[0.24] If a Booster Has Enough Parachutes On It, Count it as "Recovered"

[Jouni]

Some concrete examples:

In this case, desert wasn't flat enough (for this).

Here I was able to find a patch of grasslands near KSC that was barely level enough to land, but only after several attempts.

[Killerdude8]

Fairly simple idea(probably not too simple to implement), To get around this, Any part with an active and armed Chute, Should be recovered and compensated when taken beyond the 2.5km limit.

[vexx32]

Really? A massive booster stack with a Mk16 parachute should be classed as "recovered"?

I mean, that's an extreme example, but the simple fact of it having a chute is by no means a reliable metric to govern whether or not the part is likely to arrive on the ground intact. It's a bit more complicated than that.

EDIT: A method that might work somewhat better would go a bit like this, I think:

1. Calculate the total drag of all currently deployed chutes, as they would be when fully deployed.
   
2. Calculate an estimate of the landing velocity based on the detached section's mass and the total drag of the parts and chutes combined, as well as gravity and air resistance.
   
3. Compare this estimate with the maximum impact tolerance of each part, one at a time. For any parts where their impact tolerance is lower than the estimated landing velocity, count them as "destroyed".
   
4. Recover all remaining parts that are not marked as destroyed, then delete the detached stage.
   

Not even that is perfect -- it doesn't account for unusual vessel shapes, and assumes that every part on the ship hits the ground in exactly the same way with the same force, but it's a lot more thorough than "if it has a chute on it, it must be fine", I think.

Edited July 18, 2014 by vexx32

[r4pt0r]

Really? A massive booster stack with a Mk16 parachute should be classed as "recovered"?

I mean, that's an extreme example, but the simple fact of it having a chute is by no means a reliable metric to govern whether or not the part is likely to arrive on the ground intact. It's a bit more complicated than that.

EDIT: A method that might work somewhat better would go a bit like this, I think:

1. Calculate the total drag of all currently deployed chutes, as they would be when fully deployed.
   
2. Calculate an estimate of the landing velocity based on the detached section's mass and the total drag of the parts and chutes combined, as well as gravity and air resistance.
   
3. Compare this estimate with the maximum impact tolerance of each part, one at a time. For any parts where their impact tolerance is lower than the estimated landing velocity, count them as "destroyed".
   
4. Recover all remaining parts that are not marked as destroyed, then delete the detached stage.
   

Not even that is perfect -- it doesn't account for unusual vessel shapes, and assumes that every part on the ship hits the ground in exactly the same way with the same force, but it's a lot more thorough than "if it has a chute on it, it must be fine", I think.

id like to make a suggestion for that idea.

[Kerbart]

From a gameplay perspective, contracting is introduced to make the game more challenging. Part of that challenge will be to come up with cost efficient designs. MOAR BOOSTERS! is obviously the opposite of that. Automatic recovery of anything with armed parachutes would largely bypass the incentive. “Ok I can continue to use my 30-jumbo tank/15 Mainsail lifter stage, I just attach one parachute to it that gets armed when I jettison the first stage at 4000m. I know that unlikely to be how you intend the parachute recovery to work but it seems to me that, given the way everything works right now, it would be.

That means that there's all kind of complexity to be added, people looking for loopholes, extra code that can contain bugs, etc.

Or it's done the way it's done now; if you want to recover it, it has to be part of a controlled vessel. A much simpler solution. What might happen is that in the future certain parts (the boosters like the BACC and the S1 SRB come to mind) have built-in chutes and will be recovered automatically under certain circumstances.

[kujuman]

Really? A massive booster stack with a Mk16 parachute should be classed as "recovered"?

I mean, that's an extreme example, but the simple fact of it having a chute is by no means a reliable metric to govern whether or not the part is likely to arrive on the ground intact. It's a bit more complicated than that.

EDIT: A method that might work somewhat better would go a bit like this, I think:

1. Calculate the total drag of all currently deployed chutes, as they would be when fully deployed.
   
2. Calculate an estimate of the landing velocity based on the detached section's mass and the total drag of the parts and chutes combined, as well as gravity and air resistance.
   
3. Compare this estimate with the maximum impact tolerance of each part, one at a time. For any parts where their impact tolerance is lower than the estimated landing velocity, count them as "destroyed".
   
