|
Post by matterbeam on Nov 29, 2017 13:45:42 GMT
matterbeam Personally, I think we need to look up literature in existing liquid-metal turbines; the properties of a compressible gas coolant are extremely different from a liquid coolant. Liquid metal turbines have been built before, but I'm having trouble finding scientific papers on the subject. IIRC, the Soviet Alfa-class submarines used a lead-bismuth design. Once you can extrapolate the differences between, say, a naval gas turbine and a naval liquid metal turbine, we can then use that to extrapolate a hypothetical min-maxed Very High Temperature reactor turbine that would be mounted on a spacecraft. I am not sure than liquid coolants are a good idea. It would impose an upper limit on the possible operating temperatures, it would create a lot of waste drag on the turbine blade and you cannot make the flow too fast because otherwise you'd get cavitation. I am also unsure of how a turbine would extract energy from a closed loop incompressible flow in the first place. It cannot expand the fluid, doesn't cool it down and cannot slow it down either...?
|
|
|
Post by newageofpower on Nov 30, 2017 0:56:31 GMT
I am not sure than liquid coolants are a good idea. It would impose an upper limit on the possible operating temperatures, it would create a lot of waste drag on the turbine blade and you cannot make the flow too fast because otherwise you'd get cavitation. I am also unsure of how a turbine would extract energy from a closed loop incompressible flow in the first place. It cannot expand the fluid, doesn't cool it down and cannot slow it down either...? Hmm. Less dense coolants suffer from supercriticality issues, which tends to reduce ideal thermal properties and increases pressure stress on the system. I just tried to use a gaseous coolant in a 50 MW CDE reactor; cooling efficiency drops, so I'm forced to bloat up volume required. Furthermore, pump size jumps up dramatically, and power efficiencies for a given outlet temperature decreases. Perhaps liquid metal within the reactor core itself, then the gas or water outer loop (if we're using expansion turbines) using a heat exchanger... Now we're back to high Brayton specific power levels.
|
|
|
Post by matterbeam on Nov 30, 2017 1:43:09 GMT
I am not sure than liquid coolants are a good idea. It would impose an upper limit on the possible operating temperatures, it would create a lot of waste drag on the turbine blade and you cannot make the flow too fast because otherwise you'd get cavitation. I am also unsure of how a turbine would extract energy from a closed loop incompressible flow in the first place. It cannot expand the fluid, doesn't cool it down and cannot slow it down either...? Hmm. Less dense coolants suffer from supercriticality issues, which tends to reduce ideal thermal properties and increases pressure stress on the system. I just tried to use a gaseous coolant in a 50 MW CDE reactor; cooling efficiency drops, so I'm forced to bloat up volume required. Furthermore, pump size jumps up dramatically, and power efficiencies for a given outlet temperature decreases. Perhaps liquid metal within the reactor core itself, then the gas or water outer loop (if we're using expansion turbines) using a heat exchanger... Now we're back to high Brayton specific power levels. This is for the Advanced Solar Energy series, so criticality is less of an issue. The heat source is a grid of microchannels heated from the top and bottom by concentrated sunlight. Gas passed through the heat exchanger. I am unsure whether simply heating the gas this way will cause it to increase in pressure, or whether turbo-generators have both a compression (pre-compress the gas, heat it further, then release the pressure) and an expansion stage.
