January 24, 2020

Open Thread 44

It's our usual open thread. Talk about whatever you want, even if it's not military/naval.

A recent news story of interest is the decision to name CVN-81 after Dorie Miller, the first African-American to win the Navy Cross for his actions during the Attack on Pearl Harbor. I'm of two minds on this one. On one hand, Miller is far more deserving than most recent recipients of warships, particularly the carriers. On the other, I'd really rather we used the traditional carrier names for carriers. Right now, Lexington, Yorktown, Saratoga and Ranger are all unused. Still, this is one of the better naming decisions to come out of the Navy Department in recent years.

Overhauls for 2018 are Why the Carriers aren't doomed Parts two and three, Stability, Pre-Dreadnoughts, Basics of Naval Strategy and Russian Battleships Part 2. 2019 overhauls are Commercial Aviation Part 5, Falkands Part 10, the Spanish-American War Part 1, The NOAA Commissioned Corps, Ship Structure and Strength and The Mk 23 Katie.

Comments

  1. January 24, 2020cassander said...

    Dorie Miller is a better name than a politician, but I still say that we should name a bunch of carriers after Apollo landing sites. USS Ocean of Storms and Sea of Tranquility are great names, and they carry the implication of "USS Somewhere that you aren't cool enough to go to yet, but we are." If we haven't gotten to Mars by the time we run out of those, we should re-evaluate national priorities.

    And for the record, here's how other ships should be named:

    Large Surface Combatants: Distinguished (and dead) Americans who haven't had a ship named after them before.

    SSNs: States, these are the ships that ensure command of the sea, which is the traditional role of battleships and the most fundamental of naval missions.

    SSBNs: Sea creatures, they lurk in the depths.

    Amphibs and command ships: Victorious battles, and again, let's find some new ones. Where's my USS 73 Easting?

    Other combatants: City names.

    Auxiliaries: Who cares?

  2. January 24, 2020DampOctopus said...

    As promised:

    In high orbit over Vesta, BSS Stealth-Totally-Works-You-Guys prepares for its maiden voyage. It resembles a disc, of area A, edge-on to the sun. One side is covered by the openings of an array of mirrored Winston cones whose bases connect, via a series of radial light guides, to an internal radiator. The radiator has an area of a, and the Winston cones expand from this to cover almost the entire face of the disc. The radiator emissions are thus beamed into a solid angle of Omega = pi x a/A steradians: down the boresight the entire surface has the white-hot glow of the radiator, but from any other angle the Winston cones reflect only the cold emptiness of space.

    The radiator and the dense tankage/lifesystem module are in the core of the disc, but a fraction fr=10% of the STWYG's mass is dedicated to the light-guide array that constitutes most of the disc area. Most of this array array consists of flimsy mirrored cones, and its average areal density is only D=10kg/m2. The STWYG thus has a total mass of M = D x A / fr.

    All other external surfaces are painted black and cooled to 2.7K, blending in perfectly with the cosmic microwave background. This is overkill - the temperature could be higher and still remain undetected by a realistic telescope - but it ensures that no infrared sensor, however sensitive, should detect a contrast between the STWYG and the background sky. Keeping the external surfaces at this temperature isn't difficult, despite the warm interior: vacuum is an excellent insulator, and doesn't mass much.

    Once the STWYG has cooled down to ambient and disappeared from everyone's sensors, its Belter captain intends to take it to Earth deliver a surprise birthday present to the Earthican president. Approximating the eccentric orbit of Vesta as a circle with a radius of 2.4AU, and ignoring its puny gravity, the captain calculates that a Hohmann transfer would require a burn of 4.5km/s followed by 405 days of coasting. If the STWYG were instead to maintain a continuous, low acceleration ... the captain fiddles a bit with a Python script, gives up, and guesses that twice the delta-v (or vd=9.0km/s) should be enough to make the trip within double the time (or t=810 days), despite the inefficiency of this type of transfer. The STWYG will need to maintain an average acceleration of vd/t = 0.00013m/s, or 0.013 milligee. And the birthday present will be several years late.

    The STWYG has a small ion drive capable of supplying the required thrust. Over the course of the trip it will use fuel amounting to ff=10% of the mass of the ship. This is a small enough fraction that the drive exhaust velocity can be approximated as ve = vd/ff = 90km/s. The total kinetic energy imparted to the exhaust is Ke = 0.5 x ff x M x ve^2, and the drive will need to be supplied with Pe = Ke/t of power.

    The STWYG's reactor is an exotic beast: a gaseous-core fission reactor, with a critical mass of vapourised uranium suspended in a vortex of injected coolant to keep it away from the chamber walls. Such an arrangement would be difficult to achieve in a gravity well, but under the STWYG's sub-milligee acceleration it is sustainable. Ultraviolet radiation from the Tc=10000K reactor core passes through the chamber walls to heat the main coolant loop.

    The other end of the coolant loop passes through the radiator. For efficiency, this is made of tungsten and runs at just below its melting point of Tr=3695K. With area a, it can dissipate a power of Pr = sigma x a x Tr^4 where sigma is the Stefan-Boltzmann constant. The power-generation system operates at the maximum theoretical efficiency for a Carnot cycle: e = 1-Tr/Tc = 63%. This efficiency also relates the electrical power Pe supplied to the drive and the waste heat power Pr that must be dumped through the radiators: e = Pe/(Pe+Pr), giving us Pr = Pe x Tr / (Tc-Tr).

    Putting all of this together, the solid angle of the beamed radiator emissions, in steradians, is:

    Omega = (pi/2) x (D x vd^2) / (sigma x (Tc-Tr) x Tr^3 x t x fr x ff)

    or 0.0001 steradians under the assumptions above; 0.0008% of the whole sky. If the STWYG is to be detected by a sensor drone passing through this emission cone, the Earthicans will have to have a lot of them, or be very lucky.

  3. January 24, 2020DampOctopus said...

    I went into a bit more detail here than I originally planned, and there are still plenty of details I've skipped over. In particular, I've handwaved the delta-v requirements for the continuous-thrust case; I think I've been conservative, but I can be more rigorous if anyone's doubtful on this point. I've also ignored detection methods like active radar, thermal polarisation, occultation of background objects, reflected sunlight, thermal emissions from exhaust, etc. Keeping the STWYG edge-on to the sun minimises the problem of absorbed heat from sunlight, but it could still be significant.

    I've assumed theoretically-optimum efficiencies in the drive and the powerplant, ignored hotel loads on the latter, assumed mirrors to have perfect albedo, etc. Changing these assumptions will increase the final beaming angle, but only by a small-ish factor.

    The STWYG's ion drive is realistic, with performance comparable to some ground-tested technologies, but I have assumed one piece of supertech: the gaseous-core fission reactor (with associated cooling system). Such reactors are commonly proposed as the basis for a high-performance nuclear-thermal rocket, and less commonly considered for power generation. Keeping them stable is hard, and so is extracting power (possibly electromagnetically?) and cooling them, but at least they're thermodynamically possible ... and in engineering terms, probably still easier than fusion.

    The final equation is valid across a wide range of conditions, if you'd like to vary the assumptions. If you set Omega to pi (i.e. set a=A), you can even test the performance of a non-stealthy ion-drive craft, with an unbeamed radiator.

  4. January 24, 2020Lambert said...

    Low power and long burns aren't a panacea.

    If you need to burn for months on end, your adversary can be compiling and processing all that data. They can detect things well below the noise floor for a single exposure.
    (Is this classical uncertainty in E and t?)

    Also optics is just the limiting case of wave behaviour for small wavelengths. I'll cover the far side of the moon with a radio telescope array and be able to resolve the blackbody radiation that diffracts around your lovely mirrors.

