Why are parking orbits more efficient?

[klappertjes]

I have a question about correct/efficient ways to go places. I'm taking going to the Mun as an example here because it's straightforward for what I'm asking but this can, I assume, be applied to pretty much any destination.

I see three ways to do this:

1) Launch. Get into ~80-100km parking orbit. Fumble with manouever nodes until you get a nice encounter with the Mun with low peri and use it.

2) Wait until the Mun is in the right place of the sky (25 degrees above the horizon?). Launch. Fly straight into your Munar transfer.

3) Wait until the Mun is somewhere suitable. Launch straight up to apo at 11Mm = Mun encounter

I tend to use number 1), because I've been told (mostly by Scott Manley videos) that this is the most efficient way to do it. Sometimes Manley will do it differently, and remark that [imagine Scottish accent] "... this is of course very inefficient". I trust his judgment. But I don't understand why

If I get the Oberth effect correctly, you want to burn close to the planet as much as you can. Then surely, merging orbital insertion and Munar transfer (number 2) means you're doing most of the burning while still ascending, thus burning close to the surface and using Oberth to its fullest.

Similarly, using method 3 would do this even more, in my head. Maybe it's a little silly for the Mun, but certainly for interplanetary transfers, launching straight up at sunset or sunrise makes sense to me. Why waste fuel getting in orbit first?

I'm almost certain I'm wrong in both cases. Could someone explain to me what I'm missing here and help out a troubled soul?

Edited April 12, 2014 by klappertjes

[Rakaydos]

Unfortunately, it's not "close to the planet" you want, it's "at highest velocity."

Combining Munar transfer with orbital insertion means you're fighting kerbin's gravity the whole time you're making the transfer burn, and you're making the transfer burn while at lower velocity.

By waiting until you've made orbit, you can burn sideways, ignoring kerbin gravity, while oberthing at kerbin orbital velocity.

[vexx32]

Launching straight up and burning continuously without doing a gravity turn (which is what I think you mean by 3) means you're constantly fighting gravity. Going into a parking orbit first means you can use the momentum you've already built up and you needn't fight gravity for the remainder of the burn, making it more efficient.

I'd need to do math to prove that, but I haven't slept all night, so I'll leave it up to another to either correct me or back me up as necessary.

[Tex_NL]

The Oberth effect is not about making your transfer as close to the planet as possible. It is making your transfer when you're traveling the fastest. As it happens you're traveling the fastest when you're in a low orbit, close to the planet. In a low orbit you're often traveling faster then when you were during launch.

Edit: Double ninja

[PUNiSH3R]

I see three ways to do this:

1) Launch. Get into ~80-100km parking orbit. Fumble with manouever nodes until you get a nice encounter with the Mun with low peri and use it.

2) Wait until the Mun is in the right place of the sky (25 degrees above the horizon?). Launch. Fly straight into your Munar transfer.

3) Wait until the Mun is somewhere suitable. Launch straight up to apo at 11Mm = Mun encounter

1. Parking orbit is more efficient if you're not computing transfer angles; you can brute force an encounter, but you'll burn massive dV depending on the angles. Parking orbit is a compromise position; technically you don't need the >72km periapsis to get there, but it makes transfers easier once you're used to it and you're automatically moving faster, so the Oberth is more pronounced.

2. Mun in the 'right place' is on the horizon (90 degree intercept angle). This is what I usually do. Makes a gravity assist off the Mun for interplanetary transfers way easy as well.

3. Declines any benefit of the Oberth Effect, see Vexx above.

EDIT: Holy guac, QUAD LASER NINJA...

[No one can defeat the quad laser...]

Edited April 11, 2014 by PUNiSH3R

[Taki117]

With Method 3 you are using much more dV fighting gravity. As soon as you stop firing your engines, if you don't have any horizontal velocity, will result in you falling back to Kerbin. If you do manage to get a Munar Encounter you still need to have enough fuel to land and get back.

With Method 2, if you miss your launch window you have to wait until the next launch window, which wouldn't be for another Kerbin month.

With Method 1, if you miss your launch window it is a very short wait (No more than about 30 earth minutes depending on altitude) not to mention you aren't wasting lots of dV fighting gravity and aerodynamics.

[Tex_NL]

...

EDIT: Holy guac, QUAD LASER NINJA...

[No one can defeat the quad laser...]

Don't worry, it wasn't a quad ninja. It was just a triple ninja.

...

