to make sure the stoichometry is right for liquid silicon service, for sure, but otherwise I don’t see a real problem.
Too bad the efficiency is only 55% or so (considering losses, compressor power, and a 60% efficiency CCGT). Be like almost throwing away half the energy you put into it. Ouch.
agree 66000 m3 is tiny, a typical large capacity oil farm storage tank that. They’re building >>100000 m3 these days…
After doing a bit of digging, it turns out that alkali metals attack alumina at high temperatures so it’s obviously not a good choice for compatibility with NaK heat pipes. Also, inconel’s temperature range tops out well below1400 C so inconel heat exchangers and even pipes are non-starters. It looks like the choices are tungsten and ceramics no matter what we’re using as the heat-transfer fluid.
Magnesium’s BP is 1090 C, so it might work as a heat-pipe medium but it would be under considerable pressure and ignite instantly if it leaked. It might also attack alumina (but maybe not SiC). Antimony’s boiling point is 1440 C. That would put it under slight vacuum, which looks desirable. Tungsten pipes carrying antimony liquid/gas between alumina or silicon carbide heat exchangers?
Sounds like a fascinating lab exercise that would be a FOAK engineering nightmare.
The combustion chambers would be removed altogether. The gas just comes in pre-heated by an external source – a nuclear reactor core. It’s direct cycle.
OK- So the gas flows through the turbines compressor and is routed externally to be heated directly by the reactor and returns to the gas turbine. It then is exhausted as hot air. OR Is it a closed cycle? A closed cycle would require us to cool the gas. We are also running high pressure gas through the reactor.
I like the idea of a heat exchanger. In the event of some sort of reactor breach, the cooling medium is kept enclosed. The reactor acts as a heat source continually cooled by this circulating fluid. The heat exchanger uses the fluid to heat the compressed air which can be vented to the atmosphere.
Yes, it would have to be a closed cycle. Directly using the air would lead to both significant oxidation loss of the graphite pebbles, as well as leading to far too much C-14, N-16, Ar-41 etc. emissions to the environment. Not to mention not having containment. Theoretically I suppose you could live with some graphite oxidation loss and it would be fine for a not densely populated area, but it seems like a hard sell to have this thing directly air cooled. It would be a closed system. But, as Rod has argued, having a low temperature heat exchanger is not such a big deal, especially with benign and well understood gasses such as nitrogen, helium, or CO2. Low temperature heat exchangers are common and pretty cheap. You’ve got a couple of many thousands of Watts in your house almost certainly.
With advances in compact heat exchangers it is pretty impressive what can be done these days. The heat exchanger ends up being smaller than the reactor! And reactors, of any type really, are high to very high power density already…
Would it make any sense to use nak as a heat transfer medium between a direct closed nuclear heated nitrogen Brayton cycle and the ultimate heat sink to allow for higher heat rejection temperatures at low pressure and therefore a smaller (cheaper?) cooling tower?
If you’re using NaK as the circulating coolant you have almost no radioactivity going to the coolant-gas heat exchanger (probably less than in a BWR) and you can probably use an open-cycle gas turbine and simply dump the warm exhaust to the environment.
There are probably molten salts that have a higher boiling point, though. Those would be preferable as low-pressure heat transfer fluids, and if they were at atmospheric pressure any leaks from an open-cycle gas turbine would be inward, not outward.
I’m still thinking that sCO2 is preferable if you don’t have graphite in its loop.
There’s been some work on sCO2 direct cycle fast reactors in the past, it might interest you if you haven’t read it already:
Technically you could use NaK as intermediate cooling loop, and it would help to isolate any activity in the primary loop from the environment. Question is, is it worth the tradeoff in added cost, complexity, fire protection equipment etc? NaK is pretty hairy stuff; it burns in air and explodes in water. Since the heat sink will have water and/or air on the back end, this seems like tricky business. You can probably make the case that there’s no significant risk to the reactor: N2 does not react with NaK and a high temperature reactor with inherent core and passive decay heat removal safety isn’t going to be much bothered by fires in downstream loops. But is it worth it? Probably not.