  5. January 24, 2020DuskStar said...

    I'm going to have to take the opinion that carriers should never be named after an individual, whether or not that person was a president. To me it just gives the impression that this person is more important than a state, more important than a city - because while cities and states get SSNs and SSBNs, this person gets a carrier.

    Naming a carrier after an enlisted individual - even if they were the first African American to win the Navy Cross - is just incomprehensible to me from that perspective.

  6. January 24, 2020bean said...

    @DampOctopus

    Thanks for that. It was very interesting, although I do have to question several of your assumptions. I can't speak to the optics end, as that's not a field I've ever had a great grasp on. I don't think your assumption that you could get anywhere near background with passive cooling is a good one. Even edge-on you're going to pick up a substantial amount of sunlight, and any heat leakage is going to push up temperature even more. Given that you're absorbing north of 1 kW/m2 near Earth, that's going to push temperature way up. If this disc has 1% of its surface area absorbing solar heat, and disperses it perfectly across its entire surface (so 10 W/m2), my math says it will reach an equilibrium temperature of 115 K. That's still very low, but not anywhere near "literally disappears into the background". The fourth power of temperature law is nasty. And at that sort of scale, "vacuum is a good insulator" starts breaking down around the rest of your structure, too. Even if we neglect the heat flow through the physical connections, you're going to have radiation from the hot parts inside. Painting them silver can keep it down, but you're still looking at another 10 W/m2 of interior section, minimum. Figure another 10% of your temperature, although now the fourth power is in your favor, as you're still under 120 K.

    I also will challenge your orbits. Rule of thumb is that quasi-circular low-thrust transfers take delta-V equal to the difference in orbital speeds, so I get 10 km/s for Vesta-Earth. But that's assuming you're in free space at the start and end, and completely neglects capture. That's probably another couple km/s on the Earth end at least, unless you want to sit on the very edge of the Earth's hill sphere. John will probably be able to speak to this much more effectively. I haven't tinkered with heavy astrodynamics since college, and don't have STK access any more.

  7. January 25, 2020DampOctopus said...

    @Lambert

    Integrating a signal over time only works if you know how the source is moving, so you can stack up images at the appropriate offsets.

    Detecting the radiator emissions at long radio wavelengths, where they're unbeamed, is a nice idea. But that's in the Rayleigh-Jeans regime, in which spectral radiance is only linearly proportional to temperature, and a 3695K radiator is only ~10x as hot as the sky at metre-plus wavelengths. I can work it out thoroughly if you like, but I don't think you'll see anything even with a moon-sized telescope.

    @bean

    I skipped over cooling, but yes, there'll need to be an active system. The trouble is that heat needs to be pumped from the sunlit surface at Tb=2.7K to the radiators at Tr=3695K, with efficiency ~Tb/Tr. Allowing it to heat to (say) 50K makes this much easier, but then you have to worry about infrared telescopes.

    As a cheap solution, you can reshape your hull as a sun-radial cylinder, with an angled mirror over the sunward end. This means there's a 0.5-degree cone in which you can be detected by reflected sunlight, but this is only 0.00006 steradians, which is less than your radiator emissions.

    And it sounds like my orbits were about right. High orbit over Vesta is pretty close to free space, and I'm not concerned about Earth capture: you can deliver a cake (or other payload) as you pass LEO at 11+km/s relative velocity.

  8. January 25, 2020Lambert said...

    You've got months of supercomputer time to try to work this stuff out.
    I'd imagine it looks a bit like 'run some dumb but computationally cheap analysis over the whole dataset to generate candidates, then iterate over all plausible orbital elements for each candidate and see if anything comes up.' But I'm not a CS person.

    And once you do have a lock, you can look back through the dataset in detail to try to work out its past trajectory, as well as following it in future.

    OTOH, the sea is big and hard to find stuff in. Space is even bigger. But stuff stands out better.

  9. January 25, 2020bean said...

    Thinking this over more, this is a violation of the time requirements, which were intended to reflect typical transit times with the technology assumed for the analysis. If the war is likely to be decided by conventional fleets in a couple of months, then a stealth ship which takes 2 or 3 years to get to the destination just isn't that useful. If the available tech stretches to gas-core fission reactors, I’m not going to be using hohmann transfers. I’ll be going a heck of a lot faster. Ve of 90 km/s might be kind of high with that tech, but cutting it some leaves me with a lot more thrust. I scrounged up some old code, and am running an analysis of Vesta-Earth transfers under arbitrary conditions and delta-V values, which should give some idea of what typical transits will look like. I'll post the results when they finish.

    This design also has the problem of getting home, now that I think about it. Yes, more fuel will solve it, but you’ve now taken the ship out of service for something like five years to deliver this cake.

  10. January 25, 2020bean said...

    I've finished my run. It used a Lambert solver, and treated both Earth and Vesta as points, so neglecting capture values because that would have been really hard to set up. The output is the percentage of the time you can arrive at the destination (here Earth) within a given amount of time using a certain amount of delta-V. For 30 km/s, the window opens up at 150 days, although that's only 4.5% of the time. The window for 40 km/s opens at 120 days at 6.7%, while the windows for 50 km/s and 60 km/s are both at 90 days, although the values of 1.8% and 12% are rather different. None of these should be particularly difficult with the sort of drive tech you're describing, and with 60 km/s, I can make a 300-day transit all but .6% of the time. I don't know enough about nonimpulsive transfers to know how much leeway you have with launch windows, but if we go for a semi-worst-case (40 km/s delta-V, 50% probability of making the transit) then you actually have to do it in about 480 days, not 810.

    Also, what's the basis for the assumption that I can't detect your drive plume? Ion thrusters produce a visible glow, and I suspect that looking for them will be made easier because that glow is going to come from the ions recombining, which in turn means very distinct wavelengths. This has a design which shows 36% of power draw coming from the ionizer. If I assume it's 50% efficient at ionizing things, then you're spitting 18% of thruster power (or 30% of thrust power) out the back in the form of ions that would really rather not be ions. I'd guess that very nearly 100% of this would be released when the ions stop being ions, but even if I cut it in half again, that's still going to be something like 4 kW/N for that particular ion thruster. If I assume the same ratio of emissions to thrust for your ship, I get .52 W/kg of ship mass for the stated acceleration. Which doesn't sound like a whole lot, but that comes out to .5 kW/ton, and you're likely to have a lot of tons. You could solve this problem with mass drivers, but that's likely to lower efficiency quite a bit.

  11. January 26, 2020DampOctopus said...

    @bean

    Thanks for doing a more thorough orbital simulation. Lambert's problem means that you're solving for a pointlike transfer burn, right? And the delta-v you're quoting is for the Vesta-Earth transfer burn only, assuming no velocity-matching at the other end, the same as I did? And for each delta-v, you've quoted the lowest transit duration for which the fraction of the time that it's possible is non-zero?

    In making this comparison, remember that I've allowed the STWYG only a miserly fuel fraction of 10%. A nuclear-thermal rocket with the same gas-core reactor temperature (10000K) has a maximum possible exhaust velocity of 16km/s; with the same fuel fraction, it can't even manage a Hohmann transfer.

    Still, even if we halve the allowed transfer time t and (guessing again) double the required delta-v dv, the STWYG's radiator emission beam expands (per the last equation) to 0.006% of the sky. You can be quite a bit more pessimistic even than this and still have a fairly stealthy spacecraft.

    For getting home, I'd been assuming the STWYG would fire up another radiator, an omnidirectional one, and act like a regular fuel-efficient (but unstealthy) spacecraft for the rest of the trip.

    Regarding the drive plume, I'd been thinking only of thermal interactions between ions, which scale as the square of mass flow rate (so low at low thrust) and go down as the exhaust is better collimated. I hadn't thought of the recombination emissions you've brought up. You're right that these are particularly bad because they're spectrally narrow, so you can look for them through a narrow filter, and the redshift gives you velocity information, too. The ionization energy for cesium is 376kJ/mol, or 0.07% of the kinetic energy at 90km/s. (This implies that the ion source in your example is really inefficient; this seems plausible for a thermal ion source, but I'm not familiar enough with these to be sure.)

