Optimal Vacuum Ascent?

[Slam_Jones]

Hi,

I've been playing KSP for a while, and I feel like I should know this, but I don't.

My question is: what is the optimal ascent profile for non-atmospheric bodies? I usually go vertical to about 100m, then pitch off to somewhere between 15 - 30 degrees, but I assume there's a better way to do it. I assume a gravity-turn sorta thing would help, but I'm not quite smart enough to figure out how or why.

Any suggestions?

Edited September 9, 2015 by Slam_Jones

[Red Iron Crown]

Horizontal as quickly as possible without crashing into terrain. The optimal would be a single, instant horizontal burn to set Ap followed by another instant burn at Ap to circularize. The closer your ascent profile matches that the more efficient it will be, delta-V wise.

[Starhawk]

Optimal is to immediately pitch over and keep prograde just above the horizon. Of course, you need to pitch up quickly enough to clear all the terrain.

Happy landings!

edit: Ninja'd by RIC. And, of course, he said it much more clearly.

[Slam_Jones]

Awesome, simple enough. +rep and thanks!

[Edax]

Yes horizontal as soon as possible. Be aware though that you will not be able to clear terrain quickly though, so if a mountain is headed toward you on an assent, there's really nothing you can do unless you have powerful acceleration to dodge.

[legoclone09]

Burn with a grav turn until the apoapsis is 500m over the highest terrain and get to orbital velocities from there.

[ajburges]

OP did not specify what they were optimizing for.

Horizontal burn is only optimal from a dV standpoint. It also assumes a high TWR. For any given craft though, the quicker your gravity turn trajectory is tangent to the ground (without loosing altitude or colliding) the more efficient.

Mass optimal ascent is more complicated and involves lander design. You are optimizing for engine size, fuel, and ascent path.

Time optimal ascent is fun. Mostly 'cause you use Moar Boosters!

[FlyingPete]

Most fuel-efficient is to go horizontal immediately. On a perfectly flat, spherical planet, you could have landing gear and a side-firing engine and accelerate to orbital velocity that way. You could do this with a very low TWR if you had the patience. Of course, in the 'real world' fuel efficiency doesn't matter if you go into the side of a mountain. I tend to fire at 45 degrees or so to clear any nearby terrain, then go to almost-horizontal and complete the burn from there.

[Slam_Jones]

OP did not specify what they were optimizing for.

Horizontal burn is only optimal from a dV standpoint. It also assumes a high TWR. For any given craft though, the quicker your gravity turn trajectory is tangent to the ground (without loosing altitude or colliding) the more efficient.

Mass optimal ascent is more complicated and involves lander design. You are optimizing for engine size, fuel, and ascent path.

Time optimal ascent is fun. Mostly 'cause you use Moar Boosters!

It's partially just for general theory knowledge, but the practical application is that I have a Munar Mining/ISRU operation, and a single-stage lander that lands next to the mining base, connects via KAS to collect fuel, then heads up to the orbital station to deposit the payload. This ship would be performing multiple runs throughout it's lifetime.

Typically I try to keep my Munar TWR above 3.0, so that I have enough power for "emergency maneuvers."

[Aethon]

This will help.

Travel on airless worlds.

http://hopsblog-hop.blogspot.com/2014/06/travel-on-airless-worlds.html

[Hagen von Tronje]

I may be mistaken in this belief, but I think the "optimal" ascent profile on any nonatmospheric body is to level the navball with the horizon marker and keep it there as soon as you have enough vertical speed to clear the highest obstacle in your path.

Indeed I sometimes find myself wishing SAS had a "hold horizon" function, because I do the same thing when gravity turning off Kerbin if I've nearly put my Ap up to desired altitude but I still want to burn off the rest of this first stage before doing space maneuvers.

[Red Iron Crown]

I may be mistaken in this belief, but I think the "optimal" ascent profile on any nonatmospheric body is to level the navball with the horizon marker and keep it there as soon as you have enough vertical speed to clear the highest obstacle in your path.

You are not mistaken.

Indeed I sometimes find myself wishing SAS had a "hold horizon" function, because I do the same thing when gravity turning off Kerbin if I've nearly put my Ap up to desired altitude but I still want to burn off the rest of this first stage before doing space maneuvers.

I have long held that when the navball is in surface mode it should hold its direction relative to the horizon.

[Starhawk]

I have long held that when the navball is in surface mode it should hold its direction relative to the horizon.

This makes so much sense.

As soon as you think about it, it seems completely obvious that this should be the case.

Happy landings!

[ajburges]

I may be mistaken in this belief, but I think the "optimal" ascent profile on any nonatmospheric body is to level the navball with the horizon marker and keep it there as soon as you have enough vertical speed to clear the highest obstacle in your path.