|
|
|
Post by n2maniac on Nov 30, 2017 7:01:19 GMT
Looking at the BNTR concept described here (), it seems much more reasonable to use a Xenon/Helium mix as coolant. n2maniac : What is the 3.3MJ/kg figure? I have started looking at how turbine performance is determined, and high compression ratios are ideal. The higher the better. This is because the higher the pressure gradient, the better the temperature drop, and a large temperature drop across the turbine allows for greater Carnot efficiency. I used the following equation : Inlet Temperature/Outlet Temperature= Pressure Ratio ^ (1 - 1/y) y is the adiabatic gas constant. For helium, a monoatomic gas, it is equal to 1.6. So if you have a pressure ratio of 100, you can get a temperature drop of 5.62. If your inlet temperature is 2000K, it can be lowered to 355K at the exit. The highest compression ratio I've come across is the Boeing 777's latest jet engines, with 61:1. I am unsure how compression ratio relates to the stresses on the turbine blades, which determines how strong they'll have to be, and more importantly, how much the thing will mass. Only the latter figure really matters. Summary of my calc: get helium's speed of sound at 3000K, take 80% of that (assuming Mach 0.8) -> 2.6km/s, 1/2 * v^2 for kinetic energy per mass -> 3.3MJ/kg (which conveniently relates to flywheel required strength). Yes, these are only ballpark numbers and actual blade design would need to be done to get exact numbers, but helium is very punishing to use at that temperature. Maybe what is required is an actual crude blade / compressor design? Number of stages * pressure ratio of each stage is the total. Generally, the pressure rise on a stage is related to the square of the mach number. However, since the diameter and RPM between successive stages is similar (or constant), higher pressure stages will have a lower mach number. Also, either axial velocity or cross sectional area will have to decrease substantially with high pressure ratios, more likely a mixture of the two. Assuming per-stage compression is constant is probably not a good idea. This page has some information on the subject (including a better suggestion to go supersonic in this case), and the specific note on mach number squared can be found here.
|
|
|
Post by newageofpower on Nov 30, 2017 7:28:07 GMT
This is for the Advanced Solar Energy series, so criticality is less of an issue. The heat source is a grid of microchannels heated from the top and bottom by concentrated sunlight. Gas passed through the heat exchanger. I am unsure whether simply heating the gas this way will cause it to increase in pressure, or whether turbo-generators have both a compression (pre-compress the gas, heat it further, then release the pressure) and an expansion stage. Assuming the gas is behaving anywhere similar to an ideal gas (which, Helium is fairly close to) heating it will always increase pressure. Additionally, you'd want as much coolant per volume in your cooling system as possible (well, while maintaining safe pressure in coolant lines and pumping machinery) as that dramatically increases heat flow rates, so an optimized design would be under significant pressure. I'm uncertain about proposed Helium working fluid; a quick Google search gives me a dozen articles from 1968 to 2011 on high temperature helium embrittlement. Perhaps selecting a working fluid with a reduced tendency to diffuse into solid materials? It's not like Helium has excellent thermal characteristics as a coolant...
|
|
|
Post by matterbeam on Nov 30, 2017 13:45:26 GMT
n2maniac: I've settled on a half-half mix of Xenon and Helium for that reason: it mixes the high density (low mach number) of Xenon with the high thermal conductivity of Helium. This 50/50 molar mix of average molar mass 66.5g/mol. It should have an individual gas constant of 1071J/kg/K. At 3000K, it would have a speed of sound of 2267m/s according to this calculator: www.engineeringtoolbox.com/speed-sound-d_519.html80% of that is 1813m/s. That is 1.64MJ/kg. For titanium, this is equivalent to a flywheel of 21.56kg/m^2 if it were made of simple titanium, or 0.25kg/m^2 using silicon carbide. If I used a much denser Xenon-only gas, the speed of sound drops to 543m/s at 3000K, and the flywheel-equivalent figure using silicon carbide becomes 0.014kg/m^2. However, not the heat exchanger's surface area has to be 51 times larger... Thanks for the links. newageofpower ; Good point. Heat flow rate actually tailors off above 1bar pressure, at least for helium. 1000 tor corresponds to 1.33bar. Helium embrittlement might be something to look out for, but it should be no worse than hydrogen, and we've designed nuclear thermal rockets to handle high temp hydrogen for minutes on end...
