Hydralox with Oxygen Thrust-Augmented Nozzle?

[edzieba]

So it can work.

No, it can burn.  It doesn't mean the idea is viable.  Afterburning in a nozzle doesn't really help.

That's kind of the crux: how far can TAN be pushed in terms of thrust?
For the original TAN concept of taking an very overexpanded bell and compensating with some burning within the bell to prevent flow separation (and get a bit of bonus thrust) we can be pretty sure it works, as it has been subscale tested.
But if you want to take the same concept and generate the majority - or even an appreciable fraction - of your thrust in post-throat combustion, I'm not so sure. You don't get the benefit of the DeLaval nozzle to accelerate the combustion products, and your nozzle extension needs to be strengthened.

[Robotbeat]

The concept of adding more propellant doesn’t require any combustion in the nozzle at all, BTW. You could use nitrogen or Argon, or even powdered rock, and get much the same effect. Adding oxygen is good because it’s basically as cheap as nitrogen, is liquid, and you’re carrying liquid oxygen anyway, plus you can get more complete combustion (as rocket engines typically run fuel-rich).

[sevenperforce]

If you want to pass everything through the combustion chamber and just vary the O:F ratio then here's that design with a preburner that initially functions as a classic ORSC turbopump, but then can alter the LOX flow in an expander cycle to pull heat off of the preburner and thus increase specific impulse while decreasing LOX flow and thrust.

[acksed]

So it can work.

No, it can burn.  It doesn't mean the idea is viable.  Afterburning in a nozzle doesn't really help.

That's kind of the crux: how far can TAN be pushed in terms of thrust?
For the original TAN concept of taking an very overexpanded bell and compensating with some burning within the bell to prevent flow separation (and get a bit of bonus thrust) we can be pretty sure it works, as it has been subscale tested.
But if you want to take the same concept and generate the majority - or even an appreciable fraction - of your thrust in post-throat combustion, I'm not so sure. You don't get the benefit of the DeLaval nozzle to accelerate the combustion products, and your nozzle extension needs to be strengthened.

I went searching and actually found the paper (embedded below). It turns out the reason it was using the LANTR chamber, injectors and nozzle was it was actually spinning off from the oxygen injection part of LANTR; I accidentally recreated a knockoff 'vegan' LANTR.

Anyway, the physical tests used that, water-cooling the nozzle and chamber while pushing it to about a mean of 40% RP-1/LOX augmentation, with a peak of 77% and a mass-flow of 108%, but the authors did some calculations on two bigger engines: the NK-33 and an RS-68-equivalent, both enhanced with TANs.

Their takeaways were that the NK-33's TAN could be expanded to a ratio of 58:1 with 40% augmentation while keeping the nozzle exit pressure of 6 psi/0.41 bar. This gave an average gain of 4.5 in specific impulse (assuming 20% of its mission is sea-level and the rest in vacuum) and an increase in T:W ratio from 128:1 to 150:1.

The hydrolox RS-68-like could be pushed further, increasing its T/W from 46:1 to "greater than 60:1". The graph of impulse over thrust augmentation given shows an increase in nozzle ratio to about 50:1 enabled a thrust augmentation of about 80% and an averaged increase in Isp of 11 seconds. That's pretty damn good.

Pushing the TAN to 3 times the mass-flow in just oxygen might be too much of an ask, though.

[InterestedEngineer]

So it can work.

No, it can burn.  It doesn't mean the idea is viable.  Afterburning in a nozzle doesn't really help.

That's kind of the crux: how far can TAN be pushed in terms of thrust?
For the original TAN concept of taking an very overexpanded bell and compensating with some burning within the bell to prevent flow separation (and get a bit of bonus thrust) we can be pretty sure it works, as it has been subscale tested.
But if you want to take the same concept and generate the majority - or even an appreciable fraction - of your thrust in post-throat combustion, I'm not so sure. You don't get the benefit of the DeLaval nozzle to accelerate the combustion products, and your nozzle extension needs to be strengthened.

I went searching and actually found the paper (embedded below). It turns out the reason it was using the LANTR chamber, injectors and nozzle was it was actually spinning off from the oxygen injection part of LANTR; I accidentally recreated a knockoff 'vegan' LANTR.