Indeed I sometimes find myself wishing SAS had a "hold horizon" function, because I do the same thing when gravity turning off Kerbin if I've nearly put my Ap up to desired altitude but I still want to burn off the rest of this first stage before doing space maneuvers.

I disagree. For any given design, optimal ascent in absence of atmosphere is to burn prograde and circularize before you pass Ap. Up and to the side has more cosine losses than launch at angle theta hold prograde and control throttle to both avoid mountains and circularize at Ap.

You essentially want to turn an elliptic orbit with you at the Ap traveling a surface speed (with 1g constant radial thrust) to a higher circular orbit. We all know that the most efficient way to raise an orbit at Ap is to thrust prograde (horizontal). The radial (vertical) component is to push Ap ahead so we don't pass it and lose altitude (or hit a mountain). You want to minimize it as much as possible. Better to pay the cost as a trigometric fraction of that requisite prograde thrust.

SAS function to hold heading relative to horizon would be godsend though.

[Hagen von Tronje]

I disagree. For any given design, optimal ascent in absence of atmosphere is to burn prograde and circularize before you pass Ap. Up and to the side has more cosine losses than launch at angle theta hold prograde and control throttle to both avoid mountains and circularize at Ap.

You essentially want to turn an elliptic orbit with you at the Ap traveling a surface speed (with 1g constant radial thrust) to a higher circular orbit. We all know that the most efficient way to raise an orbit at Ap is to thrust prograde (horizontal). The radial (vertical) component is to push Ap ahead so we don't pass it and lose altitude (or hit a mountain). You want to minimize it as much as possible. Better to pay the cost as a trigometric fraction of that requisite prograde thrust.

SAS function to hold heading relative to horizon would be godsend though.

Not sure I follow. On initial ascent, prograde is wherever you point the rocket. Of course you will have to circularize with a prograde burn and you will want to do so at Ap. As I understand it, horizontal speed is retained, vertical speed is wasted after whatever it took to get you to safe altitude, hence why levelling with the horizon so your dV is going toward horizontal speed, no?

To be clear, I'm not talking about what you do to create the desired orbit out of a suborbital trajectory, I'm talking about the procedure from going from surface to suborbital so you can do that. Maybe I'm an idiot and missed something, but I think if you stick SAS on prograde starting at ascent then do nothing to adjust your AoA, you're going to go in pretty much the direction you were initially pointed and nowhere else, so unless you can somehow adjust your rocket to your desired AoA on the ground, you're still going to be flying the initial part by hand until you get enough speed to safely start maneuvering. To be fair, I have little experience with nonatmospheric HTO ascent vehicles, so maybe some of you genius designers have made easily adjustable AoA takeoff craft, in which case you rock!

But what do you mean about controlling throttle? Common wisdom seems to be that when fighting gravity, max thrust is king because less goes to waste. I can certainly see throttling up to avoid flying into a mountain face, but I wouldn't throttle down if I was pretty sure I was already clear of the mountain, I'd lower my AoA so my thrust is going in the direction I want, or I'd cut thrust entirely and just straight up prepare to circularize at low altitude when I get to Ap.

[NathanKell]

I, too, thought that the optimal in terms of delta V involved burning during ascent, rather than an impulsive burn to establish apoapsis. During the coast to apoapsis you're just taking in gravity losses (and losing Oberth effect advantage) without gaining anything in return.

However, it's plausible that a pure impulsive burn from launch and a pure impulsive burn at apoapsis (precisely half a world away) will amount to the same thing, it's just not the sort of math I want to take on. (Note: no cosine losses here, that's not the issue.)

[OhioBob]

My general rule is to accelerate vertically until my velocity is 1/8^th orbital velocity, pitch over to 45^o and continue accelerating until my velocity is 1/4^th orbital velocity, pitch over to horizontal and continue accelerating until the planned apoapsis is reached, cut the engine and coast to apoapsis, then perform a circularization burn at apoapsis.

Edited September 10, 2015 by OhioBob

[p1t1o]

So I was on the Mun the other day...

As I was planning m return to Kerbin, I realised that my landing site was almost directly at the spot where the Muns orbital line passes through on the retrograde side. So to return I just did a max-G ascent straight up, 90deg pitch until I had enough speed to clear the SOI, the resulting orbit had me on a nice fast return trajectory.

Would it have been more efficient, in terms of dV, to instead burn into a low orbit (of Mun), then execute another burn so that I swing past the front (prograde) face of the Mun, raise my Ap on Munar retrograde side and exit the SoI that way?

The vertical path saves on gravity losses, I think, but the orbital path gets a retrograde slingshot.

