New-fangled buoyancy model -anyone know how it works?

[little square dot]

I'm wondering how the various buoyancy-related parameters found in the .cfg files effect a part's floatiness, as it's not readily apparent to me. For example, I would just assume the centre of displacement to be the same as the centre of buoyancy, which it clearly isn't...or maybe it is by default, but either way I'm curious as to how offsetting them will impact a part's behaviour on the water.

Also, does a part's centre of mass have any bearing on it's buoyancy under the new model, and what the heck does [B]buoyancyUseSine[/B] do?


Cheers. =)

[NathanKell]

CoM = CoL = CoP = Center of displacement (i.e. how we know how much is submerged) = center of buoyancy (where the buoyancy force is applied) = part origin, [B]unless[/B] overridden (only a few parts override one or more of those).
There's also a buoyancy scalar--again, only a few parts set it to something other than 1.0.

To better simulate the usual condition where a fuel tank or other cylindrical part is in the water, buoyancy defaults to using a sine curve when calculating how much of a part is submerged. The first step is always to calculate maximum depth, calculate minimum depth (negative since above water), calculate depth at center of displacement, then calculate submerged portion lerping between max depth to CoD to min depth. Then, if the sine curve is used (default: true), feed that number into a sine function to get a nice curved, rather than linear, result (think of a cylinder hitting the water side-on, at first submerged volume increases slowly, then fast, then slowly again).

[Azimech]

This is very useful information for my mod, thanks Nathan! It will help prevent unwanted list.

As a small treat, here's a picture of my battleship using my mod. It's 260m. long and 28m. wide. Current mass: 8000 tons.

[img]http://i.imgur.com/cZCfxFi.png[/img]

[url]http://forum.kerbalspaceprogram.com/threads/139625-WW2-Warships-0-1-They-really-sink[/url]

[little square dot]

Makes sense. Thanks.

So I suppose the offset in the Mk1-2 command is just to ensure that the capsule remains upright? Why disable the sine curve? (sorry if I'm missing something obvious...it happens from time to time.)

[NecroBones]

[quote name='NathanKell']CoM = CoL = [B]CoP[/B] = Center of displacement[/QUOTE]

Having a brain fart... remind me what CoP stands for?

[little square dot]

Centre of pressure. I believe the parameter is used to offset the the application of drag from a part's CoM, but I could be wrong about that.

If memory serves (which it usually doesn't) I think drag is applied to a part's CoM by default, which caused some difficulties with the orientation of command pods on re-entry, which the CoP offset rectified.

[NathanKell]

You recall correctly. :)

And yep, that's why the center of buoyancy (force) is placed high for the capsule. The CoD is placed low because most of the displacing volume of a cone is near the base. The sine curve is turned off because of the conic shape, too (though the sine curve would be correct if the capsule were submerged with the conic axis parallel to the ocean surface, that's not usually how it will be submerged).

[NecroBones]

Awesome, that makes perfect sense. Thanks.,

[little square dot]

Is it just me or did the water become really sticky?

I just made my first 1.0.5 water landing using welded pontoons that used to glide nicely across the surface , but no matter how shallow my approach angle or slow my airspeed, as soon as they make contact with the water my pontoons grab on for dear life, taking me from 45 m/s to 0 almost instantly, within about 10m of my touchdown point. My plane somehow survives these violent "landings", but it's not so much water landing as water-crashing. Is there something I can add to/change in the .cfgs to make my pontoons less grabby and more glidey?

Call me old-fashioned, but I tend to prefer non-violent deceleration. :huh:

[Pds314]

It's definitely possible to land in the water with many planes. Most of my planes ditch fine and I've unwittingly created some "accidental seaplanes."

[MaxwellsDemon]

Well, my first attempt at a boat ended rather humorously... it drove very gracefully into the water and that was all she wrote. :D Back to the drawing board... I think I'm going to adopt a "pontoon boat" model instead of a hull-in-water design.

[little square dot]

[quote name='Pds314']It's definitely possible to land in the water with many planes. Most of my planes ditch fine and I've unwittingly created some "accidental seaplanes."[/QUOTE]

Landing/ditching intact isn't the problem. Ridiculously splatty landings are the problem.

