Minimizing Delta-V in landing burn

[rkman]

And what counts is the amount of delta-v spent on landing from low orbit?

[rkman]

I loaded the persistence file, hyperedited the craft in 15km Mun orbit.

Spent ~45m/s to get intersect with the surface, 690m/s total.

Used mechjeb killrot most of the descent, using pitch to affect vertical speed while keeping an eye on suicide burn countdown, used Translatron for the last 100m or so.

Edited October 7, 2013 by rkman

[Kasuha]

And what counts is the amount of delta-v spent on landing from low orbit?

What counted was the amount of delta-v (or fuel) to land from that point and speed. It's irrelevant that the orbit is "unnatural" as you need to spend just a little fuel to make it "natural". Or you may just try to brake the freefall from that point - there's enough fuel in the tank for it (I landed with 26 fuel left that way) but it's even less efficient than making the orbit elliptic first.

[rkman]

What counted was the amount of delta-v (or fuel) to land from that point and speed. It's irrelevant that the orbit is "unnatural" as you need to spend just a little fuel to make it "natural". Or you may just try to brake the freefall from that point - there's enough fuel in the tank for it (I landed with 26 fuel left that way) but it's even less efficient than making the orbit elliptic first.

Somehow i got the impression that the number to beat is the amount of delta-v spend on landing from 15k orbit, as per the video.

[Kasuha]

Somehow i got the impression that the number to beat is the amount of delta-v spend on landing from 15k orbit, as per the video.

It started with this:

Make a quicksave with the ship at the apoapsis of a sub-orbital trajectory of your choice and let's check who can land it with more fuel left.

And the answer was:

Persistence file can be downloaded here if you would like to have a go at it.

All three of us started at that point so if you want to compare your result with us, you should use that as is.

Results so far are:

Kosmo-not: 926 m/s dv, 35.24 fuel left

Kasuha: ??? dv, 33.58 fuel left

Vector: 989 m/s dv, 32.13 fuel left

Edited October 8, 2013 by Kasuha

[rkman]

I got the impression it started with the video that Kosmo-not posted where he landed from 15km orbit, and subsequent claims of others being able to do it more efficient. Then Kosmo-not issued a challenge:

"Can you propose a different landing method for that craft that uses less delta-v?"

http://forum.kerbalspaceprogram.com/threads/51577-Minimizing-Delta-V-in-landing-burn?p=677929&viewfull=1#post677929

(that is the challenge that i took up, trying to beat 840m/s delta-v landing a low TWR craft from 15km orbit - though not as low twr as the craft in the video, which explains why i did beat it).

After that you proposed what i think is a different challenge (starting suborbital instead of in orbit) which resulted in Kosmo posting the Persistence file (with apoapsis far above 15km).

If it really is irrelevant that the orbit in Persistence file is not an orbit because only a little fuel is needed to make it an orbit, then would hardly matter that i used no delta-v to make the orbit "natural".

But actually the scenario is so different from what is shown in the video that all in all it does make a big difference wrt delta-v.

Having cleared this up, what it comes down to is that i am not taking part in that challenge.

I tried the scenario in the video again with a modified version of the craft in the persistence file, using 3 LV-1R's instead of 4, resulting in an initial TWR of 1.05. This time it took me a lot more delta-v: 935m/s.

The problem is that due to the low TWR a normal descent trajectory gives a time to impact shorter than the time needed to do a suicide burn, requiring one to extend the descent trajectory, which takes a significant amount of delta-v.

For such a scenario it apparently is better to do as Kosmo did: put periapsis above ground, giving more time and thus requiring less pitching up to reduce vertical speed, which saves delta-v.

The whole "use pitch to control vertical speed" thing is only really important to very low TWR craft. It is inefficient and imo does not do much to address the OP's question.

Generally, suicide burn is the way to minimize landing delta-v, you need enough time to be able to do a suicide burn or else delta-v will not be minimized, calculating when to start suicide burn is exceedingly complicated. Either have a large safety margin in the descent, or use Mechjeb's calculated suicide burn countdown as a guide.

My take away from it is that it is best to avoid low TWR landers and instead go for a TWR of at least about 2. Afaik the LEM TWR was a little over 2.

