which is more fuel efficent??

[michaelphoenix22]

Just wondering which is better to do launch, circularize at 100 km then do an apoapsis changed to say 600km and circularize again, or is it better to launch straight to 600 km then circularize?

[Arsonik]

I don't think there would be any difference.

[Tank Buddy]

Much more efficient to get in a LKO, then do a Hohmann transfer. That way you waste less fuel going strait up, then coasting and loosing velocity. You would have gone from say, 1500 m/s when you cut your engines to somewhere like 800 m/s, so you would lose 700 m/s.

I stand corrected, apparently, there is no difference. Thanks goldenpeach.

Edited January 22, 2014 by Tank Buddy

[Sirine]

Is it better to launch straight to 600 km.

[goldenpeach]

I will do a (small) test using mechjeb to do the launch(because I use it and to eliminate any error that could be made by manually piloting the rocket).

Will edit my post to show the result.

EDIT:

delta-v used by launching to a circular 100KM orbit then transferring to circular 600KM orbit:5035 m/s

delta-v used by launching straight to a circular 600KM orbit: 5010 m/s

The result are about the same(the difference could come from non-optimal ascent path or the precision of mechjeb for ascent(or something like that). I don't think you should worry too much about that: use the one that is the most convenient for you.

Edited January 22, 2014 by goldenpeach

[Tokay Gris]

I would say it depends how you do your ascent.

Ideally and if you do a continuous gravity turn starting at, say 9km, and ending at something like 70, burning then parallel to the ground, this way rising apoapsis, it wouldn't do any difference.

Circularizing this way and then raising apoapsis would be just like pausing the burn and then continuing it again.

But if you do your ascent differently, say, burn straight up to 100, then circularizing, then raising the apogee, it would make a difference.

In theory (my theory, which isn't saying much), the "straight up, then circularize" would be more efficient the lower the first circularizing orbit is. Because once you have obtained orbit, you are not loosing delta-V to gravity.

So, put together: Since no ascent is really perfect, you end up between "no difference" and "less efficient" if you ascent to 600km directly.

Which means: Better to first make orbit low, then raise apogee.

Can anyone with more knowledge than me check this thinking? Maybe I am wrong, but it makes sense that way.

[goldenpeach]

I think I will do another test(tomorrow) but with more different orbit: launching in 75KM and going to 60 000KM and going straight to 60 000 to see if we notice a larger difference in delta-v usage.

[Soda Popinski]

I would expect it's more efficient to do a single burn and one circularization, due to the Oberth effect. With a single burn, especially with a high thrust rocket, you would do more of your thrusting while lower. On the other hand, if you thrust too much, you may lose delta-V to air resistance. Doing a launch of 75 km might actually muddy the data a bit, as you'll spend less time in vacuum before the 2nd burn.

[goldenpeach]

I made the test(with a similar ship used in the first test but with more fuel(the first one can only place the payload in orbit around 625KM):

delta-v used by launching in 75KM and transferring to 60 000KM:5592 m/s

delta-v used by launching straight to 60 000KM: 5541

I think we can say that launching straight to the desired orbit is more efficient

[Kasuha]

The best approach is to circularize as low as possible (even in atmosphere, e.g. at 60 km) and push your apoapsis to 600 km right away, then circularize at 600 km.

The reason is, launch is very inefficient maneuver because you're raising periapsis and apoapsis while you're at neither and to keep yourself at a relatively reasonable place you even burn off-prograde a lot of the time. The sooner you switch to Hohmann transfers the better.

I made the test(with a similar ship used in the first test but with more fuel(the first one can only place the payload in orbit around 625KM):

delta-v used by launching in 75KM and transferring to 60 000KM:5592 m/s

delta-v used by launching straight to 60 000KM: 5541

I think we can say that launching straight to the desired orbit is more efficient

I wonder how your direct launch profile looked like. Particularly what was your apoapsis when periapsis left the surface. MechJeb is smart.

Edited January 22, 2014 by Kasuha

[goldenpeach]

my launch ascent was:

turn start at 10KM

turn end at 70KM

the "shape" of the ascent was 45%.

Also, I made a similar test on the mun(it was an munar SSTO)(launch profile: turn start at 0KM, end at 100KM and has a "shape" of 5%) and I came with theses number(with infinite fuel on, to avoid any big difference in TWR):

using LMO and transfer: 898 m/s

using direct ascent: 887 m/s

I aimed for an altitude of 1900 KM, for the first try(LMO and transfer) I used a parking orbit of 20KM

Unfortunately, the difference is too small to come to any conclusion.

If you want the crafts file I can provide them.

EDIT: what do you mean by "your apoapsis when periapsis left the ground"?

[mhoram]

Interesting question. I tested it for a small 3-stage rocket with my ascent optimizer for fuel efficiency.

Here are the theoretical results.

From the initial 6056 Vacuum dV at the end were about 880 left. So they are in the same range as the other tests.

Edit: As one can see the periapsis is brought very fast to an altitude of -100km (below surface). This counters the gravity losses. After that, the thrust is reduced to nearly 0 (that is also the reason for the erratic up and down of pitch and AoA). Circularization is done at the end about 20 minutes after liftoff.

Javascript is disabled. View full album

Edited January 22, 2014 by mhoram

[Kasuha]

Here are the theoretical results.

Very interesting, especially the Distance and Eccentricity graphs. Most important things happen between 150~250 seconds when it both raises periapsis and apoapsis and decreases eccentricity below 0.4. After that it's mostly coasting to apoapsis where it circularizes.

I wonder if the target apoapsis was higher whether it would switch to increasing the eccentricity again, that single dent on the graph might be a fluke or it might be sign of regime change. I wonder if the simulation is not too coarse at around that time.

[mhoram]

I wonder if the target apoapsis was higher whether it would switch to increasing the eccentricity again, that single dent on the graph might be a fluke or it might be sign of regime change. I wonder if the simulation is not too coarse at around that time.

Unfortunately it is not so easy to get the best results. There are some additional parameters like number of simulation iterations and nodes in each stage. Changing these parameters can lead to worse results as in this case adding more nodes did.

About this dent, it is in my opinion more of a fluke - most other simulations had no dent.

[numerobis]

If simulating more means worse results, that suggests that you aren't converging yet, doesn't it? And so you'd need to add yet more? I'm not up to speed on optimization anymore, but that's definitely how it works for finite element simulations -- if changing the mesh changes the result, your mesh isn't fine enough.

[mhoram]

If simulating more means worse results, that suggests that you aren't converging yet, doesn't it? And so you'd need to add yet more? I'm not up to speed on optimization anymore, but that's definitely how it works for finite element simulations -- if changing the mesh changes the result, your mesh isn't fine enough.

Or the number of iterations is not high enough. Problem is that increasing the number of iterations causes in some cases problem with one of the used libraries or the memory size.