The issue is that I've yet to perform a "landing" in 1.0.5 that hasn't resulted in violent deceleration from 50 m/s to 0 within a second of hitting the water. Under the old model I could make smooth, graceful landings and decelerate gradually, gliding across the surface for a couple hundred meters or so before coming to a stop

Now I can't get pontoons to glide across the water at all while landing. My aircraft survive, however it's not a particularly gratifying experience.

I wonder if it would help to assign itty bitty little drag cubes to the pontoons

[swjr-swis]

Not sure if it has anything to do with buoyancy (or the lack of application of it), but I have noticed that a splashdown with a still-closed fairing will send anything inside the fairing in an accelerating fall towards a fiery crash at the bottom of the ocean. It seems like the fairing being closed prevents buoyancy from being applied and the parts 'think' they are still in free-fall, until hitting the bottom.

So no more using undeployed fairings as hulls if anything inside them is meant to survive.

[Chaos_Klaus]

You should add that to the bug tracker. The volume of the parts is estimated via the drag cubes. So if there is no drag applied (like in a fairing), there might be no buoyancy aswell (by accident).

For landing on the water:

Make sure your angle is reeeeally shallow. If you drop below stall speed and lose altitude too fast, you will dip into the water too hard and deccelerate quickly. But I managed a few landings that were very smooth indeed. Even at speeds above 50m/s. The first time, I actually was sure I would crash. ;) Edited November 25, 2015 by Chaos_Klaus

[little square dot]

[quote name='Chaos_Klaus']
For landing on the water:

Make sure your angle is reeeeally shallow. If you drop below stall speed and lose altitude too fast, you will dip into the water too hard and decelerate quickly. But I managed a few landings that were very smooth indeed. Even at speeds above 50m/s. The first time, I actually was sure I would crash. ;)[/QUOTE]

Thanks, but still no dice. I've tried every angle and speed on several different types of floats, but the water firmly grabbed all of them, albeit to slightly varying degrees. I came closest to a normal-ish water landing using wings as pontoons, oriented vertically, tips-down. The knife-edge approach worked ok, but I still came to a full stop within 30 metres or so of landing (at 75m/s). It also looked a bit ridiculous. Unfortunately it seems they forgot about surface tension when constructing the new model.

The strange thing is that there are stock seaplanes now, but even they don't handle particularly well on the water. The "Gull" or Seagull or whatever can't seem to manage taking off again once landed, and the massive "Mallard" flying boat must be have some crazy magic skills or something, as it doesn't even rest in the water...it perches on top, suspiciously buoyant given the lack of displacement.


Meh, like a lot of Squad's new bits and pieces on initial roll-out, the new buoyancy model feels awkward and untested in my opinion, more like a stop-gap than a completed feature. I'm confident it'll be fine-tuned eventually, but in the meantime I'm not really diggin' it. I had no problems ditching in the water with the old model, and far more success with floatplanes...not that I'm suggesting we should go back, but it just grinds my gear a little bit when I find myself having to work around new systems. Given that KSP is predominantly a game about flying things, one would think that a new buoyancy model would be a bit more geared toward aircraft that play on the water.

[little square dot]

Good gravy... I just discovered how to make perfect water landings, and it makes no sense whatsoever.

Lower your landing gear.

haha... what!? That's not quite right.

[bonyetty]

[quote name='little square dot']Is it just me or did the water become really sticky?

I just made my first 1.0.5 water landing using welded pontoons that used to glide nicely across the surface , but no matter how shallow my approach angle or slow my airspeed, as soon as they make contact with the water my pontoons grab on for dear life, taking me from 45 m/s to 0 almost instantly, within about 10m of my touchdown point. My plane somehow survives these violent "landings", but it's not so much water landing as water-crashing. Is there something I can add to/change in the .cfgs to make my pontoons less grabby and more glidey?