With Method 2, if you miss your launch window you have to wait until the next launch window, which wouldn't be for another Kerbin month.

...

Don't you mean a Kerbin day, a mere six hours?

[Jouni]

The most efficient way is probably to ascend normally, but continue burning prograde until you get the apoapsis high enough. The timing is hard to get exactly right, so in practice it may be more efficient to first circularize to a low orbit, and then do another burn at the right time to the right direction.

[Dunbaratu]

A parking orbit is NOT more efficient. It's just a heck of a lot easier to do the math. (I use them a lot in my kOS missions for that reason).

Of your three options, #2 is the most efficient (although your figure of 25 degrees is way off. You want the angle between the longitude of the KSC laucnhpad and the longitude of Mun (if you draw a line from Mun down to Kerbin and imagine the logitude where that hits Kerbin) of about 110 to 120 degrees at T zero of your launch. It's hard to be more precise than "110 to 120 degrees" because the exact number depends on your choice of how you pilot your ship into orbit. Usually if you start at 110 to 120 degrees, then you'll be hitting the transfer point at about 90 to 100 degrees, which is the right spot.

(EDIT: The boldface bit below is the real reason parking orbits are usually less efficient.)

What is more efficient is to act "as if" you were going to do a parking orbit, but instead of that bit where you shut off the engines at an altitude of about 50-60km and coast to apoapsis and then turn them on again to finish circularizing, you just keep burning anyway, raising your apoapsis very high on the other side, stretching it up to the Mun. The result is an elliptical transfer orbit who's periapsis behind you can be low enough to actually dip into Kerbin's atmosphere. That's what makes it more efficient. If you go to a parking orbit first, then you can't get a transfer orbit with a periapsis that low. An orbit with apoapsis at 12 million meters and a peirapsis at 80000 meters takes more energy to achieve than one with an apoapsis of 12 million meters but a periapsis of 50000 ro 70000 meters.

Don't aim TOO low because then you're still fighting atmosphere, but your speed at a circular orbit of 60000m is higher than your speed at a circular orbit of about 80000m. By aiming to go from 60000m to the Mun, as opposed to going from 80000m to the Mun, you are aiming to use a higher speed at your transfer point, thus taking advantage of the oberth effect, which is strong enough to more than compensate for the small drag caused by the thin air of 60000m.

The reason real space programs use parking orbits is because the loss of efficiency of that versus a direct burn isn't *that* large, but a parking orbit gains you the ability to run tests and diagnostic checks before you pass the point of no return and leave Earth orbit. i.e. For Apollo it allowed time to make sure the craft was set up right and configured properly before pushing on to the moon. Also, it puts you more in control with the precision to decide when to make the burn.

Edited April 11, 2014 by Steven Mading

[Amaroq]

The "most efficient" approach is to burn while at periapsis. Which means, a very short prograde burn at periapsis, shut off, loop around, do it again .. until eventually, many orbits later, you'd have your intercept (or escape). Nobody does that -- KSP players because it would be terribly boring, and real programs because the mass required for the extra life support would exceed the gain.

[Dunbaratu]

The "most efficient" approach is to burn while at periapsis. Which means, a very short prograde burn at periapsis, shut off, loop around, do it again .. until eventually, many orbits later, you'd have your intercept (or escape). Nobody does that -- KSP players because it would be terribly boring, and real programs because the mass required for the extra life support would exceed the gain.

It's also impossible to achieve that optimization. The optimal use of the method you're talking about would require dividing the burn into an infinite number of individual burns of duration zero seconds each, which is obviously impossible. So how close to that do you want to get? 60 orbits of 10 second burns each? 30 orbits of 20 second burns each, 10 orbits of 60 seconds each? 1 burn of 5 minutes? It's not just KSP players that refuse to optmize it all the way. EVERYone does because it's not practical to take an infinite number of orbits to accomplish the burn. It's always a trade off between boredom and practicality versus efficiency, even in real space programs. Even the recent probe from India that uses this technique only does it to the extent of using 4 iterations of the method, rather than say 10 or 20 or so.

[Mesons]

A parking orbit is NOT more efficient. It's just a heck of a lot easier to do the math. (I use them a lot in my kOS missions for that reason).

...

