Using bypass air to increase rocket thrust

[farmerben]

Take a close look at the Whiplash engine in KSP.  The engine has a skirt with radial air intakes to mix bypass air with the exhaust plume.  The bypass air is not fed directly into the combustion chamber, but mixed with the exhaust.  The heating of the bypass air inside the skirt increases thrust.   If the radial air intakes had ramps to control input air, an afterburner could be used, but increased thrust is possible even without an afterburner.  

My question is: Is it practical to combine this type of bypass air skirt with a regular rocket engine?

[Wjolcz]

Yes. It's called air-augmented rocket. It also increases specific impulse, but  I might be wrong about that one. Russians experimented with that during the cold war but AFAIK they modified only one ICBM (Gnom) to have something like that, the engineer working on it died and nobody ever tried to do anything with it.

There's also a NASA project like that.

Edited July 3, 2018 by Wjolcz

[RCgothic]

Basically it's a lot of weight and complication for most rockets that spend very little time in the lower atmosphere. More generally useful for missiles.

[Wjolcz]

57 minutes ago, RCgothic said:

Basically it's a lot of weight and complication for most rockets that spend very little time in the lower atmosphere. More generally useful for missiles.

AFAIK it's not that bad and (compared to other types of engines) relatively easy to make kind of structure/engine. I'd imagine having something like that and an aerospike would make for a pretty good SSTO design.

Edited July 3, 2018 by Wjolcz

[farmerben]

Thanks Wjolcz!  It is useful to know there is a name for it and somebody thought of it.  It appears few tests have been done with the concept.  

I suspected the biggest problem would be vibration or cooling considerations of the primary rocket engine, or overheating issues with any practical skirt.  The problem of large sized air intakes and extra weight does not seem all that serious to me.  If it is only used with the first stage for the first minute or two and then detatched, and is reusable, then the weight does not seem like a problem.  We should in fact save a lot of weight in fuel for the first stage.

All the NASA designs I have seen involving ramjets are for SSTO's.  In that context I understand why the weight might not be worth it.  But in a multistage design this is not an issue.... you detach the parts when you run out of atmosphere.  If the parts are reusable and do not require active cooling it seems like a bargain.  

It also seems to me that Whiplash style air ducts would be stable, in that a change in pitch will increase the relative airflow on the side the vessel is pitching into.  

[K^2]

1 hour ago, farmerben said:

I suspected the biggest problem would be vibration or cooling considerations of the primary rocket engine, or overheating issues with any practical skirt.

Vibration shouldn't be a huge factor in most designs, but cooling does play a role. Mostly, it comes down to extra weight and complexity, especially if you want your engines to gimbal.

The crucial thing to keep in mind is that while it's a sizable boost to efficiency, it's not free, and is effective over very narrow window. Once you ascend past certain altitude, it becomes extra weight you're carrying. If you have a lot of stages, and discard the air augmented stage early, you get a net win on fuel. But most of the cost of the rocket isn't fuel. It's engines and other mechanical components that get expended. You win a lot more by reducing number of stages, improving their efficiency across the range of altitudes. This is the direction SpaceX has been going, and part of the reason why their launches were cheaper even before they started working on reuse.

It's a great idea overall, but it doesn't mesh well with direction that chemical rockets are taking right now.

[farmerben]

I agree with all those concerns.  I don't know if the complexity, gimbaling, and cooling considerations can be adequately solved.  

But this idea of a simple skirt around the rocket engines seems to have a lot of merit compared to side mounted ramjet boosters. It should improve efficiency across all altitudes and airspeeds.  The P51 Mustang channeled bypass air around its exhaust pipe and got a noticeable increase in thrust at 500mph.  The drag created by air intakes must be contrasted with the drag created where a cylindrical fuel tank ends a meter or more before the exhaust plume begins with a conventional rocket.  Combustion is difficult to achieve at high altitudes and airspeeds, but heating bypass air works basically the same way.  Fuel may be cheap, but good fuel tanks are not.  The question is how much a skirt would cost in terms of weight and money compared to the fuel tanks you would need without them.

If a relatively thin titanium alloy skirt could withstand the temperatures, without active cooling, it might be a better deal.

[kerbiloid]

Atmosphere is 8 km thick, if redistribute it uniformly.
Density 1.225 kg/m³.
Oxygen 21%.
So, 8000 * 1.225 * 1 * 0.21 = 2 t of air oxygen per m² of cross-section.
To get more you have to use a fan, and its efficiency quickly falls while the rocket is getting up.