Anyway, the physical tests used that, water-cooling the nozzle and chamber while pushing it to about a mean of 40% RP-1/LOX augmentation, with a peak of 77% and a mass-flow of 108%, but the authors did some calculations on two bigger engines: the NK-33 and an RS-68-equivalent, both enhanced with TANs.

Their takeaways were that the NK-33's TAN could be expanded to a ratio of 58:1 with 40% augmentation while keeping the nozzle exit pressure of 6 psi/0.41 bar. This gave an average gain of 4.5 in specific impulse (assuming 20% of its mission is sea-level and the rest in vacuum) and an increase in T:W ratio from 128:1 to 150:1.

The hydrolox RS-68-like could be pushed further, increasing its T/W from 46:1 to "greater than 60:1". The graph of impulse over thrust augmentation given shows an increase in nozzle ratio to about 50:1 enabled a thrust augmentation of about 80% and an averaged increase in Isp of 11 seconds. That's pretty damn good.

Pushing the TAN to 3 times the mass-flow in just oxygen might be too much of an ask, though.

The more important metric than T/W is thrust per unit area.

Notice how the bottom of Booster is crammed with Raptor-2s, or SLS requires strapon boosters to get off the ground.  Any decrease in thrust per unit area will decrease the T/W of the entire rocket at the pad.

A large engine bell is contraindicated for the first 4-5 km of flight.  That 58:1 expansion ratio is going to require a lot of surface area. (Raptor-2 is 40:1)

[Robotbeat]

For Starship, maybe, but there are other concepts which go wider instead of taller.

[InterestedEngineer]

For Starship, maybe, but there are other concepts which go wider instead of taller.

how wide can you go?  (besides in KSP).

At some point air resistance becomes a problem, or building a wide enough launch mount (lots of people choked about Starship's 9 meter diameter, which can't be transported overland very far)

[Robotbeat]

It’s best to think of weight per unit cross section area. Starship is actually very tall and heavy for its cross section, but you don’t have to go that way. Starship gets extremely low drag because it’s only 9m wide but weighs like 5000tonnes. N-1 weighed about half that but was 17m wide. So in total it had a weight per area a factor of 6.5x smaller than that of Starship, plenty of room to reduce the thrust-per-area of the engine from Raptor’s value.

[InterestedEngineer]

It’s best to think of weight per unit cross section area. Starship is actually very tall and heavy for its cross section, but you don’t have to go that way. Starship gets extremely low drag because it’s only 9m wide but weighs like 5000tonnes. N-1 weighed about half that but was 17m wide. So in total it had a weight per area a factor of 6.5x smaller than that of Starship, plenty of room to reduce the thrust-per-area of the engine from Raptor’s value.

And now you have longer feed pipes.

Longer pipes can sympathetically vibrate with a lot more frequencies than short pipes.

Not sure how N-1 did with mass ratio, larger is more mass too.

[sevenperforce]

It’s best to think of weight per unit cross section area. Starship is actually very tall and heavy for its cross section, but you don’t have to go that way. Starship gets extremely low drag because it’s only 9m wide but weighs like 5000tonnes. N-1 weighed about half that but was 17m wide. So in total it had a weight per area a factor of 6.5x smaller than that of Starship, plenty of room to reduce the thrust-per-area of the engine from Raptor’s value.

And now you have longer feed pipes.

Longer pipes can sympathetically vibrate with a lot more frequencies than short pipes.

Not sure how N-1 did with mass ratio, larger is more mass too.

The N1's first stage had a dry mass of ~130 tonnes and carried ~1750 tonnes of LOX and RP-1, a fuel fraction of 93% (although 15 of those tonnes were expended to boiloff and throttle-up before the rocket even left the pad). With 3400 tonnes of propellant on board, Superheavy would need to be under 240 tonnes dry in order to improve on the N1's mass ratio. Of course, RP-1 is much more dense than CH4, so Superheavy is playing catch-up from the beginning.

The Saturn V's S-IC first stage, with a common bulkhead and monocoque tanks, was significantly better than the N1, at a fuel fraction of 95.4%.

[Proponent]

The Saturn V's upper stages definitely had common bulkheads, but I don't believe the S-IC did (see 61st page of the SA-507 Flight Manual, attached).

There were upgrade proposals in which the forward dome of the RP-1 tank would be replaced by honeycomb of mini domes to reduce the amount of wasted space.

