Air Launching Rockets

[Jonfliesgoats]

Here's a NASA document reviewing the costs and benefits of air-launching rockets:  Its short, to the point and informative.

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20140003206.pdf

[wumpus]

This isn't even "dropping stuff".  It  is launching the thing from on *top* of the aircraft.  Missing the tail of the aircraft is a critical feature of ejection seats (and I think the first US/UK seat (the Germans had been building them for awhile), still badly injured the pilot when he hit the tail.  So do you do a "fire in the hole" launch when you fire "clear the tail" stage to separate? (getting the CoM/CoP right for both the combined and individual aircraft appears unlikely).

Apparently there is no effort to try to keep the H2 second stage topped off.  Any guesses as to how much will be left when it stages?

The rocket gets refueled with RP-1 in flight.  This is reasonably plausible as draining a tanker of JP-whatever and replacing it with RP-1 is probably a non-issue (you might fill the thing up with an alcohol cleaner or not, I suspect it won't need any real cleaning).  This takes things out of the hands of commercial flight, but I'm sure NASA/Air Force/Navy can handle it.

The rocket also gets refueled with LOX.  Right.  At this point I'm really surprised at the lack of LOH refueling, as that is probably a bigger issue and roughly equally difficult.

Somebody is trying to take kerbal designs and use them in real life [hello Eric D. Waters/Dennis M. Creech/Alan D. Philips!, one of you is on here, I'm sure] .  Redo it after you install the "real fuels" mod.  Also using the staging to fire both decoupler and seperatron together is cheating.  You need to hit the space bar twice in succession to fire them.  Ideally you should have some randomly decide on the order and cover up that bit of screen with tape when you launch.  And whatever you do, don't build a real rocket off results from KSP, even with RO.

[Jonfliesgoats]

You know, it's amazing how Kerbal real life gets.

Check the Maya and the Mercury here:

http://www.aviastar.org/air/england/short_mayo.php

 

[wumpus]

Except that Maya and Mercury both use propellers.  In this proposal you have a rocket on top of an jet, which will want to flip the combined craft when the rocket is ignited while they are coupled and the rocket will slam into the tail if they are decoupled without igniting the rocket.  To be honest, refueling inflight is pretty kerbal already.  But refueling inflight with LOH just doesn't sound too likely.

I'm still shocked that they didn't take more care that the Maya's tail was shorter than the Mercury.  Presumably too many other things could go wrong to bother dealing with all the issues involved in shortening the tail.

[magnemoe]

43 minutes ago, wumpus said:

Except that Maya and Mercury both use propellers.  In this proposal you have a rocket on top of an jet, which will want to flip the combined craft when the rocket is ignited while they are coupled and the rocket will slam into the tail if they are decoupled without igniting the rocket.  To be honest, refueling inflight is pretty kerbal already.  But refueling inflight with LOH just doesn't sound too likely.

I'm still shocked that they didn't take more care that the Maya's tail was shorter than the Mercury.  Presumably too many other things could go wrong to bother dealing with all the issues involved in shortening the tail.

Yes speed is lower and you set top plane to have more lift and higher speed than the lower on separation, your only issue is to get an clean separation, 
Idea has been used later, Soviet used bombers to lift fighter jets. 

At high speed things is more difficult. Topping up the tanks on the rocket would be one of the easiest tasks, problem is that this would require lots of modification on the plane. 
As you say the rocket would not have the same flight path as the plane.
The SR71 drone was an plane so you could get an good flight trajectory before release. 
For an rocket on roof best way would be to release it then you was on an ballistic trajectory, rocket follows it and you move away.
Since you need an specialized plane anyway I would put rocket below and used high landing gear, optional special one you dropped on takeoff.  

[wumpus]

I didn't consider throttling down the jet engine to roughly zero thrust, if the drag of the 747 is higher than the rocket, it might separate.  The idea of "launch at idle" seems weird in a battle of thrust vs. gravity.  I still wouldn't like anything threatening the tail of my aircraft, I don't think I'm landing without it.  I also don't see Stratolaunch rushing to push this idea, which would certainly make more sense than trying to launch 3 pegasus craft in one flight.

The SR-71 (M-21 Drone) isn't a very good example. The plane was destroyed and the launch control officer was killed during a test flight of the drone and the whole idea was scrapped.  Presumably it moved to the B-52 (and much lower launch speeds), but it didn't seem to last long there either.  It is pretty much a clear example of "separation is hard".

I'd be curious if the original stratolaunch plan of "welding together two 747s" wouldn't be easier than extra long landing gear or heroic efforts for a top-mounted launch.  Presumably stratolaunch's switch to a "full custom" plane shows just how difficult the "weld two 747s together" idea really is.  

