Experiment with airship parts: optimal launch assist

[kahlzun]

Hi all!

0.20 has wiped a lot of my modded parts, and so I'm starting to rebuild my omni-fleet, which is designed to travel to (and/or) return from anywhere.

Vital to the success of the Eve launch is the Hooligan Airship mod which adds buoyant parts, which allow you to climb out of atmosphere cheaply.

I'm conducting a set of experiments to determine the optimal ascent profile when using this system, whether it is preferable to eject the parts when they are not providing lift (as they also have high drag), or to keep them on the ship as long as possible (as it effectively reduces vehicle weight, increasing TWR).

I'll post my results and experimental conditions on here as they come up, watch this space..

Edited May 25, 2013 by kahlzun

[kahlzun]

Experiment 1:

High-efficiency, medium-thrust engine.

Craft constructed as follows:

1x OKTO probe core

3x RTG

3x TT-38K radial decoupler

3x Envelope 'Ray'

1x T400 fuel tank

1x Aerospike

Experiment Info:

'Ray' deployed to 100% from KSC.

0.0 bouyancy passed at about 15300m

Apoapsis (0 m/s) at about 16650

Ascent profile for Mechjeb:

Turn Start: 10km

Turn End: 70km

Inclination: 0deg

Turn Rate: 40%

[table]

[tr]

[th]Methodology (Kerbin Test #1)[/th]

[th]Result (Test 1)[/th]

[th]Result (Test 2)[/th]

[/tr]

[tr]

[td]Balloon is maintained until the bouyancy is 0.0, then Mechjeb Ascent to 100km[/td]

[td]Activated at ~15300m.

Height: 100km X -147km

+-20m[/td]

[td]Activated at ~15300m.

Height: 100km X -147km

+-30m[/td]

[/tr]

[tr]

[td]Balloon is maintained until the vertical speed is 0.0, then Mechjeb Ascent to 100km[/td]

[td]Activated at ~16650m.

Height: 100km X -164km

+-15m[/td]

[td]Activated at ~16650m.

Height: 100km X -165km

+-215m[/td]

[/tr]

[tr]

[td]Balloon is maintained until the vertical speed is 5, then Mechjeb Ascent to 100km[/td]

[td]Activated at ~16600m.

Height: 100km X -155km

+-15m[/td]

[td]xx[/td]

[/tr]

[/table]

Plan was initially to test efficiency of keeping 'inflated' modules on the side of the ship during ascent, but probe was unable to maintain direction control during ascent.

Next experiment: NERVA engine variant.

Edited May 25, 2013 by kahlzun

[magnemoe]

I have done plenty of balloon launches, I prefer to activate engines then speed falls to 10m/s, had some stability problems close to 0, simple reasoning, you can climb around one km after you get zero lift, I then drop the balloons a bit over an second later.

It might be beneficial to hold on to the balloons for some time, mechjeb should compensate trust to terminal speed.

Did some quck tests with an realistic Eve accent craft with 6000m/s dv and 28 ton using 6 balloons.

It got 0 lift at 11472 meter, speed dropped to 10m/s at 12598 where I activated accent autopilot.

Dropping balloon after a second gave me 3845 m/s left in 100 km orbit.

Then I kept the balloon until reaching 14665 meter and around -70 in lift I was left with 3890 m/s

Main downside of keeping it longer was that the rocket started to get unstable, had I dropped it a second earlier it would probably saved some fuel.

Edited May 25, 2013 by magnemoe

[kahlzun]

Experiment Profile #2:

Identical setup to previous, however aerospike is swapped out for LN-1 'NERVA' engine.

0.0 Bouyancy height: 14700m

0ms height: 16074m

Test results (same methodology):

[table]

[tr]

[th]Methodology (Kerbin Test #2)[/th]

[th]Result (Test 1)[/th]

[/tr]

[tr]

[td]Balloon is maintained until the bouyancy is 0.0, then Mechjeb Ascent to 100km[/td]

[td]Activated at ~14700m.

Height: 100km X 100km orbit, 974 ms remain in tank

+-50m[/td]

[/tr]

[tr]

[td]Balloon is maintained until the vertical speed is 0.0, then Mechjeb Ascent to 100km[/td]

[td]Activated at ~16000m.

Height: 100km X 100km orbit, 856 ms remain in tank

+-351[/td][/tr]

[/table]

Edited May 25, 2013 by kahlzun

[kahlzun]

My conclusion from these initial launches is interesting, as it contradicts what i was expecting to witness:

I was expecting the launches that started from a higher point (ie: the 0ms launches) to be more efficient, as they have less gravity and a thinner atmosphere,

but they were reliably and consistently shown in tests to be less efficient than a 0.0 bouyancy launch.

POSSIBLE ISSUES:

*As the launch profile selected started its gravity turn at 10km, all launches were well above the turn start, and this may have contributed.

