Friday 20 January 2017

Expansion-cooled, Curved-nozzle stealth rocket

Here, we will look at a design for engines that can accelerate a spacecraft without producing detectable emissions. It is best fitted to a stealth spaceship described in 'The Hydrogen Steamer: Stealth Spaceship Concept'.

Stealth designs in space rely mainly on low temperatures to remain undetected. However, low temperatures are the antithesis of efficient propulsion: exhaust velocity suffers, and improbable propellant to dry mass ratios are required to achieve any sort of useful deltaV.

Some propulsion designs allow for undetectable propulsion, but they suffer from low specific power, creating situations where the spaceship has to accelerate for months to break an orbit. 

Here is a design which might alleviate this issue.

Expansion-cooled Nuclear Thermal Rocket

Most rocket nozzles are a balance between weight, wall temperature and expansion ratio. Lightweight nozzles are better for high power-to-weight ratio, higher wall temperatures allows for higher exhaust velocity and less regenerative cooling requirements, while lower expansion ratios allow for shorter, lighter nozzles.

For a stealth spaceship, a different set of characteristics are required. 

The Children of a Dead Earth game has an accurate physical model of these rocket engines. While it makes some sacrifices in detail, it does render an accurate picture of how exhaust gasses are cooled as they expand. 

Here is a design created in the game's Module Designer:

Details from the editor
Maximal expansion of the gasses was selected, with a low initial temperature. Performance suffers, and the nozzle weighs a lot, but the exhaust is released at the extremely low temperature of 22K. 

Coincidentally, this is just over the vaporization temperature of hydrogen, which the hydrogen steamer concept boils off to keep cool. This means it is no more or less detectable than the main hull of the stealth ship.

If we use the 2*10^-18 watt per square meter figure from this lecture on the sensitivities of the Spitzer telescope, and assuming the hydrogen plume has an emissivity of 0.01 at 22K temperature and vacuum pressure, the detection range is about 8154km. 

However, the instruments are assumed to look at the same spot in the sky for 10000 seconds to obtain a signal-to-noise ratio of 10. Also, the telescope's field of view is only 5.2x5.2 arcminutes. Considering that the sky's area is 148,510,800 square arc minutes, a much lower 'dwell time' per section has to be accepted. With 100 seconds of observation, detection range is only about 800km!

This can be compensated for by using multiple telescopes and increasing sensitivity, but there is a lower limit determined by how quickly the stealth spaceship crosses the field of view. 

Curved Nozzle
A solution to up-the-nozzle emissions
A regular straight-nozzled rocket engine is very visible as the temperature at the nozzle's throat can reach thousands of degrees. This 'hot spot' can be seen by looking up the nozzle. This means that even if the expansion of exhaust gasses creates a very cold stream, and the nozzle walls themselves are cooled and insulated, the rocket will remain visible to sensors looking directly up the nozzle.

A curved nozzle attachment can remedy this issue, and prevent direct 'up-the-nozzle' observation of the high temperature gasses at the nozzle throat. All that can be seen by an external observer is the cool gasses and cold nozzle walls.  

Care will have to be taken to prevent gases from colliding with the curved walls at compressing, which raises their temperature again. We would need a wall that curves very smoothly, which makes for a very large nozzle.


With 4.48km/s exhaust velocity, the 'ECCN' rocket engine designed in the game editor vastly outperforms traditional stealth options, such as cold gas thrusters (400-700m/s). 

While the pulsed solar-thermal thruster has higher exhaust velocity (9km/s) and a mass driver has potentially unlimited exhaust velocity (10-100km/s+), both suffer from various downsides. The pulsed thruster is limited to the amount of sunlight that can be captured by the spaceship for propulsion, leading to very low thrust. A mass driver must rely on an on-board power generation system which has to be cooled and creates a system with very low specific power.

In comparison, the ECCN here generates a respectable 110kN of thrust and has a specific power of 46MW/ton! This is over twenty times better than solar power, and nearly five times better than most nuclear reactors. 

The best part is that the ECCN has no overhead requirements (power consumption, solar panels, radiators ect.). These rockets can be mounted in multiples. Two ECCN rockets will produce 220kN, three will produce 330kN, and adding more will not compromise stealth. 

The ECCN design can also be scaled up and down. A gigawatt engine with a sufficiently long nozzle is conceivable. An ECCN with an even more extreme expansion ratio can create even colder gasses.

Note that this is just one possible design that can illustrate the concept. It is possible to use even larger expansion ratios, so that the gases start at an even high temperature and end up with an even lower temperature. For example, hydrogen propellant can start at 3000K and end at 14K. This allows for both an increase in exhaust velocity to 9km/s and a decrease in detection range by a factor 6!