    To recombine, however, the ion has to actually hit another particle, and there aren't many of those in interplanetary space. This paper (p777, top-right) describes a glow discharge in their ion thruster which appeared when the pressure in their vacuum chamber increased from 1e-8 to 3e-6 millibar. Interplanetary space has a density of about 1 particle per cm^3, equivalent to ~4e-17 millibar (at 300K). If the same recombination glow is emitted from a drive plume a billion times longer, I suspect it's going to be very difficult to detect.

  12. January 26, 2020Lambert said...

    You have to shoot electrons back into the plasma plume, right? Otherwise the ions start flying back and hitting your lovely spaceship.

    Sounds like you'll get a lot of 12 and 21eV photons coming out of that.

    What's the 'noise floor' for that kind of UV like? If the sky in general's dim enough, you might still be able to detect a fairly big plume. I'd try and crunch the numbers but there's far more important rareified gasses for me to be characterising right now.

    Also I propose that the next SSBN be named Oklahoma. ;-)

  13. January 26, 2020bean said...

    Lambert’s problem means that you’re solving for a pointlike transfer burn, right?

    Correct. I don't have the tools for more detailed analysis of long burns. That's why I cut my assumed Ve. At a constant power, you get twice the thrust and a quarter of the burn time. And because I'm not trying to be stealthy, I can probably have a lot more power to play with.

    And the delta-v you’re quoting is for the Vesta-Earth transfer burn only, assuming no velocity-matching at the other end, the same as I did?

    No. I included stopping at Earth. Didn't want to tinker with old code too much.

    And for each delta-v, you’ve quoted the lowest transit duration for which the fraction of the time that it’s possible is non-zero?

    Correct. I can post the full table somewhere.

  14. January 26, 2020DampOctopus said...

    You need to get rid of your excess electrons to keep yourself neutral, but you don't need to shoot them in the same direction. With the positive ions drifting away at 90km/s, they're never recombining with the same electrons you tore from them.

    The ionization energy of cesium is 3.9eV, which is in the near ultraviolet. It's within the emission spectrum of the sun, so I'd expect there to be plenty of galactic background from similar stars.

  15. January 26, 2020bean said...

    To expand on my earlier comments, yes, I'm using a Lambert analysis for this, primarily because it's the best I can do with the tools I have. But I basically can't see anyone with the tech you describe settle for lazing around in Hohmann orbits or anything close to them, and that rapidly brings up the issue of "what use is your stealth ship if I can resolve this whole thing with a few conventional ships before it gets there?"

    A nuclear-thermal rocket with the same gas-core reactor temperature (10000K) has a maximum possible exhaust velocity of 16km/s; with the same fuel fraction, it can’t even manage a Hohmann transfer.

    Which is why you wouldn't have a fuel fraction that low. The exact best fuel fraction is a matter that's going to take a lot of optimization to figure out, but I'm having trouble seeing a situation where you'd actually want a mass ratio below 2 for interplanetary transfers.

    I'm surprised that the ion source is so inefficient. Are you sure they're not multiply-ionizing them or something? (Of course, if it's Cesium, probably not.)

    Also I propose that the next SSBN be named Oklahoma. ;-)

    My understanding is that Oklahoma is off the table until Oklahoma City (SSN-723) is retired.

  16. January 26, 2020DampOctopus said...

    Okay ... the technology I assumed was gas-core fission reactors at 10000K (hence 16km/s exhaust velocity for nuclear-thermal rockets), and 90km/s ion drives. You've worked out transit times using 30-60km/s delta-v to be 150-90 days, which I'm guessing you feel to be appropriate for this technology.

    Without working out a continuous-thrust trajectory (which is hard), I can do a pessimistic estimate in which the STWYG more-or-less matches the transit with 30km/s of delta-v in 150 days. Assuming vd=15km/s burns at either end, to be done within a time t=7.5 days (5% of the total transit time; probably close enough to pointlike), and setting ff=0.15 to keep the drive performance the same, I get a beaming angle for the radiator emissions of 0.02 steradians.

    That's the same transit time as a similar-tech non-stealth spacecraft with a 30km/s transfer, and less than double the transit time of a 60km/s transfer. Radiator emissions are restricted to 0.16% of the sky. Stealthy enough?

    On the ion source, the paper you linked explicitly noted that cesium is invariably single ionized. Their table gives the ion production energy as 1.2keV/ion. Maybe some of the energy goes into separating the unionized atom from the rest of the cesium?

  17. January 26, 2020bean said...

    Hmm. That's by far and away the best argument I've seen for stealth. I do think that you may be unduly optimistic about your ability to pump heat around, but even with that, you can hold down your signature a lot more than I thought. I can't say that it's 15 km/s on each end, because the algorithm works by calling a bunch of separate instances of departure days and transit times, and then figuring out the delta-V total for each, but this seems a reasonable approximation.

    Re ionization, I'm not sure that cesium will be the propellant in question, but in retrospect, there's no way it's going to multiply ionize.

    Actually, one question. How well can you control your pointing angle for the directional radiator? If you're keeping edge-on to the sun, that's going to force you to sweep a reasonable area, and it's only going to get bigger if it's a fixed angle. The other guy may not be able to get 100% coverage of you, but how many platforms do they need before there's a decent chance you accidentally sweep one of them?

  18. January 27, 2020DampOctopus said...

    Regarding the pointing direction, I think you'd put the Winston cones on gimbals, so you can point them almost anywhere while staying edge- or end-on to the sun. You'd keep the pointing constant, so you're not sweeping out more solid angle than you need to.

    Yes, the power/cooling system is thermodynamically possible, but wildly optimistic. Thinking about it further, I think you'd run the core gas through something like a Carnot cycle (either sequentially, or with a travelling wave down a linear cavity), and extract power electromagnetically when it expands. Your waste heat then comes as visible-band radiation from core gas in the cool phase of the cycle, which you can pipe directly into light guides: you shouldn't need a physical coolant loop, nor be limited to solid-tungsten temperatures. Still wildly optimistic, but I can vaguely picture the engineering.

    More generally, I think the plausibility of this type of stealth depends on the relative progress in drive and power/cooling technology. If (as I've assumed) drives stay the same while power/cooling technology gets much, much better, then you can afford to be stealthy by beaming your emissions. If power/cooling doesn't progress much, but ion drives are developed with exhaust velocities of ~1000km/s, then radiators become the limiting factor, and you can't afford to restrict your beaming angle because that makes your radiators less efficient. In this second scenario everyone travels on brachistocrone trajectories, with their radiators at full power the whole way.

  19. January 27, 2020bean said...

    I can see how that's not completely impossible from a thermodynamic sense, but the engineering seems really dubious. Every single imperfection leads to heat in the system, which you then have to spend more energy pumping out, and I strongly suspect that ends in a Red Queen's Race. I'd be a lot less skeptical of a scheme to get it down to 120-150K, which is still going to reduce detectability by a factor of 16 relative to a completely passive object of the same size.

  20. January 27, 2020DampOctopus said...

    I'm not so concerned about keeping the exterior cool. The James Webb Sunshield will maintain temperatures of 50K and ~300K on either side, purely passively, despite pumping power onto the cool side to run instrumentation. That's with a five-layer foil shield that needs to survive orbital launch and unfolding. If you're building in free fall, and will never need to stand acceleration over a microgee, you could make insulation with far more, and thinner, layers. It'll be bulky, but weigh next to nothing.

    I'm more concerned with heat transfer in the powerplant: absorption in the walls of the reactor chamber, or resistive losses in the power circuitry. Coping with those might turn out to be impossible for practical purposes. And you don't make them significantly easier by allowing the exterior to warm up to 120-150K.