[Nibb31]

Like most innovative launch proposals, this one is shot down by economics more than by technology.

With air launch, your aircraft is basically a very expensive and very low performance first stage: a first stage that only gets you to a low Mach number at a relatively low altitude. Most first stages get the rocket to something around Mach 10 at 70-100km altitude. So the air launch saves you maybe 10% in the size of your rocket. 

Your aircraft also has to be pretty much a one-off design, which means that instead of spreading development and tooling costs over a large production run, you only get to build 1 or 2 aircraft. Each one gets expensive to operate and maintain, because spare parts are difficult to source, and with for example one or two launches per month (which is highly optimistic), it still spends most of its time in its super-huge custom built hangar waiting for rockets and payloads to be built.

Add in the lack of flexibility (very few airports will be equipped to fuel rockets and land such a large aircraft), failure modes (landing a fully fueled rocket after a scrub for example), and the increased weight of a rocket that can withstand vertical and horizontal structural loads, and it quickly appears that it's not worth the risk or complexity.

In the end, it simply makes sense to do away with the air launch, and just build a rocket that's 10% bigger.

And yes, I know, Paul Allen doesn't agree with me.

Edited December 6, 2016 by Nibb31

[insert_name]

Air launch is only practical for really tiny rockets. The Pegasus total mass is less than the max payload of the Atlas 5. Tdlr air launch is good for microsatellites in Leo. Not much else.

[wumpus]

5 hours ago, Nibb31 said:

In the end, it simply makes sense to do away with the air launch, and just build a rocket that's 10% bigger.

10% bigger implies almost no delta-v change (and I think launching at ground instead of mach 3 adds 100%, but nobody is talking about mach 3).  But even adding 50% more fuel (and thrust) to a rocket is nothing compared to building an albatross like stratolaunch (Whiteknight was dealing with a rocket that needed a few thousand m/s, not 9000+ like a satellite launch).

The linked article suggested 50% more cargo on a rocket  20% lighter than a falcon 9.  But it still has all the advantages of being a "paper rocket", not to mention some serious aircraft issues.  My guess is that if rockets landed a bit more reliably [think Musk's "thousands of flights per booster"], you might stick a jet engine in for stage 1 (no airplane, just stick the engine under the booster).  Until then, rockets provide more thrust for less cost.

[magnemoe]

5 hours ago, insert_name said:

Air launch is only practical for really tiny rockets. The Pegasus total mass is less than the max payload of the Atlas 5. Tdlr air launch is good for microsatellites in Leo. Not much else.

Yes, at small sized you have multiple benefits, Drag is more of an issue for smaller rockets and an air launch let you vacuum optimize all the stages. and you can use standard planes reducing the cost of the first stage. If you payload is very small you can use an fighter jet for first stage for an small extra boost. 

Main issue is probably lack of marked, looks like most small satellites is either testing systems in space or science stuff in space, an secondary payload slot is far cheaper for this as you don't care about orbit.
Note if you get low orbit satellite swarm networks this will change and you want the ability to replace satellites in random non standard orbits. 
This is an marked 

[Jonfliesgoats]

Air launch of small stuff seems favorable.  There is a company using airplanes as small as Gulfstreams to launch very small payloads.  

With regard to the Maya and Mercury I posted that as an example of Kerbal engineering, not to draw parallels with anything NASA was working on.  Eight engines, a piggy back seaplane and in-flight separation were worked out just to get 700lbs of payload from the U.K. to Canada in 1938!  Talk about some expensive air-mail!

Incidentally, the first drawings for what would become the Lockheed Constellation were made in 1937.

[Jestersage]

For those that complained about those...

http://www.astronautix.com/s/spiral50-50.html <- Done at Mach 5!

https://en.wikipedia.org/wiki/MAKS_(spacecraft)

[Kryten]

For very small launchers, air launch can give you very large performance increases. The EV-2 Caleb would've put up 7kg from about 1400kg gross weight; and later this year or early next, the ground-launched SS-520-4 is to put up 3kg from ~2600kg gross weight. That's about four times the efficiency, and it would be even higher if you had a Caleb-alike with modern fuel.

[YNM]

Looks like those engineers have just done some porn on MAKS XD

But yeah, while this all looks bonkers, actually the method is a bit efficient. There's Pegasus vs. Taurus - Pegasus, while can only launch a quarter of the payload, also mass a fifth of Taurus.

Edited December 7, 2016 by YNM

[magnemoe]

56 minutes ago, Jonfliesgoats said:

Air launch of small stuff seems favorable.  There is a company using airplanes as small as Gulfstreams to launch very small payloads.  