*Probe core chosen has issues maintaining heading when not travelling precisely prograde, especially with the balloons mounted.

[kahlzun]

Did some quck tests with an realistic Eve accent craft with 6000m/s dv and 28 ton using 6 balloons.

It got 0 lift at 11472 meter, speed dropped to 10m/s at 12598 where I activated accent autopilot.

Try dropping the balloons when you pass the 0 lift, (11472m) point, see if it makes a difference. My results here seem to show earlier drops give noticeably better outcomes.

[kahlzun]

As an unrelated side effect of this testing, I have figured out an ultra light-weight solution to all your 'heavy pod' problems:

Little dude survived reentry (IN HIS FACE) and held onto the rocket during a launch hard enough to generate further entry effects.. unfortunatly, it got stuck suborbital, and, well...

he went splat.

But, as a proof-of-concept, it works great!

[magnemoe]

Try dropping the balloons when you pass the 0 lift, (11472m) point, see if it makes a difference. My results here seem to show earlier drops give noticeably better outcomes.

You are right, this leaves me with 3960 m/s left, counter intuitive for me as I guessed the extra 1 kilometer in attitude was more important than the 50 m/s you kept.

But as you say the gravity turn stared before the balloon was dropped.

Main bonus here is probably the stability, waiting for speed to drop to 0 or keeping the balloons to long can easy give you stability problems.

[kahlzun]

I'd like to see if more people can try this, see if we can get the 'sweet spot' for assisted launch..

[FluxCap]

I've been trying this with a base of blimps with jets and nuclear rockets that forms a ring around a full orange tank and a couple of large rcs tanks, along with a second stage with rockets: trying to experiment getting cargo to orbit using airships the most efficiently.

I've been running into a lot of issues, but so far the above is the most successful design. I think the blimp base is fairly solid, but the second/cargo needs work and tweaking. Using jets to aid the balloons seem effective. Going 90 deg pitch until ditching the balloons I think is the way to go... I'm also testing "pumpkin seed" type designs, but they are rather funny looking at the moment. Quite embarrasing to share, even more so then my spammed balloon ring below.

Both sections have pods, so the idea is to let the blimp base float there while taking the second rocket stage to orbit, then float it back down. The pumpkin seed experiment didn't ditch the airship elements, and yes, had major issues with drag. Thing is, keeping them would help save fuel landing. You'd just have to do a de-orbit burn and then float along, using jets to guide it into KSP...

Some screens of the blimp circle of fire:

Edited June 4, 2013 by FluxCap

[FluxCap]

This is Mk2, and just started it...

Anyway, I based it on some things found here: http://www.jpaerospace.com/atohandout.pdf

Mk2:

[FluxCap]

Mk3 of "LTA Lifter":

Craft file: http://db.tt/3mt3Ltnt

Delivered cargo: 566 monop, 1440 liq. fuel, 1760 oxidizer (567 monop, 1700 liq. fuel, 2078 oxi w/attached leftover stage) 85K circular orbit obtained.

450 liquid fuel spent using jets and balloons to climb to 11K altitude.

1500 monop, 8010 liquid fuel, and 9240 oxidizer for all stages.

It barely meets the needs of some personal requirements I have...but, was so happy to finally get something working.

Screens:

[FluxCap]

Ascent procedure for LTA Lifter Mk3 (using Mechjeb):

Initial setup: SURF @ 90 HDG, 90 PIT. Use RCS at willl to stabilize launch. 75% buoy, 50% throttle. Fire jets using space bar. Space bar again to release clamps.

2000m: 85% buoy, 3000m: 100% buoy.

Between 5k and 6500m: increase throttle to maintain positive vertical speed. Example: 5000m: 65% throttle. 5500: 70% throttle. 6K: 80% throttle. 6500: 97% throttle. Engage RCS.

Depending on how efficient the climb is, it should be between 8600m - 11K when the jet fuel runs out and then to drop the balloons, set HDG to 80 deg, and let the craft move slightly away from the drifting balloons, and pitching down toward the 80 degree target.

Space bar to drop the "jet pack" and engage aerospikes/orange tank lifting section.

Between 17K - 27K the next stage is engaged, and ascent autopilot (or manual gravity turn). That should take the craft, at least, to an 85K orbit.

I think tweaking some fuel tanks and such can reach my goal of 100k. I wanted a reusable craft that didn't leave (much) debris, but so far this is the closest I have reached.

Edited May 31, 2013 by FluxCap

[kahlzun]

Mk3 of "LTA Lifter":

Some engines won't provide thrust if you have something in their exhaust stream, even if it survives the heat. I can't tell too well from the picture, but if the engine pods are all in a vertical line, try rearranging them so the lower ones are a little closer to the craft..