A ramjet can work at high velocities, at least 1 km/s or so.
Distance to get 1 km/s at 20g average acceleration = (1000²/2*20*9.81) = 2.5 km.
For a crewed rocket it's even more.
So, when you reach the ramjet ignition speed, air is mostly already behind.

So, air augmenting is useful in near-horizontal flight, but not so useful in vertical.

[hms_warrior]

uhm...the atmosphere is more than 8 km thick. And we don't care about oxygen. Atmosphere could be pure CO2 and air-augmented rocket would still work. (it is not a ramjet)

Edited July 4, 2018 by hms_warrior

[kerbiloid]

1 hour ago, hms_warrior said:

the atmosphere is more than 8 km thick.

When it's exponential.
If redistribute it uniformly with sea-level density, its thickness is equal to the scale factor (that's why its another name is "uniform atmosphere height", or how is it in English).
H = RT/(gM) = 8.31441 * 273 / (9.81 * 29*10^-3) ~= 7.9 km for Earth.

We care about oxygen because it's the oxidizer. We aren't interested in nitrogen.

P.S.
Gnome had a ramjet.
http://militaryrussia.ru/blog/topic-801.html

A quad of solid-fuel engines in a single hull (1st stage) would accelerate it up to ramjet ignition speed (1.75 M).
Ramjet (2nd stage) accelerated it up to 5.5 M,
Two more solid-fuel stages (3rd, 4th) accelerated the payload up to near-orbital speed.

(Also, Gnome is ICBM, so it's not at all about economy).

Edited July 4, 2018 by kerbiloid

[K^2]

16 minutes ago, kerbiloid said:

When it's exponential.
If redistribute it uniformly with sea-level density, its thickness is equal to the scale factor (that's why its another name is "uniform atmosphere height", or how is it in English).
H = RT/(gM) = 8.31441 * 273 / (9.81 * 29*10^-3) ~= 7.9 km for Earth.

Nitpick - this assumes uniform temperature, which it isn't. But yeah, that's Earth's Scale Height, and it's a useful tool in making estimates.

In fact, there is a very simple equation of estimated delta-V losses to aerodynamic drag and gravity using it. v_lost = 4gH/v_t, where v_t is terminal velocity at the surface. It makes a ton of assumptions, but it gives an ok ballpark estimate for Earth of something around 1km/s - 2km/s lost. (This estimate works way better in KSP) So at the most, we are fighting over these 2km/s. Which isn't nothing, but there are definitely bigger wins to be made. This is the same reason why we rarely launch rockets from airplanes.

hms_warrior is right that we don't care about ignition, though. Air augmented rocket generates additional thrust from launch pad. Its operating principles are very different than these of a (sc)ram jet.

[kerbiloid]

2 minutes ago, K^2 said:

Nitpick - this assumes uniform temperature, which it isn't.

The only uniform thing in this formula is R, because it's a constant.

[K^2]

You are treating T, g, and M as constants, though. Otherwise, you'd have a very nasty differential equation on your hands. And while M and g are very nearly constant at relevant scale, T is very positively not. If you were to say that Earth's atmosphere is 8km thick when layered uniformly, why should you use T as surface temperature? Why not other elevation or even some sort of average temperature?

The reason we use surface T here is because scale height is used to model pressure changes, P(h) = P(0) * exp(-h/H). This formula works so long as T remains roughly constant, and we usually use it to model atmosphere near the surface. So saying that H represents thickness of the atmosphere in some way is not entirely correct. It works as a mathematical abstraction, but in no other way.

But like I said, still useful for estimates. As a starting point, if nothing else.

[RCgothic]

Wikipedia's article on augmented rockets claim they weigh as much as 5x as much as a normal rocket engine when you include the intakes, and that the intakes are not a trivial engineering challenge. I would suggest that's the reason we don't see them on first stages right there.

[Hesp]

50 minutes ago, RCgothic said:

Wikipedia's article on augmented rockets claim they weigh as much as 5x as much as a normal rocket engine when you include the intakes, and that the intakes are not a trivial engineering challenge. I would suggest that's the reason we don't see them on first stages right there.

Making some assumptions based on Falcon 9 launch profiles, the rocket velocity at 20km altitude is around Mach 2 / 2.5

It's safe to assume that most of the air augmentation will happen below that altitude, so exotic intakes (J58 variable geometry spike) are not needed. A proper designed diverterless supersonic intake could suffice.

[kerbiloid]

2 hours ago, K^2 said:

You are treating T, g, and M as constants, though.

Not I am treating, but a reference book gives this as a common formula to estimate.
If you substitute average values, you get an estimated value.