4. Recover all remaining parts that are not marked as destroyed, then delete the detached stage.
   

Not even that is perfect -- it doesn't account for unusual vessel shapes, and assumes that every part on the ship hits the ground in exactly the same way with the same force, but it's a lot more thorough than "if it has a chute on it, it must be fine", I think.

Just for any potential modders seeing this topic, steps 1&2 are easy enough to do http://forum.kerbalspaceprogram.com/threads/50384-WEB-Parachute-Calculator-Updated%21?highlight=parachute

[LethalDose]

Holy kerballs, how many times is this going to get suggested?

And aren't the SRBs dirt cheap? Manley pointed out in one of his videos that it can cost MORE money to put the chutes on boosters than you get by recovering them.

[Sky_walker]

1. Calculate the total drag of all currently deployed chutes, as they would be when fully deployed.
   
2. Calculate an estimate of the landing velocity based on the detached section's mass and the total drag of the parts and chutes combined, as well as gravity and air resistance.
   
3. Compare this estimate with the maximum impact tolerance of each part, one at a time. For any parts where their impact tolerance is lower than the estimated landing velocity, count them as "destroyed".
   
4. Recover all remaining parts that are not marked as destroyed, then delete the detached stage.
   

Flawless landing in a mountains. 100% of time.

Awesome. I want that for every ship I fly.

And aren't the SRBs dirt cheap? Manley pointed out in one of his videos that it can cost MORE money to put the chutes on boosters than you get by recovering them.

Oh come on! You expect people to do maths? Seriously?

They'll always insist it's cheaper to land the thing and recover than simply get rid of it. After all chutes are small and fuel tanks are big, right?

Edited July 18, 2014 by Sky_walker

[Red Iron Crown]

For SRBs it wouldn't make sense, but I can imagine some LFO boosters might be worth recovering.

[LethalDose]

For SRBs it wouldn't make sense, but I can imagine some LFO boosters might be worth recovering.

Yeah, some might, and most probably won't, especially given that recovery can be for as little as 10% of what survives.

Considering how aggressive the moderators are about other topics (THAT ARE NOT SPECIFICALLY MENTIONED TO AVOID VIOLATING RULE 2.2 & 3.4!!!) that keep coming up over and over and OVER, I don't understand why they aren't locking/merging/dealing with these, or at least adding to the WNTS list. There's been at least 10 in the last week.

[Killerdude8]

and why should that prevent us from allowing this to be implemented?

imo, It makes perfect sense, some, Like myself will want almost everything dropped within atmosphere to be recovered, Efficient or not, it's a simple and neat feature.

[Jouni]

and why should that prevent us from allowing this to be implemented?

imo, It makes perfect sense, some, Like myself will want almost everything dropped within atmosphere to be recovered, Efficient or not, it's a simple and neat feature.

Automatic recovery should only be possible in those cases, where the average player would be able to land the booster safely. The bigger the booster, the harder landing with it becomes, as the are essentially tall, top-heavy landers.

[Claw]

Merging these similar topics.

Carry on,

~Claw

[Red Iron Crown]

Considering how aggressive the moderators are about other topics (THAT ARE NOT SPECIFICALLY MENTIONED TO AVOID VIOLATING RULE 2.2 & 3.4!!!) that keep coming up over and over and OVER, I don't understand why they aren't locking/merging/dealing with these, or at least adding to the WNTS list. There's been at least 10 in the last week.

[Technical Ben]

From a gameplay perspective, contracting is introduced to make the game more challenging. Part of that challenge will be to come up with cost efficient designs. MOAR BOOSTERS! is obviously the opposite of that. Automatic recovery of anything with armed parachutes would largely bypass the incentive. “Ok I can continue to use my 30-jumbo tank/15 Mainsail lifter stage, I just attach one parachute to it that gets armed when I jettison the first stage at 4000m. I know that unlikely to be how you intend the parachute recovery to work but it seems to me that, given the way everything works right now, it would be.

That means that there's all kind of complexity to be added, people looking for loopholes, extra code that can contain bugs, etc.

Or it's done the way it's done now; if you want to recover it, it has to be part of a controlled vessel. A much simpler solution. What might happen is that in the future certain parts (the boosters like the BACC and the S1 SRB come to mind) have built-in chutes and will be recovered automatically under certain circumstances.

