|
Post by apophys on Sept 13, 2017 20:09:07 GMT
Tungsten emissivity at 2500 K is about 0.4, so no major problems for the radiator in this regard. Higher would be nicer, but temperature and vapor pressure are far more pressing concerns, so 0.4 will do. An alloy of tungsten and 0.3% iridium apparently has its emissivity very insensitive to temperature, at about 0.42, so we can probably use that value for the shell. sci-hub.io/10.1007/BF02649249apophys : Very good work and nice finds! Try contacting Ian Mallet. Ah, I see on Google+ that he's working on an LDR calculator. Ian Mallett imallett has shown up here in the forum before, btw. Maybe you can direct him to this thread? Also, I found another potential filling. Silicon nitride (Si 3N 4) - melting point 2769 K from CoaDE data, and heat of fusion ~1.47 MJ/kg. There are multiple preparations of it with slightly different heats of fusion, apparently. www.azom.com/properties.aspx?ArticleID=53www.azom.com/properties.aspx?ArticleID=260 (best, so I use the average here.) www.azom.com/properties.aspx?ArticleID=77The main problem is it decomposes at 2170 K (according to Wikipedia), so unless it can be prevented from doing so, it is unusable. I'm not sure why CoaDE has it usable above this point. And I'm not sure why there is heat of fusion data at all in this case. Maybe it'll hold together under pressure?... Thorium dioxide (ThO 2) - melting point 3225 K, heat of fusion 4.61 MJ/kg (source [2]). This feels weird somehow. >.> Thorium is very dense and likely to make massive (read: bad) LDRs. Looking at the results of Ian Mallett's work on LDR droplet math, dense is actually good, since it allows smaller beads to work. plus.google.com/collection/AA3jkBThorium dioxide density is 10,000 kg/m 3, which is slightly less than lead. Wikipedia gives it a different melting point than the source I quoted: 3660 K. Nearly the same as Wikipedia's value for tungsten, 3695 K, and higher than CoaDE's, 3643 K. Thorium dioxide is a problem, not because of its density, but because its vapor pressure is probably too high on its own, and tungsten may not be able to contain it. It's possible that we could mix thorium dioxide with something to lower its melting point, but data is lacking.
|
|
|
Post by imallett on Sept 14, 2017 1:22:41 GMT
Hey all, Ian Mallett (author of linked post, again). A few thoughts, by request: - My calculator, also linked on the post, is ready for use. It is correct so far as I know EXCEPT it does not consider heat of fusion or vaporization. This is an important omission, and will be fixed eventually.
- Another consideration for heat of fusion: just like other materials, liquid metals can be supercooled. This can be due to their own radiation. If they finally do solidify, the heat of fusion is released as added temperature, which increases the radiation again. This has implications for droplet radiator efficiency. One less-obvious implication is that if hitting the collector dish triggers the phase change, the droplets can suddenly flare up to blazing heat, while inside the collector, which could be unexpected/damaging. Here's an example and more-detailed explanation with a liquid gold droplet (Cody's Lab on YouTube).
- I think the temperatures y'all are talking about are too high to be realistic—particularly for designs exploiting the heat-of-fusion (and therefore tending to have constant or more-constant temperature throughout the cycle), with no cooler section for the pump to work in. There was some discussion of this on this related thread.
A cursory literature search suggests the very hottest you could conceivably make (active-cooled!) effective machinery (like pumps) is around 1600°K, and numbers more like 1000°K are more realistic. (1000°K is the upper end deemed plausible by the NASA study. As another point of reference, it seems to be the thermal limit for even conceptual nuclear reactor designs). I don't claim either of these are hard boundaries, especially with SF-inal magitech, but I feel it needs to be mentioned, when you're talking about, y'know, 3000°K thermal cycles. That's hard or impossible even for solid state.
- Regarding the encapsulated fusion droplet radiator idea, I think the main problem will be maintaining droplet stability. You're essentially talking about hot lava in a graphite water balloon, being shot at tens, hundreds, or thousands of meters per second into a collector dish, for decades on end. I'm worried that the droplet bubbles will burst if they're too thin. If not from mechanical force, then from thermal stress and varying internal pressures of being heated/cooled—especially if you try to make the bubble thicker to compensate.