  1. I'm a little confused. If you're running nuclear thermal propulsion, regardless of what you do with exhaust, doesn't the reactor itself run into the same drawback as onboard power generation for a mass driver? Does the exhaust gas also act as coolant so you don't need radiators? (Not that knowledgeable about atomic rockets, sorry)

    1. The reactor is cooled by the propellant flow, and thus needs no additional cooling.

    2. As Unknown said, all of the rocket engine's heat is absorbed into the exhaust.

    3. Thanks for the replies. In that case, wow! I am totally on board.

    4. Its all thanks to the Children of a Dead Earth game. Very accurate simulator, lots of hard physics involved. It reveals some of the minor details that many don't pick up upon, like the fact that the exhaust temperature of a sufficiently expanded nozzle is really cold.

    5. I would actually recommend a plug/aerospike nozzle design. The original intent was to increase the transfer efficiency from exhaust to spacecraft momentum (giving more area for the exhaust to act on in order to generate reaction)... but the design would also effectively shield the high temp throat exhaust from direct observation.

    6. Looking at various aerospike nozzle designs, I don't see how they would prevent an enemy directly behind the engine from seeing up into the hotter portions of the exhaust.

    7. Actually, come to think of it, it would depend upon which "aerospike" design you are looking at. I was thinking more alongs the lines of the J-2 upgrade, and not the x-planes ersion (in case there was any confusion.
      Again, depending upon the specific design, the aerospike "plug" typically shields the throat of the nozzle, where you have the highest temperature. It will also redirect the thrust, acting as a focus point, so that the exhaust has less divergence. The key is careful design and placement. It actually acts in the same manner as the curved nozzle, but it also serves to boost performance. A 2D version of the plug/Aerospike might be ideal, as it can be used effectively and efficiently in combination.

    8. I still cannot visualize that specific configuration as being helpful. The gas only cools down to undetectable temperatures at the very end of the nozzle (full expansion). In an aerospike, most of the warm gas is exposed to vacuum, meaning it is very visible. The throat has the highest temperatures, but the gasses just downstream of the throat are still very hot!

  2. From a military perspective, the question becomes comparing the resources demanded to create low observable spacecraft vs creating large staring arrays to find them? This isn't quite like the current debate about radar evading stealth aircraft (designers always knew the aircraft were more visible to VHF radars, but optimized to be less visible to VHF and UHF tracking and fire control radars from enemy aircraft and guided missiles. I suspect that the resource allocation question will still favour more mirrors and more arrays vs more low observable spacecraft (and so long as they radiate at 22K, they still stand out against the 3K background of space).

    Staring arrays will be available both as permanent installations (space launchers based on mass drivers, laser thermal propulsion or "laser web" power beams will need accurate tracking, for example), and as "munitions" aboard spacecraft and constellations, in the form of small mirrors or swarms of drone spacecraft deployed to create "mirrors" up to one light second across.

    Even the beam expander mirrors of combat spacecraft will be pressed into use, and spaceships which can deploy mirrors ranging from 3 to 10 metres in diameter will have some pretty impressive observation capabilities, especially when informed by networks of other mirrors in the constellation. Of course a 10 metre beam expander not only locates the target, but also guides the laser beam sent against it moments later....

    1. I like where you're going with this.

      A hydrogen steamer is the sort of asset which takes a lot of exotic materials and advanced technology to create, and has to be sent away for months on end. A sensor array, on the other hand, can compensate for lesser capabilities with increased numbers and dwell times.

      I mentioned this in another post: the less resources or disadvantaged side will always focus on detection and anti-ship weapons, while the one with the most resources or perceived advantages will focus on stealth weaponry and anti-defenses weaponry.

      I forgot about munitions-like sensor arrays. They were mentioned in the Laser Web post. The drones won't achieve the resolution or spotting duration of dedicated platforms, but they will tell you if a stealth ship is nearby.

      Combat mirrors are likely to be tuned to have very high transparency (lens) or reflectivity (mirrors) for the specific wavelength used by the laser weapon. I doubt that they will be very effective in focusing the mid-infrared wavelengths of very cold spaceships.

      Also, I mentioned a design where helium-coolant heat pumps were used to cool down the hull of a stealth ship to 2.7K, consuming in the process 13x the amount of liquid hydrogen of a conventional 'steamer' that relies on evaporation of liquid hydrogen.

    2. Oddly, I am seeing this in the opposite direction, EarthForce can access huge arrays of mirrors (both dedicated military mirrors and coopting or coercing the industrial pointing/tracking arrays of mass drivers, laser and microwave generators and so on) to create giant "virtual" mirrors of almost arbitrary size. Add the deployed constellation and you also have "binocular" vision, or the ability to look at potential target volumes from a multitude of directions (and the idea of a light second wide mirror in Earth's Hill Sphere coupled to a light second wide mirror of a fully deployed constellation on the way to Mars should give you a good idea of the capabilities being described).

      While its true that the expander mirrors of combat spacecraft might not be optimized for that particular wavelength (unless it is a conscious decision to make the laser weapon emit in that part of the infrared so you "can" point and shoot), my take is the mirrors will be part of a hunter/killer array, with the huge virtual mirrors identifying potential targets or points of interest, and the laser mirror going for the fine definition of the target. And of course, if you have several Laserstars in the constellation, you now have a smaller virtual mirror focused on that particular point....