It bypasses the reasons to even have any stages. Every ship will become a SSTO? Really, automatic recovery (of some sort) is required now. It's just how to do it that is up for discussion. It could be entirely automatic, it could be entirely simulated. And it could be anywhere in between.

[Red Iron Crown]

Obligatory "there's a mod for that" response: http://forum.kerbalspaceprogram.com/threads/86677

That was quick.

[Killerdude8]

It bypasses the reasons to even have any stages. Every ship will become a SSTO? Really, automatic recovery (of some sort) is required now. It's just how to do it that is up for discussion. It could be entirely automatic, it could be entirely simulated. And it could be anywhere in between.

So much win in this post.

This is why we need this, Otherwise if it isn't an SSTO, You are being horribly punished for launching and attempting to return and recover objects that can VERY clearly survive the descent.

So you go either SSTO, or you lose money.

[RuBisCO]

Well I tested it and 2 stage recovery works in 0.24 (I don't think anyone was doubting it would not but just in case) as long as the first stage remains above ~69 km while the second stage makes it to orbit, you can go back to the first stage and land it in the ocean or even fly it back to the launch pad ala SpaceX.

I disagree with making every part recoverable by just having parachutes on them: it takes some of the engineering joy, eer I means "challange", out of making parts that you can see actually survive landing, but if others want that I'm not stoping you: someone make a mod.

[Technical Ben]

I accept there are differences of opinion. So to help understand what people consider helpful and what is not, can I ask?

Would a "tested lander" or "tested recovery" option be helpful? So only if you have a 100% success with a recovery/lander design could it become automatically recovered (and only if landed near KSC)?

This would require the player to generate a "mission/objective" and test their recovery/landers before deploying. Then it would be marked as "recoverable" and only if it separates/activates in the same mode as the tested version?

[cpast]

So much win in this post.

This is why we need this, Otherwise if it isn't an SSTO, You are being horribly punished for launching and attempting to return and recover objects that can VERY clearly survive the descent.

So you go either SSTO, or you lose money.

Yup. Staging costs money. It's cheaper to minimize staging. What's the issue with that?

[regex]

Yup. Staging costs money. It's cheaper to minimize staging. What's the issue with that?

It's also more challenging to minimize staging. Players who can do that should be rewarded appropriately. This suggestion pretty much amounts to no costs for anyone who simply slaps some parachutes on a stack which means the game will have to be rebalanced around the fact that everything will now have parachutes. It makes money useless and devalues the skill of those who can create effective SSTOs.

[Godot]

It's also more challenging to minimize staging. Players who can do that should be rewarded appropriately. This suggestion pretty much amounts to no costs for anyone who simply slaps some parachutes on a stack which means the game will have to be rebalanced around the fact that everything will now have parachutes. It makes money useless and devalues the skill of those who can create effective SSTOs.

I don´t see a problem with this:

1. As soon as you leave Kerbin (and aim for Mun, Minmus or other planets) it becomes harder and harder to recover parts via staging ... 100% recovery is usually only possible for things on Kerbin / Kerbins orbit.

2. Recovery at the moment is buggy ... the recovery gives too much money and also doesn´t take into account that the recovered parts are evoid of all fuel ... as soon as squad fixes this, recovery should be come much less lucrative than it is now

3. Squad could introduce 2 classes of parts ... those that regularly landed, either on a runway (on wheels) or with landing legs ... and those that had a rough landing (i.e. recovered debris and/or landed command modules) ... with the first category giving more money upon recovery than the second category

4. The proposed mechanism should only work up to a certain altitude (for example 50 km ... in order to account or detsuction of parts via reentry heat (users of deadly reentry know what I mean ) ... everything that geta separated above this altitude automatically gets destroyed when it leaves the 2.5 km radius

This way (especially with 3. SSTOs are still a good option)

Edited July 18, 2014 by Godot

[Technical Ben]

Currently the game punishes you HARD for a mid air launch (airplane assisted etc) for anything not creative mode.

See the recent rocket launch designs too. They are looking to add re-usability in just this way, except it has no parachutes (it's a powered landing). As I don't know how KSP could sim both (other than timewarp backwards for multiple instances, and yes, that can be done. ), parachutes seem at least a partial solution and a partial goal for those wanting to save the first stage.

Of cause secondary or additional stages would burn up on re-entry or need full flight control for return.