If it could be made workable, then the design parameters would favor small expansion coefficients, with a high thermal conductivity and emissivity on the bubble surface and high specific heat (or heat of fusion) on the working fluid. A slow speed would aid longevity, and some messing around with the cube-square law and droplet kinetic energy suggests that making the droplets larger to take advantage is a plausible win.
- For working fluids, I was looking at Antimony (Sb), which has a respectable heat of fusion, a plausible temperature range, and forms well-behaved compounds. However, as a reminder, working fluids do not have to be pure-elemental. I saw mention of a few compounds, and I do expect the optimal working fluid will be some molecule. Higher mass tends to mean higher boiling point and lower vapor pressure, too. With eutectic mixtures, you can even tweak the temperatures to fit to some degree.
Ian
|
|
|
Post by apophys on Sept 14, 2017 5:34:26 GMT
My calculator, also linked on the post, is ready for use. It is correct so far as I know EXCEPT it does not consider heat of fusion or vaporization. One thing that looks like a bug: Your droplet size/spacing optimization results in distances between droplet centers being the radius of a droplet (meaning that droplets overlap, a physical impossibility). I've seen a NASA study on LDRs (first page, posted by matterbeam), but didn't see any mention of an upper temperature limit on machinery. I'm not sure what requirements could lower the maximum temperature so greatly below CoaDE's; a cursory search didn't get me much. CoaDE reactors have coolant outlets at 2500 K. Their turbomachinery is made of amorphous carbon (thus no need for active cooling), and thermocouples are tungsten-osmium. In light of this, boron, aluminum oxide, and aluminum nitride are all fairly reasonable candidates for encapsulated working liquid. For the ~3000 K melting point materials, I was just looking to see the theoretical upper limits of encapsulated droplets. By the way, you may not have heard of the fairly insane boiling uranium reactor concept developed in the CoaDE community recently, or the nuclear-diesel variant of matterbeam ... Extracting work from the expansion of gaseous uranium compounds or actual gaseous uranium metal, it could have a coolant outlet up to 4100 K. Antimony is not workable at all as a non-encapsulated liquid; its vapor pressure is much too high to sustain LDR operation for very long. See wikipedia; at 10 Pa of vapor pressure it is still solid. The NASA study that I saw limits vapor pressures to 1e-9 to 3e-7 mm Hg depending on construction, which translates to about 1.3e-7 to 4e-5 Pa. It could work as a filling to a capsule, but there are certainly better options for that. Tungsten has the lowest vaporization rate of any material known, lower than TaHfC, and much lower than graphite (at 2839 K, graphite has 10 Pa vapor pressure; this is the lowest pressure for which I have data, but it means graphite certainly cannot be used for 2500 K). So my best candidate for ~2500 K is still boron in a tungsten capsule. Emissivity of tungsten is good enough (~0.4 at 2500 K). Expansion coefficient may be a non-issue with tungsten's excellent yield strength. Alloys of tantalum, osmium, or rhenium with tungsten are an option if for some reason pure tungsten is unsatisfactory. For high-ish temperature working fluids of an LDR without encapsulating droplets, I found neptunium and uranium to have very good vapor pressures while liquid (probably usable around 950 to 1400 K for neptunium and around 1400 K to 1500 K for uranium). Specific heats are quite poor, unfortunately. I haven't looked in depth for lower-temperature materials yet, since my primary goal was getting things to work for CoaDE's reactors in the best way possible, and secondary goal to explore the tech limits.