[Proponent]

I’d go 26:1 for O:F ratio for a first stage for hydrolox. You want super high bulk density and low Isp in the beginning of flight, in this case 287s Isp. 26:1 O:F would give you bulk density nearly that of methalox.

Maybe there are other reasons for going that high, like reducing temperature or increasing thrust, but otherwise I would think you wouldn't want to go any leaner than the mixture ratio that maximizes impulse density (i.e., the product of specific impulse and bulk density).  That ratio will of course depend on chamber pressure and expansion ratio, but when I was playing around with a simple variable-mixture-ratio hydrolox SSTO model a few years ago, it came out at about 17.

EDIT: Corrected link to old post; now shows how O/F changes from lift-off to shutdown..

[Robotbeat]

Again, the reason I was interested in it was long term reducing the energy and/or propellant costs. It can make a big difference for the first stage.

Keep in mind that if you neglect stage dry masses (and gravity and drag) and consider just energy put into the reaction mass, an exhaust velocity equal to 62.75% of the mission delta-v maximizes the conversion efficiency (64.76% in this case) of reaction mass energy to payload kinetic energy IF you pick a constant exhaust velocity. IF you let the exhaust velocity change over time, the optimum exhaust velocity is whatever your current velocity is relative to your starting point (except for right at first, as you want to keep your total propellant mass finite LOL), and the efficiency can then approach 100% (assuming basically an infinitely long insulated nozzle extension… and perfect combustion efficiency…).

Maybe we’re all just condensed combustion exhaust from an enormous 4 dimensional rocketship, expanding since the Big Bang haha…

[1]

I went searching and actually found the paper (embedded below).

A key feature of this paper, assuming I'm reading it correctly, is this line:

These analyzes indicate that providing the LOX and hydrocarbon fuels to the TAN injectors from the boost pump discharge is feasible with minimal impact on the main turbopump assembly (TPA), preburners, and main chamber operating performance.

IMO, this suggests that the overall O/F ratio if the engine as a whole does not appreciably change. To me, this makes sense, as drastically increasing passive reaction mass comes with an opportunity cost in total available energy. If an increase in energy efficiency does not overcome a reduction in total available energy, then it's a poor trade. Better to just burn the less efficient engine for a longer time.

Pushing O/F more towards 'O' may have uses, but I wouldn't drift too far away from stoichiometric in that direction, either.
26:1 to me feels waaaaay to high. But maybe 10:1 early during boost, assuming the nozzle can survive? Might be reasonable. Haven't run any numbers on my end.

[Robotbeat]

Thing is that passive propellant like oxygen is virtually free, energy-wise, compared to fuels. For a given fuel flow (power input), you generate way more thrust by adding a huge amount of passive propellant.

It‘s like a jet engine with a larger bypass ratio, except the bypass air is oxygen that you’re carrying with you. A jet engine with a high bypass ratio is more energy-efficient than a turbojet. But like that analogy, it’s not advisable to use this configuration at high speeds because it’s no longer energy efficient to have accelerated all that reaction mass with you. There’s an optimum mission velocity for this sort of thing. Early in flight, it’s optimum to do this. Later in flight, it’s more efficient to operate at higher exhaust velocity (even though the thrust per unit power will be lower at higher exhaust velocity), ie without thrust augmentation by passive propellant but instead nearer to stoichiometric.

[InterestedEngineer]

Thing is that passive propellant like oxygen is virtually free, energy-wise, compared to fuels. For a given fuel flow (power input), you generate way more thrust by adding a huge amount of passive propellant.

It‘s like a jet engine with a larger bypass ratio, except the bypass air is oxygen that you’re carrying with you. A jet engine with a high bypass ratio is more energy-efficient than a turbojet. But like that analogy, it’s not advisable to use this configuration at high speeds because it’s no longer energy efficient to have accelerated all that reaction mass with you. There’s an optimum mission velocity for this sort of thing. Early in flight, it’s optimum to do this. Later in flight, it’s more efficient to operate at higher exhaust velocity (even though the thrust per unit power will be lower at higher exhaust velocity), ie without thrust augmentation by passive propellant but instead nearer to stoichiometric.

I'm having a tough time thinking of rocket oxidizer as "passive"

Helium?  Sure.

Hydrogen?  Mostly, save for a bit of embrittlement.