With regard to the Maya and Mercury I posted that as an example of Kerbal engineering, not to draw parallels with anything NASA was working on.  Eight engines, a piggy back seaplane and in-flight separation were worked out just to get 700lbs of payload from the U.K. to Canada in 1938!  Talk about some expensive air-mail!

Incidentally, the first drawings for what would become the Lockheed Constellation were made in 1937.

The Maya and Mercury system makes sense, an problem with seaplanes is that they have far lower maximum takeoff weight that the standard version of the plane as water has higher friction than wheels, the pontoon are also heavier, Lifting the small plane up to cruise attitude would allow far longer range. 
Some seaplanes has wheels on the pontoons to take of from an airfield with lots of fuel and supplies, fly to remote location and unload, then fly lighter plane back 
Fails if hunting or fishing has been very good

Today its not uncommon for fighter jets on long missions to do mid air refueling after reaching cruise attitude. Less kluge than Maya and Mercury but mid air refueling was not standard until well out in 1950 even if done earlier. 

[cantab]

"Both designs utilized a Boeing 747-400F as the carrier aircraft...All concepts, when fully loaded, exceeded the allowable Gross Takeoff Weight (GTOW) of the aircraft platform."

That's not very encouraging. You're taking one of the largest and highest-capacity aircraft available, and it's still not enough.

As I see it the big advantage of subsonic air launch is operational and infrastructure simplicity. No big specialist launchpad, just any decent airport. No worries about weather at the rocket launch location (though obviously bad weather could ground the plane). Operate out of the temperature latitudes but actually launch the rocket much closer to the equator. But if you're proposing to use cryogenic fuels and aerial refuelling, that's adding a ton of complexity, hazards, and consequent cost. There are good reasons aerial refuelling isn't used in civil aviation, and that's even before we consider the extra problems of handling cryogenics.

Pegasus is well-established, and actually one of the most-launched orbital rockets. 23 ton rocket putting around 500 kilos into LEO. Pegasus II was much more ambitious, but so much so that it was to require a brand new aircraft. That's a major extra development cost. Let's say we instead use the highest-payload "off the shelf" aircraft we can have, a 747-8F with a maximum payload of around 140 tons, then scaling up the Pegasus design we're looking at a 2-3 tons to LEO payload. Meanwhile in that same original paper, the "PD-2" design without using aerial refuelling but using RP-1 and LOX rather than solids gets about 6 tons to LEO.

That I think is about as big as is worth going with air-launch.

[Nibb31]

Not "any decent airport". You need an airport that can handle an oversize aircraft (many airports had to do major changes to support the overhang of the A380 because of the risk of ingesting grass and dirt in the engines). This is an even bigger challenge for something as wide as Stratolaunch.

Also, you need an airport that can handle LOX loading and whatever your rocket runs on. This includes safety perimeters around the rocket at all times for loading, takeoff, and contingency landing (you probably need a system to dump fuel safely in case of a scrub).

You need an airport that has payload integration facilities, which means a large cleanroom facility with provisions for hypergolic fuel storage and handling.

So basically, once you add in the installations that need to be specially built or modified, you end up with a very limited number of airports.

[Kryten]

54 minutes ago, cantab said:

 Pegasus II was much more ambitious, but so much so that it was to require a brand new aircraft.

Pegasus II was built around Stratolaunch, not the other way around.

44 minutes ago, Nibb31 said:

You need an airport that has payload integration facilities, which means a large cleanroom facility with provisions for hypergolic fuel storage and handling.

You don't need to payload integration at the same airport as the launch. Pegasus does integration at Vandenberg then does a ferry flight to whatever airport is best for the launch, and LauncherOne is set to do the same thing with integration at VG's hangar in Mojave.

Edited December 7, 2016 by Kryten

[mikegarrison]

On 12/5/2016 at 10:38 AM, wumpus said:

Missing the tail of the aircraft is a critical feature of ejection seats (and I think the first US/UK seat (the Germans had been building them for awhile), still badly injured the pilot when he hit the tail.

There was at least one airplane that ejected the crew through the bottom of the airplane, but that was a bit of an issue during low altitude or ground ejections.

[wumpus]

Three days till Pegasus XL/CYGNSS (air launch).
[it isn't quite dead yet.]  And the economics work better when you don't have to redesign an aircraft and you launch smaller rockets.

[monophonic]

32 minutes ago, mikegarrison said:

There was at least one airplane that ejected the crew through the bottom of the airplane, but that was a bit of an issue during low altitude or ground ejections.

Ah yes, the early F-104 variants. There is an anecdote about that. Basically, at one air force or another that bought the type, just when the pilots had learned to roll the plane upside down before ejecting, they replaced them with later versions. Ones which featured upwards firing ejection seats. You can guess where the first few low altitude ejections in the new variant went...