What is more efficient is to act "as if" you were going to do a parking orbit, but instead of that bit where you shut off the engines at an altitude of about 50-60km and coast to apoapsis and then turn them on again to finish circularizing, you just keep burning anyway, raising your apoapsis very high on the other side, stretching it up to the Mun. The result is an elliptical transfer orbit who's periapsis behind you can be low enough to actually dip into Kerbin's atmosphere. That's what makes it more efficient. If you go to a parking orbit first, then you can't get a transfer orbit with a periapsis that low. An orbit with apoapsis at 12 million meters and a peirapsis at 80000 meters takes more energy to achieve than one with an apoapsis of 12 million meters but a periapsis of 50000 ro 70000 meters.

Couldn't have said it better myself. One thing to keep in mind, though, is that when you have a lower periapsis, your orbital speed at apoapsis is less. This affects how much dV you will need to reduce your velocity in the Mun's SOI if that is your plan. The speed at apoapsis and a non-instantaneous burn are the reasons why your option #3 (burn radially the whole way) is less efficient. Even if you could do your burn very fast once you're above most of the atmosphere, when you reach apoapsis near the Mun, you will enter the Mun's SOI with its orbital speed, whereas if you had an easterly orbit your entering speed will be lower since you're moving the same direction as the Mun.

For interplanetary burns, if you can do your whole transfer burn high in the atmosphere in a matter of seconds, option #3 is the most efficient because no energy is wasted on gaining angular momentum. Launching at just before sunrise or just after sunset would be optimal (since you will have some small amount of angular momentum thanks to Kerbin's rotation). Unfortunately, since interplanetary transfer burns tend to take orders of magnitude longer than a second, option #3 is virtually never the best choice. The problem is you are continually losing kinetic energy as you move farther out from Kerbin's surface, so your benefits from the Oberth effect are continually decreasing. You would need to finish your whole burn before reaching your high atmosphere end-burn altitude (from option #2) in order to gain an advantage out of choosing this launch option.

Personally, I always choose to go with parking orbits because I can't bear to look up or calculate great launch windows. I work all day, I don't want to have to work to play Kerbal too, and the extra dV I can just engineer into my rockets and deal with the extra encumbrance of flying unwieldy rockets.

Edited April 11, 2014 by Mesons
Website problem caused double-post

[Eric S]

Options #1 and #2 aren't as far apart as people tend to think if you're doing them efficiently. Because of the way orbital mechanics work, you're going to get the most efficiency if you focus on horizontal velocity rather than vertical velocity once you're far enough out of the atmosphere. This is the point of the gravity turn. Now, if you're doing that, at some point you're going to be in LKO even if you don't aim for it because you'll be moving at the right speed to be in an orbit. The biggest difference is that #2 is going to have a lower periapsis since the periapsis doesn't have to be out of the atmosphere, but not hugely so because if you have too low of a periapsis, you're running into the gravity losses of #3. If you're minimizing your gravity losses, the periapsis will be in the mid to upper atmosphere at the lowest.

[Dunbaratu]

Options #1 and #2 aren't as far apart as people tend to think if you're doing them efficiently. Because of the way orbital mechanics work, you're going to get the most efficiency if you focus on horizontal velocity rather than vertical velocity once you're far enough out of the atmosphere. This is the point of the gravity turn. Now, if you're doing that, at some point you're going to be in LKO even if you don't aim for it because you'll be moving at the right speed to be in an orbit. The biggest difference is that #2 is going to have a lower periapsis since the periapsis doesn't have to be out of the atmosphere, but not hugely so because if you have too low of a periapsis, you're running into the gravity losses of #3. If you're minimizing your gravity losses, the periapsis will be in the mid to upper atmosphere at the lowest.

It's true that the optimal way puts your periapsis in the upper atmosphere not the lower atmosphere, but that's got nothing to do with gravity losses. It's because of drag losses. Go somewhere without air and the most efficient takeoff is one that is as low as possible, just barely enough to clear the terrain. The most effecient trip to orbit from the Mun, for example, is to make an orbit of only 8km, skimming along the ground just above the highest mountain peak. The primary reason not to do that in game is purely the boredom problem that the game won't let you time warp very fast when you're low to the ground.

[klappertjes]

Whoa. Never had this much response to a question before

You all had me on "Oberth effect is not about close to the planet, but at high speed" but the rest has definitely been an interesting and enlightening read. I'll dismiss method 3 (by which I meant what Vexx32 said) as the nonsense it is, and proceed to do 1, or 2 if I'm feeling good about my timing.