  21. January 27, 2020bean said...

    The exterior issue is solar radiation, not heat leakage from inside. The JWST sunshield works because it's able to get rid of all the heat via passive radiation on the hot side, which is not an option the STWYG has. Insulation can help, but it's not going to remove that particular problem.

  22. January 27, 2020DampOctopus said...

    Oh, that was the issue for which I suggested an angled mirror as a sun shield. That'll block >99.9% of solar radiation, so the rest of the hull exterior can get down to ~50K like the James Webb even without active cooling. The downside is that you're reflecting sunlight, but that reflection is a cone with the same opening angle as the apparent size of the sun: 0.5deg at 1AU. If you turn to the right angle, you can put that reflection inside your radiator-emission cone.

    (We can tell this thread has grown too long when it starts needing an index.)

  23. January 27, 2020Suvorov said...

    Every single imperfection leads to heat in the system, which you then have to spend more energy pumping out, and I strongly suspect that ends in a Red Queen’s Race.

    Convert the heat into electricity!

    "So, the bad news is that we've made 83% of the ship's mass into thermoelectric generators, but the good news is that we can power the entire ship's life support systems for for six months with a gallon of kerosene and the body heat of two average adults!"

    On a less ridiculous note, I think it's interesting we've focused around using the stealth spacecraft as offensive weapons. But I think the STWYG has a promising career as a defensive vessel, much like how small countries use their SSKs. It makes me wonder how difficult and expensive the techniques would be – the expensive part of stealth aircraft is getting the shaping and coating right, but since spacecraft engineering of the STWYG is probably taking place in orbit and spacecraft already have to be engineered to reasonably fine specifications, it makes me wonder if building a stealthy spacecraft designed to control friendly space might be a cost efficient decision. You skimp on delta-V since you're not expecting to patrol far from your home posting, and you embark on long "deterrence patrols" to keep hostile powers from parking themselves in your planet's orbit and beginning orbital bombardment.

    Depending on where your planet was, you'd need to worry a lot less about solar radiation, too – the outer solar system just isn't as "warm" a place as the inner solar system. You might could even keep your ships in the shade of a gas giant (if they were launched from one of the moons.)

    Finally, I wonder how easy it might be to mess with the optical systems mentioned using decoys. Bean's mentioned previously that it's usually going to be possible to determine mass from a drive plume, but if your enemy was searching for ion trails and particle interactions, or faint blackbody radiation, they might be easier to decoy. Some sort of drone running a "dirty" ion engine designed to cause plenty of particle interaction and give off a low level of radiation might be cheap enough to send enemy super-computers on a wild goose chase, saturate their high-power sensors, give them false assumptions about your stealth performance, etc. etc. (I'm not sure how plausible such a "fake" ion trail would be, though. Could you still determine ship mass or nah?)

    Ironically, that suggests that decoys might work better for stealth ships which try to disguise their plumes than for regular warships that can only be imitated by ships using the same drive systems and in the same mass class.

  24. January 28, 2020DampOctopus said...

    I made a mistake above: mirrors with >99.9% reflectivity exist, but only over narrow bands. For a broadband source (e.g. sunlight) you're probably only going to get ~98% reflectivity with e.g. silver.

    @Suvorov

    I don't think decoy drive plumes work: the amount of recombination radiation depends directly on the exhaust mass, and you need the same exhaust velocity to get the redshift right, so your decoy has to have the same thrust and power as the real thing.

    I like the idea of stealthy defenses, and the analogy with SSKs. You might make them automated platforms idling on just enough power to detect a wake-up signal. You could even skip the drive entirely, making them more like mines, and use an SSK-equivalent to deploy them without giving away their locations.

  25. January 28, 2020bean said...

    Stealthy defenses definitely make more sense than stealthy attackers. If you have time, then it's a lot easier to keep your signature down, and you can hide in the clutter of an active orbit. You really don't even need all the fancy cloaking stuff. Just disguise them as junk, pushed into some far-out orbit. Although there's the issue of them being able to see you when they attack. Nukes are the only way around that particular problem, and I have doubts about practicality. But kinetics coming in from multiple directions, even with a few minutes warning, could be a serious problem, depending on the exact tech balance.

  26. January 28, 2020Suvorov said...

    I don’t think decoy drive plumes work: the amount of recombination radiation depends directly on the exhaust mass, and you need the same exhaust velocity to get the redshift right, so your decoy has to have the same thrust and power as the real thing.

    Hmm. In theory you could increase recombination by increasing the density of the medium: pumping gas out the "front" of your decoy and then accelerating through it. Or am I missing something?

    You really don’t even need all the fancy cloaking stuff. Just disguise them as junk, pushed into some far-out orbit.

    "Naval mines in space" would also be a lot easier to make decoys for. If the "other side" has to shoot every tiny object floating in orbit, they're going to be burning through their ammo and heat sinks pretty rapidly – although again the tech balance will matter.

    But kinetics coming in from multiple directions, even with a few minutes warning, could be a serious problem, depending on the exact tech balance.

    Even if they dodge them, that's fuel they don't have to get home.

  27. January 28, 2020bean said...

    @Suvorov

    Convert the heat into electricity!

    Not sure if this is sarcastic or not, but in case it isn't, it doesn't work that way. You can only turn heat gradients into electricity, not heat itself. And that means getting rid of the waste heat.

  28. January 28, 2020Suvorov said...

    Not sure if this is sarcastic or not, but in case it isn’t, it doesn’t work that way. You can only turn heat gradients into electricity, not heat itself. And that means getting rid of the waste heat.

    I mean, I'm not an expert on the subject, but "converting heat into electricity" is how Nasa explains their RTG power sources. My understanding is that the heat gradient is the means by which you tap into the heat energy to convert it to work.

    My understanding is that, since no system is going to be 100% efficient, you're always going to have some waste heat, but you can harness (some of) that heat to do useful stuff (like work) instead of heating your spacecraft's hull.

    Hence the absurd imagery of a ship full of layers of waste heat recyclers.

  29. January 28, 2020bean said...

    Thermocouples may look more like "turning heat into electricity" than, say, a steam engine, but they still obey the same laws. Basically, the maximum efficiency of any system to turn heat into work (electrical or otherwise) is equal to the ratio of the temperatures of the hot side and the cold side. The problem on a spaceship is that you still need to get rid of the excess heat, and while you may have less heat to get rid of after running it through a thermocouple, it's also at a colder temperature, which means you need a bigger radiator. And because power/area (and thus power/mass) scales with the fourth power of temperature, this rapidly becomes a game you can't win. The math comes out to suggest that minimum radiator area comes when the cold end is 75% of the temperature of the hot end. If I go to 50% temperature, I only have a third as much heat to get rid of (because efficiency goes from 25% to 50%), but I still need about 70% more radiator area.

  30. January 28, 2020Suvorov said...

    The problem on a spaceship is that you still need to get rid of the excess heat

    Right, I was leaning into your idea of a Red Queen’s Race – in this case increasingly absurd amounts of thermocouples to reduce the temperature.

    and while you may have less heat to get rid of after running it through a thermocouple, it’s also at a colder temperature, which means you need a bigger radiator.

    Ooohh, that's an interesting point.

  31. January 28, 2020DampOctopus said...

    @suvorov

    Let me go into a bit more detail here on the thermodynamics. When you have a heat gradient, you can allow heat to flow from the hot end to the cold end, and tap off a fraction of the heat energy to do useful work. RTGs use the heat gradient between a radioactive source and a radiator, steam engines use the heat gradient between their boiler and their surroundings, etc.

    Conversely, when you have a heat gradient, you can do work to pump heat from the cold end to the hot end. (The energy of the work gets added as extra heat on the hot end.) Peltier cooling does this with electricity, refrigerators use compressors in a physical coolant loop, etc.