|
|
|
Post by imallett on Sept 14, 2017 7:04:20 GMT
Merrrrp that's indeed a bug. I used the wrong constant in the optimizer. Now fixed. Thanks!The 1000°K number actually came from the paper Liquid Droplet Radiators for Heat Rejection in Space (actually AIAA, not NASA). The 1600°K number comes from melting points of turbine superalloys—not a perfect analogy to pumps, but both are thermally-limited rotating fans, and it was the highest temperature machine I found a reference to actually existing. Because of this, and because these are real machines pushed to presumably tight limits, I'm remain skeptical that real machinery could actually work at significantly higher temperatures. I'm not a (real) engineer, but I know enough to know that nitty-gritty engineering challenges ruin parades. I don't think we can just cite materials that have a higher melting point than our target and assume we can build effective, reliable turbomachines out of them. That said, I can think of various ways to try to pump high-temperature fluids with only minimal moving parts touching high-temperature fluids. For example, maybe you can make the collector and emitter spin, with some kind of spiral track flinging droplets back and forth. The bearing surfaces can be further away, on the outside, at reasonable temperatures. Or maybe the collector and emitter can be oblong, and they can propel the droplets back by shaking or spinning. Or something. The point is, anyway, that high temperatures are big problems, and a material having a high temperature tolerance is just one aspect of the engineering requirements. Moreover, if we assume high-temperatures and the non-standard materials that implies, the other challenges probably get worse, because these materials are probably worse in other respects. I thought about material choices a while back for the sake of the related thread linked previously. Ideally, one basically wants to consider every possible material, remove the ones we can't use for various reasons (gaseous, radioactive, etc.) and sort the rest according to some criterion for "optimal" involving vapor pressure, specific heat, melting point, heat of fusion, chemical stability, price, etc. A computerized search would be ideal.
|
|
|
Post by apophys on Sept 14, 2017 15:56:17 GMT
|
|
|
Post by n2maniac on Sept 15, 2017 6:05:34 GMT
I've brought this up before and I'll bring it up again while we are on the topic: Another important parameter: chemical compatibility: does the solid react with / dissolve in / absorb / let diffuse out of it the working fluid at the working temperature?
|
|
|
Post by imallett on Sept 18, 2017 16:01:30 GMT
I'm impressed with that pump—in-particular that they somehow made a seal out of ceramics. I'd be curious to know if they could make it higher; ceramics tend to be able to do that.
|
|
|
Post by lucubratory on Dec 24, 2017 9:15:31 GMT
apophys : I see now. While pressure increases with both temperature and the number of moles of vapor, meaning the mass loss rate isn't really linear with the vapor pressure, I do realize now that liquid droplet radiators have a tough time with high temperatures. Solid particle or 'sand' radiators might be the only option. Basically for warships it looks like there is no stopping point between solid radiators and dusty plasma. I'm pretty sure that if dusty plasma is possible, economic nuclear fusion is possible with the same level of technology, so the whole field would change incredibly dramatically. Much larger ships would be feasible fielding much larger crew complements, weapons, and manufacturing capabilities. Fusion torchships aren't a given with fusion reactors any more than NSWRs are a given with fission reactors, but the sort of control over magnetic fields and plasma that dusty plasma radiators require could result in viable nuclear torchships of some variety. Manufacturing very large scale space habitats that could hold hundreds of thousands of people and power everything they need would be much more economical and near term. A lot of the fundamental assumptions we make about space combat would need to be re-evaluated.
|
|
|
Post by matterbeam on Dec 27, 2017 22:37:40 GMT
Basically for warships it looks like there is no stopping point between solid radiators and dusty plasma. I'm pretty sure that if dusty plasma is possible, economic nuclear fusion is possible with the same level of technology, so the whole field would change incredibly dramatically. Much larger ships would be feasible fielding much larger crew complements, weapons, and manufacturing capabilities. Fusion torchships aren't a given with fusion reactors any more than NSWRs are a given with fission reactors, but the sort of control over magnetic fields and plasma that dusty plasma radiators require could result in viable nuclear torchships of some variety. Manufacturing very large scale space habitats that could hold hundreds of thousands of people and power everything they need would be much more economical and near term. A lot of the fundamental assumptions we make about space combat would need to be re-evaluated. The dusty plasma radiator is manipulated very differently from a fusion reaction. In the radiator, the 'dust' is hot, charged particles at temperature ranging from 4000K to 400K. The plasma they sit in is charged gas at much cooler temperatures (1000K to 200K) that is being manipulated by rather weak magnetic fields over long (several meter) ranges. By acting on the plasma, you can create a force that is transmitted to the dust. This way, you can extend, spread, organize and collect the dust without physical structures. Fusion plasma is massively different. It is extremely hot and under incredible pressures. The magnetic fields involved are also very powerful, but they have to be arranged very carefully as any 'gaps' in the magnetic fields will allow the fusion plasma to squirt out and fail to fuse. Which the lessons learned from developing dusty plasma radiators will be useful for fusion technology, the challenges involved are quite different and solving one might not imply that the other becomes any less easier!
|
|