CH4?   Sure, minimal coking. Of course one can run fuel rich on a Raptor so it's not an `advanced topic`.

Oxygen?  Hot oxygen burns nearly everything except for the magic rare fairly low temperature high density compounds in the Oxygen path of Raptor.

why not just go straight for Krypton if mass is what is needed.  It's pretty cheap, $1.40 per kg


Cool pricelist btw:   http://www.leonland.de/elements_by_price/en/list

[1]

It‘s like a jet engine with a larger bypass ratio, except the bypass air is oxygen that you’re carrying with you.

Yes, but this is a critical difference. The jet engine actually gets its oxygen/air for free, so it's free to haul ONLY fuel. The rocket doesn't have this option. It must either haul both, or else it hauls dead/passive mass.

Here's what I believe to be the point of contention:

There’s an optimum mission velocity for this sort of thing.

Be very careful here. I agree that there's an optimal distribution for a fixed amount of energy available within a vehicle, but I caution against using 'exhaust velocity as close to zero as possible' a your preferred optimization point. If changing reference frames suddenly makes it appear that efficiency has dropped because the exhaust velocity is no longer zero, then your thought experiment is in danger of leading you astray in the general sense. Kinetic energy, after all, is not a conserved quantity.

Let's change the thought experiment somewhat. Instead of strap-on oxygen tanks, imagine that you have strap-on potassium tanks (while you're at it, also imagine that the potassium somehow behaves like a fluid). Now what? Well, first, a similar result should occur in as far as you're injecting a higher density substance into the exhaust stream, trading lower ISP for higher impulse density.

But, now that elemental potassium is free to react very violently with both the fully combusted hydrolox (water) as well as any other free oxygen in the exhaust.  So, by reacting with the combusted exhaust products, you'll increase the total energy of the exhaust in the nozzle. Will this therefore increase the pressure in the nozzle and therefore thrust of the vehicle? It might, especially early in flight and if the engine exhaust is slightly overexpanded.

By doing the 2K + H₂O -> 2KOH + H₂ reaction, more of your total energy may get bound back into that free hydrogen. Will that also now increase your ISP? If so, then this may be less "efficient" from an exhaust velocity standpoint. But if your total available energy increases, and if a portion of that increased energy goes into your vehicle and allows it to get better performance as a result, then we might not care.

Hope that makes sense somewhat. Overall I agree with you that there's likely an interesting trade to be made; I just don't believe going super oxy-rich is optimal all things considered.

[Robotbeat]

It‘s like a jet engine with a larger bypass ratio, except the bypass air is oxygen that you’re carrying with you.

Yes, but this is a critical difference. The jet engine actually gets its oxygen/air for free, so it's free to haul ONLY fuel. The rocket doesn't have this option. It must either haul both, or else it hauls dead/passive mass. ...

So you'd think that'd be a big benefit for jet engines, but it's actually not. You DON'T get that for free, you have to slow down the air first, or rather bring it up to the speed of the aircraft. (...which is the same thing the rocket does, when you think about it!)

SCRAMjets can get away with not doing that as much, BUT the impulse you can get out of a given mass of air that you DON'T slow down first is much less than that which you do. (This is a sort of difficult point to intuit, I think, but it is absolutely true.) So the benefit is negligible.

It's just the pure gravity losses of the extra weight that you lose out on, but that's much less if you have a nice high-thrust first stage with high T/W ratio.

[Robotbeat]

...

Be very careful here. I agree that there's an optimal distribution for a fixed amount of energy available within a vehicle, but I caution against using 'exhaust velocity as close to zero as possible' a your preferred optimization point. If changing reference frames suddenly makes it appear that efficiency has dropped because the exhaust velocity is no longer zero, then your thought experiment is in danger of leading you astray in the general sense. Kinetic energy, after all, is not a conserved quantity. ...

I promise you the physics works on this. See Geoffrey Landis explain it at 18:20 here:

[redneck]

I seem to recall that much of the thrust gain from augmentation was in reducing the losses from over expansion. One cold flow experiment from years back was a small jet reverse flow into the over expanded nozzle. Kicked the flow over to the side and entrained ambient air into the nozzle to achieve an aerospike like effect in a conventional over expanded nozzle. Noticeable increase in thrust.  All low pressure stuff (<200 psi) with compressed air.