More interesting, I think, is the application of leak-free heat exchangers like PCHEs. Then you can have any sort of heat sink you want, water or air.
So which alternative would better fit the standard designs of today’s gas turbines: Would it be the closed cycle or the heat exchanger to heat air? I am still thinking that using as much commercially available off the shelf stuff will allow fewer headaches. The heat exchanger to heat air would still allow that.
How would load be controlled? Would the output of the reactor be varied or would you simply dump some heat elsewhere and route less heated fluid to the gas turbine? This would allow for a startup to the turbine allowing proper metal expansion.
Existing gas turbines are remote controlled and so simple they can be made up of skid mounted equipment. Complex systems will not be competitive. Ideas like Sodium cooling may be a tough sell.
Eino, it should be quite feasible to have a gas cooled reactor provide hot air via a printed circuit heat exchanger. This gives you a lot more isolation from any activated products in the reactor coolant, and in case of an accident with some fuel failure it provides another barrier. I’m not sure if Rod is aware of the C-14, N-16, etc. activation products that would give issues for turbo-machinery maintenance.
The air could be at about the same pressure as the closed cooling gas: then there’s nil pressure difference across the exchanger so very low primary stresses. Then all you’d worry about is secondary (thermal) stresses which can be addressed by design (series exchangers, geometry, close pinch-point design etc.)
The heat exchanger would probably add about $~100/kWe installed.
With a closed cycle gas reactor loop you can then have any gas you’d want for the reactor while keeping air for the turbo-machinery. My short list would be helium, neon and CO2 for candidate gasses. Neon is more available than helium, easier to contain, and doesn’t make tritium (unless I’m missing some reaction). CO2 is interesting from many angles, no activation problems, experience with CO2 cooled reactors (UK), good heat transfer, might allow gear-up to a sCO2 type reactor later on. As discussed for a high temp reactor with graphite it does require some considerations for graphite loss via carbothermic reduction but this is pretty simple (SiC coating of the pebbles).
We’re talking about 3 different things::
1. Indirectly-heated open-cycle nuclear gas turbine using air, probably combined-cycle.
2. Directly-heated closed-cycle nuclear gas turbine.
2a. NItrogen working fluid using OTS GT powerplant parts.
2b. CO2 working fluid using soon-to-be-available Allam-cycle machinery.
3. Non-nuclear combined-cycle turbine, powered by stored heat, combustion or any combination.
#3 works hand-in-glove with either #1 or #2, which like to run flat-out. You use the heat-storage units as dump loads during the periods of low demand, then operate the CC plants on carbon-free stored heat while you have it. While your plant is using dump loads it gets paid for providing spinning reserve and regulation.
(2/2) I think #2b may be the killer app, because an Allam-cycle machine is going to be tiny and probably cheap. If 80% of the cost of a modern plant is outside the nuclear island, shrinking a 50 MW turbine to something that will fit in a pickup truck and eliminating all the water-purification systems required for steam has to slash a lot off the bottom-line cost. As for re-engineering, just buy them in 50 MW pieces from NetPower and gang them together with synchronous alternators. Your nuclear plant could use any number of turbines and keep operating if one trips off-line, or several.
Guesstimating: if the nuclear heat supply for a NuScale is 1/5 the total cost of $6000/kW(e), that’s $1200/kw(e) or $360/kw(th). If you can get 45% efficiency on the Allam cycle, that becomes $800/kw(e). If the BOP cost is cut by 2/3 because of elimination of whole areas of equipment, that’s $1600/kW for a total of $2400/kW(e).
Yup. Absolute killer. And that’s before you get into value-added things.
 You might not even need to sync them if you fed them all from a common cooler. If one unit lost its load, it would hog the compression work when it tried to overspeed and limit its speed that way.