You should know that more accurate formulas for density and pressure are more complicated, and this value is by definition an approximation.
This is exactly that case when nobody needs that

2 hours ago, K^2 said:

very nasty differential equation on your hands. And while M and g are very nearly constant at relevant scale, T is very positively not. If you were to say that Earth's atmosphere is 8km thick when layered uniformly, why should you use T as surface temperature? Why not other elevation or even some sort of average temperature?

, as a low accuracy estimation is well enough.

2 hours ago, K^2 said:

But like I said, still useful for estimates. As a starting point, if nothing else.

Were you going to use it in a real life rocket project?

[K^2]

30 minutes ago, kerbiloid said:

Not I am treating, but a reference book gives this as a common formula to estimate.
If you substitute average values, you get an estimated value.

Again, for scale height. Which doesn't mean what you implied it means, specifically because of non-constant temperature.

[kerbiloid]

3 minutes ago, K^2 said:

Again, for scale height. Which doesn't mean what you implied it means, specifically because of non-constant temperature.

I don't understand, what are you trying to prove.
This estimation is enough good for both scale height of the atmosphere, and thickness of a uniform atmosphere.
Nobody uses it for precise calculations.
If you have a better estimation, feel free to share it.

[Wjolcz]

3 hours ago, RCgothic said:

Wikipedia's article on augmented rockets claim they weigh as much as 5x as much as a normal rocket engine when you include the intakes, and that the intakes are not a trivial engineering challenge. I would suggest that's the reason we don't see them on first stages right there.

The 5x part sounds a bit dodgy. Can we have a link or at least quotation?

We had a discussion about this a while back. I too think rockets can fly just fine even without any additional intakes/cowling used to boost their performance but it's mostly the lack of R&D being the reason why we don't see air augmentation on rockets. That and the fact that reusable stages became feasible very recently. Similar case with the aerospike. It could fly and be used instead of regular rocket engines. It's just that it would need more R&D and nobody wants to do that since regular nozzles work just fine.

With all the New Space companies I wouldn't be surprised if somebody tried to build something like that eventually.

Edit: Got it. But it says "citation needed" so we will never know unless somebody decides to do more R&D.

Spoiler

Finally, the air ducting weighs about 5× to 10× more[citation needed] than an equivalent rocket that gives the same thrust.

Edited July 4, 2018 by Wjolcz

[kerbiloid]

(I would expect aerospikes first on warheads).

[farmerben]

That is a good starting point.  But an air augmented engine will have a unique flight profile.  More research and modelling would be needed to determine exactly what that is.

I would guess that in the low atmosphere one would want to go a bit slower.  We would want our thrust bonus to exceed the gravity losses of a slower climb, and otherwise stay slow to reduce drag.  

In the upper atmosphere (~20-70km) we would want to pitch more horizontal and throttle up.  This maneuver would happen later than a scramjet design, but earlier than a pure rocket.

Edited July 4, 2018 by farmerben
quote didn't attatch

[farmerben]

6 hours ago, Hesp said:

Making some assumptions based on Falcon 9 launch profiles, the rocket velocity at 20km altitude is around Mach 2 / 2.5

It's safe to assume that most of the air augmentation will happen below that altitude, so exotic intakes (J58 variable geometry spike) are not needed. A proper designed diverterless supersonic intake could suffice.

That is a good starting point.  But an air augmented engine will have a unique flight profile.  More research and modelling would be needed to determine exactly what that is.

I would guess that in the low atmosphere one would want to go a bit slower.  We would want our thrust bonus to exceed the gravity losses of a slower climb, and otherwise stay slow to reduce drag.  

In the upper atmosphere (~20-70km) we would want to pitch more horizontal and throttle up.  This maneuver would happen later than a scramjet design, but earlier than a pure rocket.

[YNM]

Would it be possible to "just" make something like hush kits to rocket engines ? Or are hush kits the perfect explanation as to why such air-mixing measures are all in vain ?

[Xd the great]

14 minutes ago, YNM said:

Would it be possible to "just" make something like hush kits to rocket engines ? Or are hush kits the perfect explanation as to why such air-mixing measures are all in vain ?

If you are clearing a large area for a rocket launch, there is no need to reduce sound pollution. And people get to enjoy the rocket booming.

Plus hush kits means more weight. And probably a drop in performance.

[YNM]

7 minutes ago, Xd the great said:

If you are clearing a large area for a rocket launch, there is no need to reduce sound pollution. And people get to enjoy the rocket booming.

Plus hush kits means more weight. And probably a drop in performance.

The devices roughly explained here are very close to hush kits in form. Hence my question.

I know the actual hush kits are made for noise reduction, but could we use them for performance boost instead, or is that just a figment in people's mind ?

Edited July 4, 2018 by YNM