[tavert]

For such a scenario it apparently is better to do as Kosmo did: put periapsis above ground, giving more time and thus requiring less pitching up to reduce vertical speed, which saves delta-v.

The whole "use pitch to control vertical speed" thing is only really important to very low TWR craft. It is inefficient and imo does not do much to address the OP's question.

Generally, suicide burn is the way to minimize landing delta-v, you need enough time to be able to do a suicide burn or else delta-v will not be minimized, calculating when to start suicide burn is exceedingly complicated. Either have a large safety margin in the descent, or use Mechjeb's calculated suicide burn countdown as a guide.

My take away from it is that it is best to avoid low TWR landers and instead go for a TWR of at least about 2. Afaik the LEM TWR was a little over 2.

Low TWR results in a larger more noticeable difference between constant-altitude and retrograde landing methods. But constant altitude is always more efficient than retrograde, for any finite TWR - the difference goes to zero asymptotically in the limit of infinite thrust.

You will use more delta-V with a lower TWR. Exactly how much more? See http://forum.kerbalspaceprogram.com/threads/39812-Landing-and-Takeoff-Delta-V-vs-TWR-and-specific-impulse for precise numbers. But you can achieve more delta-V capacity with the same total craft mass at a lower TWR. You really want to look at overall payload fraction, for which there's an optimum TWR, and it's lower than you think.

[Runboy]

The best way (To my knowledge)to bleed off as much Delta-V during Landing is aim DIRECTLY at the retrograde vector.By experience,this is the best method for me.

Leave a reputation if this was helpful (The sheriff star symbol on the bottom left)

[numerobis]

Sorry Runboy, no gold star -- read the thread first.

[iStickyDuck]

To minimize Delta-V used in landing, lithobrake... You can safely protect your entire ship if you use those panels correctly...

[rkman]

Low TWR results in a larger more noticeable difference between constant-altitude and retrograde landing methods. But constant altitude is always more efficient than retrograde...

But you can achieve more delta-V capacity with the same total craft mass at a lower TWR. You really want to look at overall payload fraction, for which there's an optimum TWR, and it's lower than you think.

I have come to the conclusion that the video shows an extreme case of what is generally a more efficient landing technique than what most people seem to do (first kill a lot of orbital velocity, then drop down and burn some more retro), but the craft in the video can not make the most of it because of its low TWR (1.06 initial) - to low to be efficient.

Although it is the most efficient for that particular scenario, it is also pretty much the only possible way to actually land (not crash) that craft from that orbit.

Just to avoid crashing it needs to spend delta-v tangent to the direction of travel to keep the craft off the ground long enough to be able to lose enough horizontal speed for safe landing. That is a lot of delta-v that it would not need to spend if it had higher TWR.

As such it is imo not a very good demonstration of what is in principal an efficient landing technique.

It is trivial to demonstrate that a landing very similar to the one in the video (near horizontal trajectory for much of the descent) with a craft that has higher twr, requires significantly less delta-v than 840m/s used in the video. (690m/s I reported earlier was with twr 1.33 initial; the ship from the persistence file - edit: i can't reproduce that 690m/s, it's more like ~740m/s)

Edited October 9, 2013 by rkman

[tavert]

1.06 is a bit on the low side, at least for the Mun and a non-nuclear lander. How much higher you should go for better payload fraction depends on which engine you're using. If you assume infinitely divisible engines and fuel tanks (unrealistic assumption, but it allows us to make general conclusions without getting specific about size of lander), here's the payload fraction for a single-stage constant-altitude landing on the Mun:

And if you also want to take back off with the same stage:

For LV-N or LV-1, the optimal TWR for the Mun is somewhere between roughly 1.1 and 1.6. For other engines, it's around 2. Little lower if you just care about landing, little higher if you count takeoff too.

Edited October 9, 2013 by tavert

[Wallace]

another way to save a little bit of delta V is to have robust landing gear for lithobraking. If you are not adverse to mods, then try out these bouncyboo airbags. http://kerbalspaceprogram.com/cp-airbag-system/ You'll have to experiment to find the splattage speed for your lander though.