[dronkit]

How about simple aritmethics recovery (as described in this thread) for inactive debris and physics bubble for "active debris" (ie. debris with a command part and electricity) so to be able to land assisting planes, maybe with mechjeb, KOS or back warp or something?

[Exxmorphing]

An idea to address those who state the issue of how parts fall over when they land.

I'm not sure if this idea is complicated enough for its own thread, and it would be great if someone would tell me if it is so. However, we're already discussing mainly on this thread, so I'll post it here for now.

And yes, I know that it's extremely ambitious and highly theoretical. Flame Shields up.

It may be possible for an algorithm to be implemented that uses the dimensions of section off-ed portions of debris for a more realistic recovery system. While I believe that this is quite feasible, it is just a theory.

Let's say, just to give a tangible example, that every 2 meters or so of an object's determined height, the game will check the general spatial dimensions of the object and possibly mass, thus statistically "separating" it into "portions". This could take place when it's no longer part of the main vessel, or when it goes out of the 2.5km limit. Those values could then be inserted into an algorithm, to which the objects within each section, whose boundaries are defined by the the locations of the measurements, are given certain chances of survival.

First, objects in higher up sections have a lower initial chance of getting destroyed, and the lower objects have a higher chance. Naturally, these chances are also affected by the calculated speed of the object (the speeds of it with the parachute deployed at the highest possible point) when it hits the ground and the impact resistance of each object.

As an example, a part with a resistance of 4 m/s will have a 90% chance (Figures are more or less random) of getting destroyed when it is in the bottom most portion of the overall object when falling at 6 m/s, but only a 20% chance of getting destroyed when in the top section.

But with the point of toppling over, any calculations made above will not be the final chances of destruction. This theoretical algorithm will then compare the general dimensions, and if possible, mass, of the selected points for a chance that the object will "fall". Certain areas of Kerbin, probably biomes, are then to affect that chance.

For example, if the measurements of the bottom portions of an object is quite skinny, say 5 meters in diameter, while the overall height of the object is high, say 40 meters, the chance of "falling over" is to be quite high; let's say 60%. If the object happens to land in a mountainous area, those chances are bumped up to 95%; if in a flat area, it may simply stay at 60%; if hilly, it may go up to 75%.

Once that's calculated, parts of the object are assigned different chances of getting destroyed, depending on which portion they are in. Generally, objects in the lower portions have a lower chance of getting destroyed, while objects in the higher portions have a higher chance. If a certain portion has a large area/width compared to the surrounding portions of the object, the objects in that measured area ("portion", as I've been referring to them as) may get a higher chance of getting destroyed, but a lower chance for the objects in the surrounding areas/portions. These calculated chances are then applied to each object, with their own individual impact resistance

If possible, it may also be wise for the mass of each overall area to play a roll in the chance of destruction.

Here's another example: An object with, say, 5 portions, is to be calculated for damage. An object with a resistance of 10 m/s in the bottom-most portion has a 12% chance of destruction if the object is to fall over or topple. That same object in the top most portion, portion 5, is to initially be given a 60% chance of destruction. However, portion 4 has a large area or width when compared to the other portions due to some outcropping objects. The same object in portion 4 has a 75% chance of destruction, but the same object in portion 5 is to have its chance of destruction changed from 60% to 40%.

Once this is calculated, chance takes its roll on the object that is to fall to earth. First, the initial fall destruction is to be resolved, and some objects are to be tagged for destruction, and others not. Afterwards, whether or not the object will have toppling damage applied will then be figured out. If luck has it that it does fall, then objects will have another chance of destruction, but with the new calculations meant just for toppling over.

I find it necessary to not combine the chances (falling and toppling) and make it all one big chance of destruction.

Any object that survives the roll o' death is then recovered.

As an option, if the object fell on mountainous terrain, all of the objects will have another chance for damage, but this time not in the measured portions, but rather proportionately equal.

Also, if an object is to exceed a single portion, should only the highest chance count, or should it have 3 chances for destruction?

While I'm sure that my idea is possible, its only a theory, and hella ambitious for one, and thus may end up being not very feasible at all.

It would be great if someone more learned on this could provide some insight on this and some actual values that could be plugged in.

Edit: There would also need to be a way for the game to determine the bottom and top of the object.

Edited July 20, 2014 by Exxmorphing
Made some of my idea clearer