I feel like the orbital route might be more efficient, mainly a gut feeling I get because the vertical ascent just seems lazy!

But it was more fun

Thoughts?

[March Unto Torment]

My general understanding of things - and I could be wrong - is that a direct ascent profile is always the most gravitationally efficient, since you never need to worry about circularising or establishing horizontal velocity.

Certainly, the minimum dV I've ever used for transferring to and from the Mun was a direct ascent colliding with its surface, and then the exact same back. For reasons I don't fully understand, immediately after leaving the Mun's SOI, you're on a trajectory for Kerbin's atmosphere. The only problem is that you can hit the atmosphere going way too fast; multiple aerobraking passes may be necessary.

[Warzouz]

Certainly, the minimum dV I've ever used for transferring to and from the Mun was a direct ascent colliding with its surface, and then the exact same back. For reasons I don't fully understand, immediately after leaving the Mun's SOI, you're on a trajectory for Kerbin's atmosphere. The only problem is that you can hit the atmosphere going way too fast; multiple aerobraking passes may be necessary.

That's because the speed you get/lost during your escape Mun SOI burn is not lost when entering Kering SOI. If it had been, you would find yourself on a near Mun orbit. It's not.

And diving into Kerbin atmo is not that dangerous. It depends on your return vehicle. Mine (Mun/Minmus) usually have airbrakes and don't even have any heatshield. For Now, only Moho returns are specific : I usually do a slowdown to 4500m/s before entering atmo, or my airbrakes explodes. I usually dive between 10km and 35km depending on my speed.

[mhoram]

@p1t1o:

Lets calculate how much dV is needed to escape Muns SOI for the two scenarios under the assumption that your Low Mun Orbit is at an altitude of 0m and burns are instanteanous and you start at the equator:

1. Vertical Ascent from the Mun surface to the SOI

In order to leave Muns SOI, you will need to burn about 808 m/s vertically

2. LMO + Ejection Burn

To get into a LMO (at altitude 0) you need 570-9 m/s dV (9 m/s because that is the your initial speed induced by the rotation of the Mun)

The ejection burn costs 237 m/s

In sum 798 m/s.

So under the stated assumptions the difference is roughly the Mun-Rotation velocity and you are more efficient by using the circularize first method.

Here is the calculation:

https://www.dropbox.com/s/uujozwc7ipxfgnr/MunExiSoiCalc.xlsx?dl=0

Also getting into orbit first makes planning the time of the ejection burn much easier.

Edit: Another explanation:

The escape velocity of the Mun is 807 m/s (every object that travels at this velocity above the surface will eventually leave Mun's SOI)

Accelerating in the same direction as your initial 9m/s Mun-Rotation speed hepls to reach this velocity.

Edited September 10, 2015 by mhoram

[p1t1o]

@p1t1o:

Also getting into orbit first makes planning the time of the ejection burn much easier.

Yah! Thats definitely true! It was really a 1-off situation.

Thanks for taking the time to get the figures

I'm surprised the dV values were so close together - just seperated by the Mun's rotational 9m/s, I was expecting to see a component from the slingshot effect, or perhaps thats the other 1m/s between 798 and (808-9)?

[mhoram]

@p1t1o: in both scenarios there is no gravitational slingshot involved.

A gravitational slingshot is the effect, that by entering and leaving the SOI of a moon (brown orbit) you change your current orbit (blue) into an other (purple) without needing to spend any dV.

[cantab]

My understanding is the most efficient ascent or landing is a "constant altitude" one. Take off and immediately pitchover so that you maintain zero vertical speed as you build horizontal speed. As you gain speed you will be able to pitch closer and closer to the horizons. Once you attain a ##x0 km orbit cut engines, coast to apoapsis, and circularise there. With a high TWR lander you'll be aiming quite close to the horizon from the start, but with a low TWR one you'll be aimed much more vertically.

In practice you'll need to adapt this to ensure you clear any mountains. I think it would be most efficient to maintain a constant non-zero vertical speed that's just enough to clear any terrain ahead of you, then let it drop to zero when you're high enough.

[p1t1o]

@p1t1o: in both scenarios there is no gravitational slingshot involved.

A gravitational slingshot is the effect, that by entering and leaving the SOI of a moon (brown orbit) you change your current orbit (blue) into an other (purple) without needing to spend any dV.

There should be a slight slingshot effect on the latter case, circularise-before-exit, any open trajectory around a body should benefit from the effect, once you've made your exit burn you are now in an open trajectory that fulfils the criteria you laid out - you are on the same trajectory as if you had entered the SOI at a velocity which would put you at your [just over escape]velocity where you are at when you made the burn, if you follow my meaning. Since relative velocities we are talking about are quite low, I'd expect that the effect would be minimal.