    Those are your options: let heat flow down a gradient, and get work out; or push heat up a gradient, and put work in.

    Now say that the STWYG has a hull temperature of 100K, and you want to make it colder. To stay stealthy, you want the heat to go to the radiator. But the radiator is at 3695K, which is much hotter than the hull. To move heat from the hull to the radiator - from the cold end of the gradient to the hot end - you have to do work. You can't get work by doing that, regardless of what amount of mass you dedicate to thermoelectric generators.

    Worse, to do the work to pump heat from the hull to the radiator, you need to run the reactor harder. That generates more heat, some of which leaks to the hull, so you then need to do even more work to pump that extra heat to the radiator. That's the Red Queen's Race that bean is talking about.

  32. January 28, 2020Suvorov said...

    Now say that the STWYG has a hull temperature of 100K, and you want to make it colder. To stay stealthy, you want the heat to go to the radiator. But the radiator is at 3695K, which is much hotter than the hull. To move heat from the hull to the radiator - from the cold end of the gradient to the hot end - you have to do work. You can’t get work by doing that, regardless of what amount of mass you dedicate to thermoelectric generators.

    If one opts out of doing work to move the heat to a radiator, the skin of the ship ends up being the radiator by default, right? (Which cooks the crew, overheats the reactor, and ruins our stealth.)

    I assume RTGs just let the heat radiate off the body, but they're generating a modest amount of heat compared to a fission reactor. Or do they pump it around too?

    That’s the Red Queen’s Race that bean is talking about.

    Presumably one could run a different Red Queen's Race of turning an entire spacecraft into a series of heat gradients to cool the hull and watch in horror as the ship's size and mass ballooned in service of ever-more-inefficient waste heat recyclers (which is what my joke earlier was suggesting.)

    On a somewhat unrelated note concerning technology assumptions – how much would superconductors change things? I assume they would make both railguns and lasers more practical, and they'd presumably cut down on waste heat and power generation at the margins, but would they have any major impact on spacecraft design besides that?

  33. January 28, 2020Lambert said...

    Lifehack:
    Once you're this deep into discussion about temperature gradients, you ought to just treat the ship as a black box and calculate energy and entropy fluxes.

  34. January 28, 2020Doctorpat said...

    From a fictional point of view, a stealth spaceship has one huge advantage: You can't be in radio contact with the home planet, so you are forced to have at least one human on board, authorized to make "This is an act of war" type decisions.

    Unless you are prepared to let your computers launch an attack on the pan-galactic empire just because they didn't have their IFF turned to the right frequency. Even though they had their armada flying in a formation that clearly spelled out, in English, Spanish and C++, their peaceful intentions.

    In reality we may well accept a sufficiently advanced computer having that responsibility, but it's certainly a strong enough argument to justify the narrative of needing (brave, sexy, slightly rebellious against authority) heroes out there controlling the ships.

    Whereas having a naval officer with that responsibility is a practice going back centuries, especially before radio.

  35. January 28, 2020Suvorov said...

    Unless you are prepared to let your computers launch an attack on the pan-galactic empire just because they didn’t have their IFF turned to the right frequency. Even though they had their armada flying in a formation that clearly spelled out, in English, Spanish and C++, their peaceful intentions.

    Imagine explaining why your navy blew up half-a-dozen freighters because of a software database error. Or, worse yet, explaining that the aliens had destroyed half your navy because their ship types were not yet in your vessel's databases because, you know, first contact and all.

    In reality we may well accept a sufficiently advanced computer having that responsibility, but it’s certainly a strong enough argument to justify the narrative of needing (brave, sexy, slightly rebellious against authority) heroes out there controlling the ships.

    Plus, what a perfect setting for drama. Your ship can't run, can barely fight, you're inside all the time, you're extremely cramped, you crew is probably small (I'm guessing automation would probably reduce the crew to 4 - 8) and if anything breaks you're either slowly cooking to death or dehydrating.

  36. January 29, 2020quanticle said...

    I can't believe we've gone this far without anyone linking gwern's analysis of space warfare, so I'm rectifying that oversight now.

    While gwern's analysis is geared towards interstellar conflict, I find that much of it could be reapplied to conflict within the solar system. Namely, what prevents any power from, say, going out to the Oort Cloud, breaking off a chunk of comet, and dropping it into the inner system? By the time it got to Earth or Mars, it would have the kinetic energy of a decent-sized nuclear weapon. The Kuiper Belt and Oort Cloud are far enough away that they would be difficult to monitor and the mass of the cometary or asteroid fragment could very easily suffice to shield the drive plume output of whatever method you use to alter the orbit of said fragment.

  37. January 30, 2020Suvorov said...

    I can’t believe we’ve gone this far without anyone linking gwern’s analysis of space warfare, so I’m rectifying that oversight now.

    This was a good read, thanks for posting it.

    Namely, what prevents any power from, say, going out to the Oort Cloud, breaking off a chunk of comet, and dropping it into the inner system?

    I'm partial to the thought that by the time this is easy to do, highly-populated places (so, Earth) will have interceptors in place to deal with this specific threat. Comets, I am given to believe, are largely water ice; a few nukes should do the trick. But that's not much consolation to, say, the Martian colony.

    I do think, though, that this technique in general is over-anticipated. Of course the physics of space warfare suggest dropping rocks on everyone's cities, but the physics of modern warfare dictate nuking everyone, and nobody does it – and not just because they are worried about mutually assured destruction, either. Under what situation do you want to drop an asteroid on a world? It lets you "win a war" if you drop enough to kill everyone on the opposing side, but nobody is going to start a war if you can do that unilaterally, and if they can return the favor, no one is going to do it except as a last resort, and presumably they'll tell you when you're crossing their red lines. Generally speaking, it's not going to be helpful to the war effort (war production will be in space most likely.)

    [Gwern's wrong, by the by, about the impossibility of staging a false-flag nuclear strike - the U.S. could probably pull it off.]

    I think you mostly have to worry about 1) terrorists; 2) aliens [Dark Forest scenario] or 3) if you're on a militarized/war industry habitat or location.

    TLDR; like the old fears of nuclear exchange leading to complete destruction of the human race, I think a similar threat from dropping asteroids isn't especially realistic.

  38. January 30, 2020bean said...

    Rocks are not free, citizen!

    Even if you want to bombard the opponent's planet (not a good assumption), dropping rocks from the Kuiper Belt or Oort Cloud is probably not the best way to do it.

    1. Moving hundreds of thousands or millions of tons of rock isn't easy. And it's a lot harder to put said rock on a collision course than it is to divert it.

    2. It's going to take time to boost the rock, then a lot more time for it to impact. Any method of winning a war that requires years or decades to work is a bad plan if the war can be won by more conventional means in a shorter period of time.

    3. Doing this in a stealthy manner is essentially impossible. More power makes detection easier, and you're going to need a lot of power to move the rock in question.

    I could do enough damage a lot faster with one ship and a bunch of nukes. And that also translates to cheaper. Rocks are not and will never be free.

  39. January 31, 2020bean said...

    I've come down sick, so the RTW2 post will be delayed. Hopefully I'll have it up tomorrow.

  40. January 31, 2020ADifferentAnonymous said...

    Presumably one could run a different Red Queen’s Race of turning an entire spacecraft into a series of heat gradients to cool the hull and watch in horror as the ship’s size and mass ballooned in service of ever-more-inefficient waste heat recyclers (which is what my joke earlier was suggesting.)

    This could make sense as a way of getting the most out of a valuable, very strong power source--a Matrioshka brain is basically this, with a star as the power source (though you wouldn't call one a "ship")

  41. January 31, 2020Lambert said...

    Sounds like a triple expansion engine with extra steps.
    The idea of getting work out of a series of increasingly large, increasingly cool fluids is by no means new.