|
|
|
Post by Rocket Witch on Nov 30, 2017 15:19:54 GMT
Rocket Witch : Thank for your input. If I'm getting this correctly, I can take the figures for an actual turbine, replace its turbine blades with a ceramic, calculate the ratio of specific strengths, and reduce/increase the mass of new turbine by that ratio, assuming the actual and new turbine are operating under identical conditions. How useful is this calculation? I'm not sure. The turbine dimensions and shape can't change either, or the forces acting on it will too; a ceramic turbine able to withstand large temperature deltas may have to be designed quite differently. What do you think about protected turbine blades? A very high specific strength material, such as zylon fibres, covered by a ceramic 'shield' so that it is not in direct contact with the gasses. Active cooling would run in the gap between the shield and the blade core, to prevent heat from being carried across. I'm not sure how you can spin something very fast and pump liquid through it. I guess you'd need valves in the hub that feed some coolant through each time a hole in the turbine blade section lines up with it, like how a rotary cannon feeds rounds. Depending on how fast the turbine is expected to get, I'm not sure this can work reliably. To my knowledge, rotary guns misfire all the time (every few cycles of all the barrels lining up with the feed in turn, the occasional one may not receive a round) and a coolant pump system set up in the same way may be similarly prone to. This could be a serious problem due to ueven cooling distorting or fracturing the turbine blades. There may be some design problem parallels with gas core nuclear thermal rockets here. A cladded or composite turbine will face issues at the interface between the materials, and that's before adding the coolant plumbing. The thermal expansion coefficients and stress responses (foremostly Young's/elastic modulus and elongation at break) of the two need to line up very closely across the massive temperature delta proposed, making your choice of compatible materials quite limited. This is probably what makes reinforced carbon-carbon quite popular, as it uses two forms of the same element, and the graphite half in particular has, being graphite, high lubricity between molecular layers which can easily break and reform bonds. One positive feature however is that thermal conductivity and specific heat tend to rise as materials heat, so the cooling apparatus may be proportionally smaller/lighter.
|
|
|
Post by matterbeam on Nov 30, 2017 16:26:07 GMT
Rocket Witch : Thank for your input. If I'm getting this correctly, I can take the figures for an actual turbine, replace its turbine blades with a ceramic, calculate the ratio of specific strengths, and reduce/increase the mass of new turbine by that ratio, assuming the actual and new turbine are operating under identical conditions. How useful is this calculation? I'm not sure. The turbine dimensions and shape can't change either, or the forces acting on it will too; a ceramic turbine able to withstand large temperature deltas may have to be designed quite differently. What do you think about protected turbine blades? A very high specific strength material, such as zylon fibres, covered by a ceramic 'shield' so that it is not in direct contact with the gasses. Active cooling would run in the gap between the shield and the blade core, to prevent heat from being carried across. I'm not sure how you can spin something very fast and pump liquid through it. I guess you'd need valves in the hub that feed some coolant through each time a hole in the turbine blade section lines up with it, like how a rotary cannon feeds rounds. Depending on how fast the turbine is expected to get, I'm not sure this can work reliably. To my knowledge, rotary guns misfire all the time (every few cycles of all the barrels lining up with the feed in turn, the occasional one may not receive a round) and a coolant pump system set up in the same way may be similarly prone to. This could be a serious problem due to ueven cooling distorting or fracturing the turbine blades. There may be some design problem parallels with gas core nuclear thermal rockets here. A cladded or composite turbine will face issues at the interface between the materials, and that's before adding the coolant plumbing. The thermal expansion coefficients and stress responses (foremostly Young's/elastic modulus and elongation at break) of the two need to line up very closely across the massive temperature delta proposed, making your choice of compatible materials quite limited. This is probably what makes reinforced carbon-carbon quite popular, as it uses two forms of the same element, and the graphite half in particular has, being graphite, high lubricity between molecular layers which can easily break and reform bonds. One positive feature however is that thermal conductivity and specific heat tend to rise as materials heat, so the cooling apparatus may be proportionally smaller/lighter. The questions you raised are also what I am trying to determine. How wide does a turbine have to be handle a gas at a certain pressure, flow, temperature and composition? I cannot find any single description online of the turbine component alone. An actively cooled turbine is pretty common today - I think however that they use gasses such as steam or hydrogen within the blades themselves. The hydrogen would then escape out of the turbine blade and mix into the main coolant stream. For our space application, I would use a dense, initially cold gas flowing through a hollow blade from the shaft to the tips, but escaping out of their ends into a collection drum to be recycled.