Thanks a lot everyone

Edited April 12, 2014 by klappertjes

[Eric S]

It's true that the optimal way puts your periapsis in the upper atmosphere not the lower atmosphere, but that's got nothing to do with gravity losses. It's because of drag losses. Go somewhere without air and the most efficient takeoff is one that is as low as possible, just barely enough to clear the terrain. The most effecient trip to orbit from the Mun, for example, is to make an orbit of only 8km, skimming along the ground just above the highest mountain peak. The primary reason not to do that in game is purely the boredom problem that the game won't let you time warp very fast when you're low to the ground.

Sorry, I didn't explain the point there very clearly.

The comment about gravity losses was dealing with the case of "don't raise your periapsis, just put your apoapsis at the height of the moon." Unless you've got an insanely high thrust, this usually involves focusing on vertical or prograde thrust by following a very weak gravity turn.

Let's see if I can explain it better.

For the "efficient direct ascent" Munar transfer, you'd start off with a regular gravity turn, and when you are building horizontal velocity, rather than stopping when you've got an apoapsis in the LKO orbital range, you continue until your apoapsis is at Munar height. My main point there is that if you're doing it efficiently, at some point before you achieve an apoapsis at Munar height, your apoapsis is at LKO height and your periapsis will probably be above (or at) the surface rather than below the surface of the planet.

The upper vs lower atmosphere comment was more about the fact that I generally find that if my periapsis is still below the upper atmosphere when I'm raising my apoapsis for a transfer then something is wrong, because I should be raising my periapsis until it becomes my apoapsis in that case. If I'm ahead of my periapsis approaching my apoapsis and burning for a transfer to the Mun, then the drag losses I'll encounter at my periapsis won't matter since my orbit will be changing before I get there.

I can actually explain the point using an "efficient return from the Mun" type mission, which doesn't involve drag losses, where the goal is to go directly from the Munar surface to a transfer orbit back to Kerbin without making a Munar orbit. Without atmospheric drag losses, the most efficient ascent is, as you said, just enough vertical thrust to keep you from colliding with the terrain. So at first, you're building up vertical velocity. This doesn't last long, I'm usually starting my gravity turn as soon as I have vertical motion. Then you're applying just enough vertical thrust to maintain your apoapsis while you raise your periapsis. At some point, you'll have raised your periapsis to the point it becomes the apoapsis and your current position becomes the periapsis. Even though you're not stopping there, you're now in orbit of the Mun.

If you were to stop there, orbit, then continue raising your apoapsis when you arrive back at your periapsis, you have not cost yourself any efficiency. You were already in an orbital trajectory, so it cost nothing to orbit.

I think the point I'm trying to focus on is that even with a gravity turn, a direct ascent via raising your apoapsis without raising your periapsis isn't going to be efficient due to gravity losses, whereas raising your periapsis until it becomes your apoapsis and then continuing to raise it until it reaches a Munar transfer orbit passes through the point of being a low kerbin orbit, so it isn't necessarily more efficient than a launch to orbit then do a transfer burn. All you're really doing is combining your circularization burn with your transfer burn. You can save a little delta-v by having this circularization point being in the upper atmosphere since you won't be there long enough to accumulate significant aerodynamic drag losses, but having the circularization point in the lower atmosphere means you'd have more aerodynamic drag losses which would nullify the benefit of not raising your periapsis higher (in this case, you're right, it would be because of the atmospheric drag).

[Dunbaratu]

Sorry, I didn't explain the point there very clearly.

You did. I just don't think you're correct. The atmospheric drag is small enough in the upper atmosphere that aiming to circularize at, say, 55000m instead of 70000m gives better results. Granted, you can't maintain that circular orbit at 55000m because even the thin atmosphere is enough to slowly degrade the orbit, but that's sort of my point. If you're doing a direct burn you don't have to maintain it circularized. You only have to if you're going to establish a stable parking orbit first. If you're doing a direct burn then you're only going to circularize temporarily. As you continue to push through circularization - periapsis swaps with apoapsis and you don't stop burning, you keep going to push the apopasis all the way up to the Mun. Thus you have a transfer ellipse drawn from 55000m to 12 million meters instead of one drawn from 70000m up to 12 million meters.

What the lack of parking orbit buys you is the ability to circularize while still at a lower altitude because you don't need that circularized state to be stable. You'll push right past circularization and into transferring, which will pull you up out of the atmosphere very soon anyway. A parking orbit demands you be higher than that at the point you begin your transfer burn, and therefore you're doing it from a slower starting velocity.

[Aedile]

The only thing number 3 is good for, is saving on mission time, really lousy way to do things.