 Some of these operate along with electric generation, like sale of hot water for district heating. The cooler inlet temperature of NetPower’s machine is probably close to that of Dostal’s recompression cycle, or about 170 C. With printed-circuit heat exchangers, that heat can be converted to hot water at no less than 150 C for distribution; that’s less than 80 PSI gauge vapor pressure so it wouldn’t be hard to keep it from boiling. Dilute this water down to 80 C at the point of use with recirculated water, and you’ve got all your space heat and DHW taken care of.
Other things would use hot working fluid straight off the reactor, a “dump load” for heat. If the reactor is producing CO2 at 650 C, something like Frank Shu’s supertorrefier operating at 550 C is a natural complement. If you can convert an entire city’s waste stream to clean syngas and stable char using excess heat on nights and weekends….
Rod must be off somewhere; nobody’s rescuing things from moderation purgatory.
“The air could be at about the same pressure as the closed cooling gas: then there’s nil pressure difference across the exchanger so very low primary stresses. Then all you’d worry about is secondary (thermal) stresses which can be addressed by design (series exchangers, geometry, close pinch-point design etc.)”
It’s starting to sound pretty good now. External connections to the turbine mechanism would be a gas out from the compressor and a gas in to drive the turbine. I’m sure some sort of shroud would be needed to ensure the incoming air will drive the turbine in a balanced fashion. Hot air could simply be vented upon startup and when lower loads are required as when accommodating windmills.
Would it be good to have a “bank” of these and one reactor to provide the heat for the entire bank? It may be easier to build smaller units and no full outage would be required for turbine maintenance. Simply shut the one down that you are working on. I don’t suppose long outages would be needed for the Triso fuel reactor since like Canadian reactors, the fuel is constantly being replenished.
If the fuel is cheap enough, there would be no need for the added efficiency of the combined cycle and steam turbine.
I think this may be what makes NetPower a killer in the nuclear industry. The 50 MW unit size is convenient. The real problem I can see is that there’s no nuclear heat source, either available or planned, which could feed it properly. Something like a NuScale reactor unit is about the right thermal capacity to feed 2 (and could be up-rated to feed 3 and maybe 4 with the forced coolant flow), but the recuperator outlet temperature is too high for mild steel reactor pressure vessels to tolerate. Different materials like stainless steel are called for.
Who’s got the forges to make a 2.5 meter ID stainless reactor vesssel?
Key observation. Eliminate capital cost and interest, because the added operating cost is trivial and it reduces market risk.
2.5 m ID ss forgings are quite available, but in limited thickness/weight. Though you could buy smaller sections and electron-beam weld them together. Building up from plate is always an option too.
Jacketed systems are interesting too, basically cool the vessel by incoming compressed gas. Just good engineering design – always try to make the high pressure boundaries closer to room temperature. It’s generally easier to do than gas turbine blade cooling as the vessel isn’t moving.
If the jacket cooling is passive, it can do double-duty as emergency heat removal. Internal insulation can be applied to minimize heat transfer in normal operation.
This paper by Talbert et al provides an innovative concept for a supercritical water cooled reactor. It looks like it is applicable to sCO2 as well – silicon carbide particle fuel, etc. I really like this type of concept, except for the massive thick pressure vessel that seems like a bottleneck to building a lot of these.
I thought that forgings were used because the NRC has a problem with welds. Otherwise we wouldn’t have to go to Japan to buy reactor top heads and things.
I looked at that, and even the recuperator outlet temp was too high for mild steel (which seems to be used because it resists embrittlement). You’d have to use something like straight compressor outlet gas.
That might not be a bad option, actually. You’d lose thermal efficiency mixing it with the inlet gas, but what’s a bit more uranium when it’s so cheap?
I think the weld issues are only for the core beltline region, having a ring forging eliminates the axial welds which is nice for in service inspection work. and maybe the control rod drive region, having no welds makes life easier there.