    Modern steam and gas turbines often have multiple expansion stages, and combined-cycle gas turbines boil water to get a second bite at the thermodynamic cherry.

  42. February 01, 2020bean said...

    There were some civilian ships that had quadruple-expansion engines, although it was more common to have a triple-expansion engine with two low-pressure cylinders. And most steam ships had high-pressure and low-pressure turbines. Occasionally, you'd see a medium-pressure turbine. I think Queen Mary had those. And gas turbines often have multiple spools.

  43. March 11, 2020AlexT said...

    Hi, great blog!

    Re stealth in space: an interesting, if a little math-heavy, discussion can be found at

    https://groups.google.com/forum/#!msg/rec.arts.sf.science/FjEbx0j8zuU/LrRorgB8uioJ

    the bottom line being that exhaust plumes in space are very difficult to detect, and thermal rockets which run on hydrogen are a good way to both cool the ship and accelerate while maintaining very low observability.

    Combine with a reflective sun shield, shut down non-essential systems while coasting and leak hydrogen slowly to get rid of any remaining waste heat and a ship can move interplanetary distances unobserved.

  44. March 11, 2020bean said...

    I believe that’s the same one we discussed at some length in Escorts, where the topic first came up. I’m skeptical on several fronts. Expanding that much seems difficult to impossible, and the delta-V numbers were worked on Hohmann transfers, which really don’t apply here. I can’t shake the lingering fear of a thermodynamics violation, and my intuition on that is generally pretty good, although I have other things to spend my time on (this blog) that are going to keep me from running it down properly.

    I also enjoyed the questions over John Schilling’s fate. And the fact that he’s alive and well, and commenting here.

  45. March 11, 2020AlexT said...

    @bean

    "I believe that’s the same one we discussed at some length.."

    You might be thinking about the proposed solar thermal rocket from toughsf. That's not what I was referring to, although they do have in common the idea of open-cycle cooling while coasting.

    The interesting point concerns plume detectability. Or rather, its undetectability. Stealth detractors dismiss any notion to camouflage a ship under thrust, and assume the plume acts as a giant infrared flashlight. As it turns out, this is very far from accurate. At least for hydrogen thermal rockets, the exhaust radiates next to nothing and is very difficult to spot.

    This doesn't "break thermo" since entropy still increases, irretrievably. But that giant cloud of near-vacuum density hydrogen isn't radiating worth a damn.

  46. March 11, 2020bean said...

    Expanding the hydrogen may not be a thermo violation, but I doubt you could get an invisible engine without one. The radiation may not be coming from the exhaust plume, but you can't hide the engine forever.

  47. March 12, 2020AlexT said...

    Expanding the hydrogen may not be a thermo violation, but I doubt you could get an invisible engine without one.

    The big problem would be ensuring that the fuel flow through the engine is sufficient to pick up all the heat the engine is generating. Looking at public test data for NERVA, it seems to me the kinetic energy of the propellant (based on mass flow and exhaust velocity) matches pretty well the reactor's thermal power.

    https://en.wikipedia.org/wiki/NERVA#Reactortestsummary

    The longest running test has flow rate of 32.8 kg/s, exhaust velocity 8.25 km/s, giving kinetic energy output about 1116 MW, with the reactor's power 1137 MW thermal. This would mean it's being cooled regeneratively during thrust. So I guess it's doable.

    Of course, this concerns hiding the engine while thrusting. During coast, different considerations apply. Directional radiators, a good reflective sun shade, leaking a little lukewarm hydrogen etc, coupled with tightly controlled waste heat generation, can ensure the ship radiates very close to the cosmic microwave background in almost all directions.

  48. March 12, 2020bean said...

    Sure, most of the energy will go into the propellant. But most and all are very different matters when trying to do stealth, and even a few percent being radiated will mess you up big time. And I suspect that getting all of the energy into the propellant is at the very least an extremely difficult engineering challenge, maybe impossible. After all, heat flows from hot to cold, so any place where the engine is cooler than the surrounding hydrogen will need to itself be cooled by more hydrogen. This doesn't end well.

  49. March 13, 2020AlexT said...

    any place where the engine is cooler than the surrounding hydrogen will need to itself be cooled by more hydrogen

    The operating principle would be that cryogenic hydrogen cools the nozzle, gets warmer, then it cools the reactor core walls, gets even warmer, then the equipment and the living quarters, by now it's at say 300K, now it cools the high-temperature equipment and the rest of the engine, then into the reactor core where it absorbs most of the heat, gets hot as hell and out the tailpipe it goes.

    Probably the devil to design the plumbing and complicated to keep working. Worth it, though.

  50. March 13, 2020bean said...

    Oh, I absolutely get the theory, but I'm much less sure about the practice. The big problem is that LH2 isn't that good of a coolant, so you'll need to run a lot through the nozzle to keep it down to cryogenic temperatures, probably more than you're sending out the back end. Regenerative coolant in rockets can warm up from tank temperature to near engine temperature, but you've got maybe 200K to play with here. Insulation can help if you use it in short bursts, but the heat has to go somewhere in the long run.

  51. March 13, 2020AlexT said...

    LH2 isn’t that good of a coolant, so you’ll need to run a lot through the nozzle to keep it down to cryogenic temperatures

    That depends on how cold the exhaust gets, and how quickly, as it expands, as well as how much it heats up the nozzle as it passes through. Also, on how cold the nozzle needs to stay, as per mission requirements. I'm arguing that it's theoretically possible. The rest, as they say, is engineering.

    you’ve got maybe 200K to play with here

    I assume that means the max temperature allowed for the nozzle above the hydrogen's storage temperature? Allowing that, I'd cool the nozzle by allowing the LH2 to flash to gas (maybe cooling the outer skin of the ship) then send that gas, at constant pressure, through the nozzle, to get the best heat capacity out of it. And, of course, build the nozzle such that that's enough. As easy as that.

  52. March 13, 2020bean said...

    I assume that means the max temperature allowed for the nozzle above the hydrogen’s storage temperature?

    No, I'm talking about the exit temperature of the hydrogen from the nozzle cooling system. The warmest part of the nozzle will be warmer than that, although maybe not by much. The problem is that you're now looking at parts of the nozzle radiating at something like 200K, which is a lot more than you can afford. Although now that I think about it more, you might be able to have the hydrogen make multiple passes, reducing the radiation temperature. That's going to have major size and weight impacts on the engine.

    I’m arguing that it’s theoretically possible. The rest, as they say, is engineering.

    Speaking as an aerospace engineer, "the rest is engineering" can cover a lot of extremely difficult work. And there are plenty of things which are theoretically possible which fail on engineering grounds. I need to poke around more on stuff like nozzle heat transfer properties before I can say which camp this falls in.

  53. March 14, 2020AlexT said...

    parts of the nozzle radiating at something like 200K, which is a lot more than you can afford

    Say 200K on average. Correct me if I'm wrong: a 200K blackbody radiates about 90 W/m^2. For a 6m nozzle, so an exit area of about 30 m^2, that means 2.7kW. Good luck spotting that at 1AU.

    Also, to eat up that 2.7kW (assuming 5 J/g/K heat capacity for LH2) would take what, 500g per second warmed up by 10 degrees? That, out of a total flow on the order of tens of kg per second going from cryo to several thousand degrees. So, if I can keep the nozzle at 200K, I might as well keep it at 30K and let that 30 m^2 * 50 mW/m^2 = 1.5 Watts shine for all the Universe to see.

    “the rest is engineering” can cover a lot of extremely difficult work

    What I meant was "the theory sounds doable, the practicalities will be interesting" but it came out wrong. Sorry.