|
|
|
Post by n2maniac on Dec 1, 2017 5:03:44 GMT
Regarding cooled turbines: This is common, either bleeding out cold gas from tiny holes or built hollow for sodium heat pipes.
|
|
|
Post by newageofpower on Dec 1, 2017 6:07:22 GMT
Minutes is not a severe issue. However, I assume your turbines exist to produce electrical power, yes? If we're supplying electricity for electric propulsion, you may need months of endurance...
|
|
|
Post by matterbeam on Dec 1, 2017 11:19:09 GMT
Minutes is not a severe issue. However, I assume your turbines exist to produce electrical power, yes? If we're supplying electricity for electric propulsion, you may need months of endurance... Well, unless we are aiming to achieve continuous acceleration over interplanetary distances, then the turbines will function for two weeks at most. Even a mediocre 10 milligee racks up to 118km/s over two weeks.
|
|
|
Post by newageofpower on Dec 1, 2017 15:25:06 GMT
Well, unless we are aiming to achieve continuous acceleration over interplanetary distances, then the turbines will function for two weeks at most. Even a mediocre 10 milligee racks up to 118km/s over two weeks. B-but muh 500km/s Apophys superliner! How will I get my Saturnian beauty products shipped on schedule? (Also, 10mg acceleration is actually fairly high for an electric drive craft, and resisting embrittlement for minutes is still dramatically different from weeks.)
|
|
|
Post by n2maniac on Dec 2, 2017 20:45:38 GMT
Assuming the gas is behaving anywhere similar to an ideal gas (which, Helium is fairly close to) heating it will always increase pressure. Additionally, you'd want as much coolant per volume in your cooling system as possible (well, while maintaining safe pressure in coolant lines and pumping machinery) as that dramatically increases heat flow rates, so an optimized design would be under significant pressure. I'm uncertain about proposed Helium working fluid; a quick Google search gives me a dozen articles from 1968 to 2011 on high temperature helium embrittlement. Perhaps selecting a working fluid with a reduced tendency to diffuse into solid materials? It's not like Helium has excellent thermal characteristics as a coolant... will always increase pressure -> Heat exchangers operate roughly isobaric, turbines roughly adiabatic Thermal conductivity of a gas is roughly independent of pressure between the lower limit where the gas collides with the walls of the test apparatus more then itself (molecular flow), natural convection / turbulence reduces the boundary layer, or the gas becomes a supercritical fluid. If the mix is He/Xe, a Ne/Xe or other noble gas mixture would likely also be suitable. Helium is the second best gaseous coolant (a property of its high speed of sound, which also causes other turbine problems). matterbeam Stepping out more to the higher power density side (read: mad science side), UF6 (or another gaseous fission fuel) seems like it would be an interesting choice if corrosion issues are solvable. You get rid of a hot side exchanger, actively cooled blades could be working with temperatures much higher than their safe operating limits (aided by UF6's somewhat poor thermal conductivity), the low speed of sound (barely 300m/s at 3000K, though that assumes no dissociation) means relatively low centrifugal stresses on the turbine blades (and thin, very effectively cooled turbine blade walls). Just reject all the heat to the highest reasonable operating temperature for the blades and radiators, add the active cooling heat load in there (some of which is probably going to be cooling loads on the reactor walls and control rods), and let the upper temperature soar based on operating power demand and just how well surfaces can be cooled. Or, if power density is more important (and heat rejection is cheap), thicken the blades and bump up the pressure approaching the supercritical region for the gas.
|
|
|
Post by n2maniac on Dec 2, 2017 21:35:11 GMT
Hmm, good reference on the lack of thermal stability of UF6 above about 2400K (check the figures at the end): ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19740007544.pdfMost of its decomposition products are also volatile at that temperature, so it is probably reversible (so long as hot, free fluorine isn't scavenged by literally anything else).
|
|
|
Post by newageofpower on Dec 2, 2017 23:26:30 GMT
so long as hot, free fluorine isn't scavenged by literally everything else There. I fixed that for you.
|
|