Only time I do vertical launches, is return from my mun base, which is nicely positioned to point retrograde (tidally locked Mun, kethane rifinery), and is manly for the ease of use(no MJ).

[Eric S]

You did. I just don't think you're correct.

Well, I'm still failing to communicate in some manner, because every point you made there is not only something I agree with but something I said (or at least tried to say).

If you have a lower atmosphere periapse that you're not raising, then you're either burning at the horizon doing the #2 launch low enough that atmospheric drag will play a major role (which is what I think you're thinking I'm talking about), or you're burning prograde higher than the horizon, at which point you're doing a #3 launch, which will involve heavy gravity losses (which is what I was trying to say).

Or maybe I'm not making my original disagreement clear enough, because the point I was trying to make isn't contradicting anything brought up explicitly in this thread. Of all the people that I've seen talk about the efficiency of the #2 launch style, you're the first one to acknowledge that the savings come from the fact that you're possibly effectively circularizing at a lower altitude, one that wouldn't be maintainable (for long) with a #1 style launch. The people I've seen previously discuss the efficiency of this launch style seem to think the improved efficiency comes from them not circularizing, which they still are if they're doing this right, they're just not stopping the burn once they've circularized.

I'm not saying that it isn't more efficient, I'm disagreeing (with them) on why it's more efficient, and in some cases, the amount of delta-v that would be saved this way. I haven't done the math, but I think you'd be hard pressed to save more than 100 delta-v on this, and only about 30-40 m/s of that will be post-circularization transfer burn. I've seen people that think this method saves them however much delta-v it takes them to circularize and that just isn't the case for "typical" launch profiles.

[Dunbaratu]

Well, I'm still failing to communicate in some manner, because every point you made there is not only something I agree with but something I said (or at least tried to say).

I'm not saying that it isn't more efficient, I'm disagreeing (with them) on why it's more efficient

THAT is what you didn't communicate. Up until now you kept saying that the people saying #2 is most efficient were incorrect. At no point did you (A) mention that they were correct but for the wrong reason, or ( mention that you were arguing with other people's explanations rather than mine. As the original proponent of that position in this thread was me, that's kind of a necessary distinction to make.

And the reason I didn't mention that you push past circularization to get to a transfer orbit at the start is that I wasn't assuming people didn't already know this. It's obvious.

I think you were being unnecessarily condescending of people because the explanation you're claiming you were arguing against is one I haven't seen anyone arguing for.

Edited April 13, 2014 by Steven Mading

[FREEFALL1984]

In my opinion the most efficient method would be to launch, perform a perfect gravity turn and keep burning horizontally to kerbin until your perfectly timed intercept. That would be the most time efficient (time in space, not time in waiting on the pad for the correct launch time) but it wouldn't be any more fuel efficient than entering a parking orbit and waiting for the right window.

[Mesons]

The only thing number 3 is good for, is saving on mission time, really lousy way to do things.

Only time I do vertical launches, is return from my mun base, which is nicely positioned to point retrograde (tidally locked Mun, kethane rifinery), and is manly for the ease of use(no MJ).

Very clever! That's a great idea!

[Eric S]

And the reason I didn't mention that you push past circularization to get to a transfer orbit at the start is that I wasn't assuming people didn't already know this. It's obvious.

Obvious to you and to me, but like I said, I've seen more people try to justify the efficiency of this launch profile by stating that you're not wasting delta-v circularizing than that actually understood where the efficiency comes from. Then again, there are people that have argued that it's obvious that #3 should be the most efficient launch profile. Orbital mechanics aren't intuitive to most people, so the term obvious may not fit as well as it does in everyday life.

I never said you couldn't save delta-v doing this. I did say that it wasn't necessarily more efficient (if your point of circularization is above the atmosphere being the exception to improved efficiency) and that the difference between the two wasn't as strong as I've seen argued.

[Dunbaratu]

Obvious to you and to me, but like I said, I've seen more people try to justify the efficiency of this launch profile by stating that you're not wasting delta-v circularizing than that actually understood where the efficiency comes from. Then again, there are people that have argued that it's obvious that #3 should be the most efficient launch profile. Orbital mechanics aren't intuitive to most people, so the term obvious may not fit as well as it does in everyday life.

I said it was obvious because you don't need to know one bit of orbital mechanics. Just the basic geometric picture shows it. If you imagine a hoop of of weak bendy metal in the shape of an ellipse, and you grab it by the short axis and pull it apart to make it stretch the other way, clearly it's going to pass through circular on the way there.