  54. March 14, 2020bean said...

    Hmm. This isn't as implausible as I thought. I'm still skeptical about how much all of this will weigh, though. And about performance in general. Your stealth ship is of little use if the war is over by the time it gets there, and in an era of nuke-electric, nuclear-thermal just isn't going to do faster interplanetary trajectories.

    Also, I'm now going to move the goalposts, and wonder about star occlusion. (This has been building for a while, and I just felt like doing it this morning.) Let's take Kepler-438b as the basis for our math. It has a radius 1.12 times that of Earth, and orbits a star with 0.52 the radius of the sun, 640 light-years away. So it covered 0.03% of the star during transit. At 1 AU, an object 9 meters across is the same apparent size. Yes, obviously it's going to be moving a lot faster, so you won't get the same signature and your detection will be worse, but with things like TESS available today, I don't think it's at all outrageous to assume that this could be a major means of detection if stealth works reasonably well. Keppler-438 is of magnitude 15, and a quick approximation gives 117,777 stars of that magnitude throughout the sky. This looks to cover about 1e-21 of the sky, at least if we take Keppler-438 as our baseline for star size/distance. Not sure what that translates to in terms of detection on the ground, but I suspect that you'd benefit because the system ecliptic lies close to the galactic ecliptic, increasing the density of stars in the areas you're likely to care about, particularly if the other guy is stuck with comparatively wimpy hydrogen NTRs.

  55. March 14, 2020AlexT said...

    in an era of nuke-electric, nuclear-thermal just isn’t going to do faster interplanetary trajectories

    I would argue that, in the era of stealth thermal ships, using nuke-electric is suicidal. It may get to some places faster, although I'm not sure exactly how far it has to be before a good kick-and-coast stops being competitive. But that only means it arrives faster to its death, if enemy stealth ships are waiting.

    So, not something for the battle fleet. A cruiser/raider might be a good fit, to chase civilian ships and be a nuisance. Still, it would have no way of knowing when a missile will show up out of nowhere and terminate its career. And that missile can be launched from very far away, since a ship on electric thrust can't be too unpredictable.

    Assuming electric propulsion is practical, I'd go with a mothership-rider setup, with a large electric high-isp ship that deploys small, stealthy combatants with just enough delta-v to fight a battle.

    star occlusion

    I'm not sure that enough of the sky is bright enough to spot things against it. Take a trajectory covering say a 90 degree arc of the sky from an observer's view point. How many stars, on average, would it occlude, even over the densest parts of the sky?

    Also, how would the detector tell a stealth ship from an asteroid? How many asteroids cause occlusions, randomly, throughout the sky at any given time?

    Thirdly, one occlusion would provide bearing, but no range, unless another significantly distant detector happened to get an occlusion at the same time on the same ship but a different star, which sounds unlikely. Or, the same tracker would need multiple occlusions, preferably not too far apart, to make a guess about the object's trajectory. Is that likely?

  56. March 14, 2020bean said...

    This is sounding like "submarines will render all surface ships obsolete". I suspect similar drivers will be in play here. On a strategic level (planet to planet) you're going to be much slower. On a tactical level, detection is much easier because you know where they're starting from, particularly if they're parasites coming from a carrier. There are lots of ways to thwart stealth, particularly at close range.

    I don't know exactly how many occlusions you'd get over a typical trajectory. I'll look into it more at some point, but right now, it's firmly in the "maybe" category.

    Also, how would the detector tell a stealth ship from an asteroid? How many asteroids cause occlusions, randomly, throughout the sky at any given time?

    That's easy enough. Pretty soon, you'll know where the asteroids are, and be able to predict their transits.

    Thirdly, one occlusion would provide bearing, but no range, unless another significantly distant detector happened to get an occlusion at the same time on the same ship but a different star, which sounds unlikely

    A couple of aspects. First, couple the broad-area surveillance telescope to a more powerful one that gives you a lot more stars after you get a hit. Second, you're going to have some size and angular velocity data from the initial fix. Not direction, but you'll know how much the star's power dips, and how long the dip lasts.

  57. March 15, 2020AlexT said...

    There are lots of ways to thwart stealth, particularly at close range.

    Certainly. It reverts to the age-old dance of counter- and counter-counter- and so on.

    Pretty soon, you’ll know where the asteroids are, and be able to predict their transits.

    That sounds a bit iffy. Not only are there immense numbers of 50m-or-less rocks, they are on all sorts of orbits of various eccentricities. Is that blip an enemy ship, or a TNO that made a flyby of Jupiter 2 years ago and is now en route to Sirius?

    how many occlusions you’d get over a typical trajectory

    Assume a 90 degree circle at 1 AU, say 150 million km. So, long-range. Length of the circle - say 22510^9m. Area of occultation for a 50m-wide object ~ 112510^10 m^2. Suppose we're looking 15 degrees off the ecliptic. At 1AU, the background surface is 2pi150Gm (length of the circle) multiplied by 2sin(15deg)150Gm (height), so a grand total of 73513 Gm^2, say 73*10^21 m^2. The trajectory would black out ( 112.5 / 73 ) * 10^-10. Say on the order of 2/10^10. Allow 2M stars in the detector's catalog, spread uniformly over those 15 degrees off the ecliptic. The odds of the trajectory occluding one of those stars is about 1:10000. I think I've made pretty conservative assumptions, especially concerning the number of stars. Unless I've forgotten something, I'd bet on the space sub.

  58. March 15, 2020Doctorpat said...

    I think we've ended up with the "space sub" style stealth spaceship now being plausible enough for fictional purposes, even if there are still wrinkles to iron out before actually starting to build one.

    I think that it could result in some very interesting hard SF, with the bonus that you can mine a literal century of real and fictional sub warfare accounts for ideas, scenarios, and general inspiration.

  59. March 15, 2020bean said...

    That sounds a bit iffy. Not only are there immense numbers of 50m-or-less rocks, they are on all sorts of orbits of various eccentricities. Is that blip an enemy ship, or a TNO that made a flyby of Jupiter 2 years ago and is now en route to Sirius?

    A TNO is going to have a much lower angular velocity, not to mention that a 50m wide TNO isn't going to look like a 50m wide ship because of how much further away it is.

    I can't find a big enough problem with your math on the stars. I do think it's somewhat low, as the stars might be bigger than a 50m object at 1 AU, but not enough to change the math a lot. Of course, if it gets closer, the odds change significantly.

  60. March 15, 2020AlexT said...

    Just realized I need to get better at writing math with markdown. Sorry for that mess of a post, glad it made some sense even as poorly formatted as it turned out.

    I agree that occlusion does seem a likely passive detection method, with its limitations.

    But the whole tansis thing is just too inaccurate for the wide, unqualified acceptance it receives.

  61. March 15, 2020Kyzentun said...

    detection via star occlusion

    I don't see this as practical.

    Star power varies for many reasons already. Starspots, exoplanets, companion stars, whatever drives luminous blue variables to change so much, and so forth. Some stars are believed to expel a large mass that is opaque, obscuring their visible radiation. (wikipedia article on luminous blue variables is quite fun)

    Exoplanets are detected by a repeated pattern of the same dimming. Single events are not enough to detect an exoplanet until they repeat enough to form a pattern.

    Occlusion by a craft is going to be a single event for any given star. So for any star it crosses, the database would need to have an extremely detailed model of all the things that regularly affect that star. Light curve resulting from each type of event, compared to light curve from transit of a given object at a given distance and size.

    Gut feeling: A database with that level of detail for even 10% of the sky is beyond current astronomy by 100-1000x.

  62. March 15, 2020bean said...

    Exoplanets are detected by a repeated pattern of the same dimming. Single events are not enough to detect an exoplanet until they repeat enough to form a pattern.

    I'm not sure this is true. TESS is only a 2-year mission, and it doesn't watch the whole sky at once. I'd guess they're doing more complicated discrimination based on how the signature changes, which would work for us here. The timescale of the changes is probably seconds, and I suspect that everything except a few pulsars is stable on that timescale.