      Once again, the various space navies will need to do the cost benefit analysis, and in general, the side with the lower resource base will be better off applying resources to detection. In the Space environment, I suspect that the fully mature polities (like Earth) will be able to leverage the advantage of the existing space infrastructure in ways that less developed colonies cannot.

  3. Stupid question: Would a hydrogen/helium steamer be better suited to work in the outer solar-system than in the inner system? My gut says outer (further from the sun and thus harder to occlude it, easier to cool, larger orbits requiring more staring array platforms). Of course, given that HYSSs can increase to huge sizes and remain effective...

    I'm also thinking that a Mercury colony could set up a series of very wide mirrors to sweep the area with sunlight to heat up such craft and their hydrogen puffs (possible nickname for such weapons- 'the eye of Sauron') and secure their space against stealth incursions.

    1. No stupid questions ;)

      In the outer solar system, there is less sunlight, so a stealth ship would need to expend less liquid hydrogen to stay cool. This increases their endurance.

      An 'eye of sauron' would have to have a range of about 5AU. Setting up a laser web chain that long might be problematic... the same concept can be recreated at a much closer distance, by basing it on a nuclear-powered generator embedded in ice. With such a huge heatsink, the reactor could be 3-4x more powerful than space-craft based reactors.

    2. I think he was intending "eye of Sauron" as a more localised defence against stealth, since he says "secure THEIR space against stealth". If this were the case, then it should work... but it might not be necessary given the solar flux intensity native to that region in the first place.

    3. "a more localised defence against stealth"

      Precisely- a local observation system in case a HYSS manages to both deal with occlusion, and to cool itself to a temperature suitable for that area/ local phenomena.

  4. Somewhat tangental to the discussion is a new entry in the ever popular Atomic Rockets:

    Pulsing the nuclear reactor, and perhaps more importantly finding a way to "harvest" the energy in the neutrons to heat the reaction mass allows you to get far more power out of the reactor than conventional NTR's, and brings solid core technology more in line with the sort of performance that classic age SF authors often seemed to imply with their "nuclear" spaceships.

    This particular version of the NTR isn't going to be "stealthy", but then again, many military forces right here on Earth haven't decided that the extra expense of stealth technology is worthwhile to invest in, despite it dating back to the end of the Carter Administration , almost 40 years ago.

    I suppose anyone willing to do the R&D could combine the elements of the pulsed solid core NTR with the expansion cooling system described in the main post, and it would be interesting to see if the results are worthwhile, or fall into the "more trouble than it is worth" category.

    1. This has prompted me to look into pulsed mode of operations. Look at the comment just below: no curved nozzle is needed, just expand the propellant pulse behind a shutter.

  5. I'm a bit confused. The exhaust is expanding while traveling down the nozzle, but when it encounters the curve at 4.5 km/s, it will be compressed against that curve, heating up(how much?), and while being deflected by the curve, it will push the nozzle sideways wrt. to the original exhaust direction, and if the curve is an average of 90 degrees as in the picture, a single engine like this would just rotate the vessel. Of course the ship could have a number of these in a radially symmetrical configuration. Still, I would expect the curve to be heated quite a lot from deflecting 110 kN worth of 4.5 km/s exhaust, and it would have to be quite sturdy to even withstand it.

    One idea that comes to mind is that since the curve must be a long-ish way from the center, it could maybe safely expose a much hotter up-the-nozzle view if the ship is spinning along its axis, because it would be only visible from any one direction for only a small fraction of time.

    1. You are quite right that a straight 90 degree bend would cause shockwave heating of the exhaust gasses and reveal the spaceships' position... but there are design 'tricks' that can counter this problem.

      The first is to have a very gradual curve. This limits the shockwave compression.
      Another is to over-expand the gasses. Reduce them to 2K instead of 20K, for example. This allows the compression at the curvature to increase the gas density by x10, and therefore gas temperature by x10, without exceeding your stealth requirements.
      There might be other options someone better ersion in hypersonic flow physics would be able to suggest. Maybe you can bleed in liquid hydrogen at the outer curve of a curved nozzle to mix in with the hypersonic gasses and cool them down the evaporation.

      The other concerns you raised are negligible. 110kN is not a lot of force. It is 11 tons' weight distributed over the nozzle's surface area. Thin sheets of metal can handle these forces, I assure you.

      Something I am considering right now is a pulsed thruster.

      Hot propellant is injected into the de Laval nozzle behind a shutter. It is expanded until it is cool. Just as it reaches the tip of the nozzle, the shutter opens for the gas to go through. The view into the nozzle is devoid of hot heat sources. Once the shutter closes, a fresh load of hot propellant can be injected.

  6. What are the limits of expansion cooling? If you can expand to 2K, why wouldn't you? You want to be at the lowest possible temperature.

    If super expansion works, wouldn't you be better of heating the hydrogen coolant to 300K and expanding it through your nozzle, rather than dumping it at 22K? How curved can the nozzles be - can we have two to cancel out any thrust we don't want?

  7. You could also use nozzles that are aerospikes or linear aerospikes with a S ducted or the Serpentine nozzle to help reduce that instead of needing a curved nozzle