[AeroGav]

Skimming this PDF,  I got the impression it was just a research paper by a couple of interns rather than a serious study.  At least I hope so.  A bit of back of the envelope optimisation of the rocket stage.

In a lot of ways it's profoundly depressing that we're talking about taking retired 1960s airliners to carry rockets powered by 1960s cryogenic engines (RL-10) in 2016.

Also the entire focus of the study was clearly to get the largest possible payload to orbit, impress everyone that they can compete with a Delta Heavy.   Very NASA !

Instead of actually trying to make the economics of re-usable launchers work,  which would mean targeting the modal launch payload mass to get as many launches per year as possible to try and offset the costs of developing your re-usable stage.   Instead they are talking about taking off with  the rocket empty and having the 747 (with enormous rocket on it's back) to rendezvous with (multiple?) air to air refuelling tankers ,  just so they can get past the maximum takeoff weight restriction of the 747 and show the SLS guys we can launch heavy manned deep space payloads too !   

I don't know if it's true or not, but i once heard that after landing the Space Shuttle , it would be more economic to send it away to be crushed and build a new one, than have the army of highly paid technicians in hazmat suits (UDMH, NO4 not too nice on the skin) crawl over it with a fine tooth comb for months and certify it safe for another flight.      How many billions did they spend developing the Ford Focus?  If they only manufactured five of them,  and each car made only 25 journeys before getting scrapped or totalled, the economics wouldn't look so good either.

 

 

[magnemoe]

18 hours ago, monophonic said:

Ah yes, the early F-104 variants. There is an anecdote about that. Basically, at one air force or another that bought the type, just when the pilots had learned to roll the plane upside down before ejecting, they replaced them with later versions. Ones which featured upwards firing ejection seats. You can guess where the first few low altitude ejections in the new variant went...

Ejections seats was designed to avoid the tail, back during WW2, pilots had to jump out of planes this did not work well with jet planes because of higher minimum speed so they got ejection seats. 
The first versions was more like cannons than rockets, modern ones are pretty smart, as I understand they gimbal well and will rotate the seat if you eject upside down, they can also eject on ground. 

Russia has an helicopter with ejection seats, they cut the blades of the rotor just before ejecting, ejecting down would be an bad idea on an combat helicopter who often fly low. 
 

[wumpus]

1 hour ago, AeroGav said:

Instead of actually trying to make the economics of re-usable launchers work,  which would mean targeting the modal launch payload mass to get as many launches per year as possible to try and offset the costs of developing your re-usable stage.   Instead they are talking about taking off with  the rocket empty and having the 747 (with enormous rocket on it's back) to rendezvous with (multiple?) air to air refuelling tankers ,  just so they can get past the maximum takeoff weight restriction of the 747 and show the SLS guys we can launch heavy manned deep space payloads too !   

In any air launch system I've seen, the means to get in the air is extremely reusable (balloon lift may be an exception).  The entire system would likely be more reusable than the shuttle system.  One thing to remember is that NASA has flown at least one space ship (shuttle orbiter) on top of a 747 and presumably still has it (possibly mothballed).  Converting the hard point to decouplers would be an issue, but I've claimed the real issue would be separation (possible if you idle the 747's engines).

I doubt that filling the RP-1 (in the rocket and the plane) is much of an issue.  It seems to be standard procedure for both the Air Force and Navy.  Filling the LOX is another story (presumably it leaks out too fast), and I really have to wonder if they need to refuel the LH2 from inside the plane.

While it might not have been interns, I'm pretty sure it was a junior bunch.  Certainly it helps for NASA to have studied all the issues, just in case they miss something pretty important.  Although I wouldn't be too surprised if some of the bits (like having to match Delta-Heavy) were sent down from on high to make sure that it was effectively impossible.  I'd have liked a project manager to tell them to go back and compute the size they would get without in flight LOX refueling and this type of thing makes me wonder if it wasn't to show how impossible air launch really is (it might work for smaller, but current tech seems to be doing ok there as well).

It isn't that far from Orbital's original recipe.  They have a plane with hardpoints (just ignore the tail in the way).  They have the means to lift (just ignore the LOX refueling).  I'm sure there was a herculean effort by the Orbital people to get that idea in space, and NASA seems only willing to write a white paper.  Considering the "details" above, that's probably for the best (don't forget that Pegasus is solid rocket, at least stage 1).

[Jonfliesgoats]

Great observations, all!

This is a company with a proposal to launch cubesats from F-104s.  I think they are associated with Starfighters Inc.  They are the smallest scale air launch to orbit program I can find with Google.

http://cubecab.com/

Edited December 9, 2016 by Jonfliesgoats