  63. March 15, 2020John Schilling said...

    I've had time to catch up on this now, and first off: thanks for bringing that old rec.arts.sf.science discussion to my attention. I had left the sf newsgroups at about the time of Racefail, and never looked back, but there was apparently still some good stuff going on as late as 2011. It might have been worth a return visit for that discussion.

    I think my old analysis still holds up pretty good. The original poster seems to be hinging his argument on my figure of 95% efficiency being too low, but that was only a handwavy estimate for thrusters of SFnal interest in general, from old LOX-kerosene rockets to fusion torch drives. I was trying to show that the math is so very much against stealth in space that it doesn't much matter what number you put there, 90%, 95%, 99%, whatever, and apparently I didn't make that clear enough.

    But, in trying to claim that the proper number should be (for radiant emissions, at least) 99.994%, the OP cites a NASA paper that is only concerned with near-field downstream plume emissions. Rightly so for their application, but even that paper admits that radiant emissions from the flow upstream of the nozzle exit plane will be more than a factor of twenty higher, and blackbody radiation from the nozzle interior will probably be higher still. All of which will be visible across most of the rear hemisphere. NASA doesn't care, because NASA isn't trying to hide their spaceship, just keep it from melting its own solar panels.

    No, you can't just keep the engines always pointed away from the enemy, because A: that's a huge strategic tactical constraint and B: the enemy isn't in just one place, any peer competitor will have seeded the environment with remote sensor platforms - and if you're entertaining the possibility that stealth can be made to work at all, those sensor platforms will certainly have it. Also, as shows up later in the usenet discussion, far-field plume radiation is also significant.

    I was also amused by the bit where the OP correctly notes that plumes may be nigh-invisible in the most literal sense because most of the radiant emission will be in the IR rather than the visible, at the same time he is arguing that I can't possibly be right because the plume from the Saturn V wasn't dazzlingly, impossibly bright to visible-light cameras.

    The suggestion here that we should be considering nuclear-thermal rockets with hydrogen as a reaction mass, is a step in the right direction if you're trying to go stealthy. Molecular hydrogen has much weaker emission lines than CO2 or H2O. But it's still a diatomic molecule with rotational and vibrational bands that emit in the infrared. And naive calculations based on ideal nozzle expansion to an emission of nice, cold hydrogen are going to miss quite a bit due to frozen-flow considerations (some of the hydrogen molecules will be excited to toasty levels in the chamber and simply not have time to de-excite before they leave the engine), and by frictional heating of the boundary-level flow. Note that these trade against each other; a big long nozzle to let the core flow reach equilibrium, will have correspondingly high boundary-layer heating.

    AlexT cites a wikipedia article giving a specific impluse of 841 seconds, suggesting 98.2% efficiency, for NERVA XE-Prime. But as usual with wikipedia, you want to go one reference further. The actual report gives the specific impulse as 841+ seconds, and thus the efficiency as something greater than 98%, but that's for ideal vacuum specific impulse. I couldn't find the details of how they calculated that, but it's perilously close to saying "Assuming 100% efficiency, the efficiency will be about 100%".

    There are for obvious reasons no measured vacuum Isp values for NERVA, and very few good sea-level measurements. But the same report gives the nominal expected performance for a couple of engine designs, and the thermodynamic efficiency is at best 84%. Not all of the remaining 16% is going to be radiated in a detectable way, because hydrogen, but it's likely to be significant.

    For better results, you could use helium. No rotational or vibrational energy states there, and the first electric state is high enough that there will be little excitation at NERVA temperatures. When I poke around with plausible stealthy-spaceship concepts, that's where I start looking. But liquid helium is a mediocre propellant, a lousy coolant, and a bitch to store in quantity for long periods.

    When I started discussing this on usenet many years ago, the analogy I made was to WWI-era submarines - technically feasible, capable of against an unprepared enemy, but not going to dominate force-on-force naval combat. And, once the enemy widely deploys the appropriate detection platforms, very limited even in niche applications like commerce-raiding. It is perhaps unfortunate that this was so quickly simplified to "There Ain't No Stealth in Space", but that's still closer to true than not.

    I really do need to write up a decent primer on space warfare before I get sucked so deep into the black world that I can't safely comment on it at all.

  64. March 15, 2020bean said...

    Ah. Thanks for that. I was suspecting a trap, but didn't quit have the time and skills to go looking for it.

    I really do need to write up a decent primer on space warfare before I get sucked so deep into the black world that I can’t safely comment on it at all.

    If you need a co-author who's not in the black world, I'd be happy to help.

  65. March 15, 2020AlexT said...

    @John Schilling

    The crux of the matter is how detectable are hydrogen plumes. I've yet to find an actual statement that "this big a plume with this much hydrogen at this temperature has this shape and radiates this many watts".

    Since it's the only non-chemical thermal rocket engine test data I know of, I'm going to keep looking at nerva test results. The reactor's thermal output almost equals the kinetic energy of the exhaust in the direction of thrust. The rest could be engine heat, exhaust divergence, nozzle radiation, or radiation out of the exhaust itself. But how much of each?

    a big long nozzle to let the core flow reach equilibrium, will have correspondingly high boundary-layer heating

    Sure, but the nozzle can be cooled regeneratively. The only part that can't be fixed is the damn plume.

    I was also amused by the bit where the OP correctly notes that plumes may be nigh-invisible in the most literal sense because most of the radiant emission will be in the IR rather than the visible, at the same time he is arguing that I can’t possibly be right because the plume from the Saturn V wasn’t dazzlingly, impossibly bright to visible-light cameras.

    I think his point was that the cameras didn't melt.

  66. March 15, 2020John Schilling said...

    "Since it’s the only non-chemical thermal rocket engine test data I know of, I’m going to keep looking at nerva test results. The reactor’s thermal output almost equals the kinetic energy of the exhaust in the direction of thrust."

    Where are you getting the kinetic energy of the exhaust? Again, the numbers in the wikipedia page are the hypothetical values for the ideal case. Assuming 100% efficiency, the efficiency is greater than 98%. The NASA report behind the wikipedia page, when it makes predicts for actual flight engines, gives kinetic energy of no more than 84% of the reactor's thermal output.

    "I think his point was that the cameras didn’t melt"

    The cameras were specifically behind quartz windows to protect them from the heat. Assuming 5% of the total power of five J-2 engines manifests as uniform radiant heat, and eyeballing the cameras as ~7 meters from the center of the downstream plume, that's 533 MW total radiated power, 866 kW/m^2 at seven meters, and a radiant equilibrium temperature of 1,703 deg C - just below the melting point of quartz. But it will really be quite a bit less than that, because the radiation won't be uniform. As noted above, most of it is only visible from below the nozzle exit plane, and the cameras are well upstream of that.

  67. March 15, 2020John Schilling said...

    "If you need a co-author who’s not in the black world, I’d be happy to help."

    I may take you up on that.

  68. March 15, 2020AlexT said...

    the numbers in the wikipedia page are the hypothetical values for the ideal case

    ..and now that I looked at the report, I agree and won't mention them any more.

    radiant emissions from the flow upstream of the nozzle exit plane will be more than a factor of twenty higher, and blackbody radiation from the nozzle interior will probably be higher still

    These, as well as the angle of observability, can both be mitigated by a larger, better cooled nozzle. Difficult and awkward, but worth it. Can't say more without numbers and test data.

    far-field plume radiation is also significant

    Not very much, according to that discussion, especially if it's hydrogen. Numbers are needed once more.

Comments from SlateStarCodex:

Leave a comment

All comments are reviewed before being displayed.


Name (required):


E-mail (required, will not be published):

Website:

You can use Markdown in comments!


Enter value: Captcha