In terms of jacketed system, you could also have the insulation on the inside and then have passive cooling on the vessel wall. Not too hard to keep it <<200 Celsius like that. Nice benefit is it reduces thermal stress which is usually more limiting than the primary (pressure) induced stress, considering the thickness. That's nice for startup/shutdown cooldown/heating transients, as well as safety.
I’m having a bit of a vision here. Something in the form factor of a NuScale unit, but using full-height fuel elements with the forced coolant flow and doing away with the convection chimney. Stainless or SiC cladding would be required as Zr reacts with CO2 starting at about 500 C; perhaps metal fuel with a sodium bond a la EBR II. 660 MW(t), 300 MW(e), 6 NetPower turboalternator units per reactor. The only thing that needs to go outside the reactor building is the cooling water, though you’d probably want the reactor coolant itself to go to cooling towers (dry) during the summer. A natural circulation CO2 stream at a pressure of a few bar goes between the internally-insulated reactor vessel and the containment can in its water bath, cooling the RPV by convection.
Electrical output would be controlled by turbine bypass, perhaps also with regenerator bypass. In case of loss of grid, the reactor and turbine units would continue to operate while dumping heat directly to the heat sink. Heat sink would be guaranteed by separate alternators on the turbine shafts (6x) to feed the bus for the cooler water pumps, if used. This creates a nuclear unit which rides through grid outages and has black-start capabilities.
It would also generate something on the order of 330 MW(t) of hot-water supply at full electrical output. At maybe 1030 BTU/scf, this would displace roughly 1.1 million scf/hr of natural gas (HHV).
* Fits a NuScale reactor cell.
* 6 times the electrical output.
* At least 3x the heat-output potential of a NuScale.
* Black start and grid outage ride-through.
* 6x redundancy on the power generation and internal plant power.
NOW is the time for NEI and ANS to justify themselves to the public and jump up to the plate and phone — not the local media — but the network hubs and do a education-spiel to allay this media “over concern” (and veiled permanent shutdown urgings) of Turkey Point and St. Lucie in the way of Irma. Should I hold my breath?
It remains to be seen if the “over concern” is justified or not. And it has nothing to do with factual information. You know as well as I do that the slightest “mishap” will be milked for every ounce of fear factor your foes can place on the mishap. Its obvious that no media PR counter offensive will be launched by the industry. So it all hinges on how well these plants weather the storm, and rather or not they can avoid providing ammunition to the fudistas. If anything major occurs, its lights out for the NE industry in North America.
NE is about to face a critical test. You all better hope the Florida nuke plants have their sh*t together, because, if not, it will spell the end of NE in North America. Your foes will have a field day if the plants do not weather this storm without any mishaps.
Having worked in a gas-cooled reactor station [Heysham 1] in the UK using CO2 as the working fluid, I have one query:
In the UK AGRs, which are all cooled with CO2, Nitrogen is used as a secondary shutdown system should the primary system [conventional borated steel control rods] fail. If you are going to use nitrogen as the coolant, how do you get round the neutron absorption effect?
Neutron absorption is a challenge for nitrogen, but it isn’t dissimilar to the challenge of overcoming neutron absorption effects using light water instead of heavy water.
We had a pretty good discussion of some of the options on Atomic Insights several years ago. Here is a link to the post and the associated comment thread.
Recent Comments from our Readers
The Clinton Nuclear Plant also in Illinois was shutdown essentially for almost 2 years before it was taken over by…
Good Podcast – Very informative One thing that was not discussed is how to deal with a particular fear that…
Renewables people are masters in marketing. Unreliable intermittent generators whose output is all over the place, and usually badly correlated…
Looking at their lineup, Westinghouse seems bound and determined to keep Gen IV in its “place” which is apparently the…
So they are developing a scaled down version of the AP1000, which is a scaled up version of the AP600,…