Call me old-fashioned, but I tend to prefer non-violent deceleration. :huh:[/QUOTE]

It would have to be a highly engineered real life air craft to land on water at over 150kmph and not quickly decelerate in real life. I believe SQUAD wishes to have the game as close to real life while still being fun. A lower expectation of travel/landing speeds on and in KSP's water would be my recommendation. Seaplanes touchdown at about half the speed of what you described. Also to move an object floating requires moving the mass of the object and the water it displaces, which would be about the mass of the object again each time it travels its own length.

[benjee10]

Yeah, I thin the problem is that you're coming in too fast. Landings are silky smooth at 50m/s or below for me.

[Dawnstar]

My MK1 pods always keep falling when they touchdown in water. The pod, the instruments, and the parachute(s) just keep on going until they hit the bottom of the sea.

A read of the forums indicates that a few other people have the same issue and that most people appear not to. What I haven't yet read is what might be causing this bug and how to fix it. Any thoughts?

[little square dot]

[quote name='bonyetty']It would have to be a highly engineered real life air craft to land on water at over 150kmph and not quickly decelerate in real life. I believe SQUAD wishes to have the game as close to real life while still being fun.

A lower expectation of travel/landing speeds on and in KSP's water would be my recommendation. [/QUOTE]

Thanks for the suggestion.

Unfortunately I started out fairly slow by KSP standards, ie. always <50m/s, typically around 40 m/s, but my floats were catching the water pretty badly at lower speeds. I was advised to come in faster, but quite honestly it doesn't seem to make a difference.

I've tried all manner of approach speeds, from ludicrous to stalling as I contact the water, and it's always the same result. SPLAT! "We're somehow OK!"
The plane lands intact, although I'm not quite sure how it remains in one piece considering that the entire landing covers about 10-20m of water, if that. Sadly it's about the furthest thing from realistic landing physics, and while the old buoyancy model was just plain weird and took forever to figure out, the landing physics were a million times better, so long as your aircraft was equipped to land on the water. The old water seemed to have plenty of surface tension, perhaps too much actually, but I was able ride out the first half of my landings on the backs of my pontoons, gliding across the surface and gradually settling down in a reasonably realistic manner that was quite satisfying.

The new water seems to lack surface tension, making it incredibly awkward to land on in any sort of a half-way smooth manner, and on the flip-side it's far too forgiving with aircraft. I mean, I shouldn't be able to touch the water at or above 100m/s for starters, and if my aircraft decelerates from 70m/s to 0 over a distance of 10m, I should probably to be doing a front-flip...and if I do a front-flip, my aircraft should probably be damaged.

[quote]
Seaplanes touchdown at about half the speed of what you described. Also to move an object floating requires moving the mass of the object and the water it displaces, which would be about the mass of the object again each time it travels its own length.[/quote]

Yes, but real seaplanes land on water that behaves more or less like water.

Landing on water in real life involves coming in slow, cutting your throttle about 20 feet above the water, pulling your nose up, landing as close as you safely can to stall speed. The aft end of the pontoons make contact first, and the surface tension allows you to skim along with your nose slightly raised until you've bled off the lion's share of your speed. The plane doesn't come to rest on the entire pontoon until you're taxiing, essentially.

In KSP on the other hand, your pontoons immediately catch once you touch the water, and it's simply not possible to ride out any portion of the landing on the backs of your floats. Well, it hasn't been for me anyway, and I've seen quite a few videos of people having similarly sloppy landings.

Also, seaplanes can't reasonably be considered a floating object until they cease generating lift and settle onto the water...which doesn't really happen until the landing is all-but complete.

It's all about that surface tension my good man. Have you ever been water-skiing?

[COLOR="silver"][SIZE=1]- - - Updated - - -[/SIZE][/COLOR]

[quote name='benjee10']Yeah, I thin the problem is that you're coming in too fast. Landings are silky smooth at 50m/s or below for me.[/QUOTE]

How much distance do you cover when touching down at 50m/s? I can land reasonably smoothly if a come in REALLY slowly, but I'm still decelerating very quickly... there's just less of it to do.

Also, as mentioned earlier, I've managed extremely smooth high-speed landings with my landing gear down, which is just bizarre to me. My pontoons catch, yet somehow massive tires just glide straight across, haha..

[COLOR="silver"][SIZE=1]- - - Updated - - -[/SIZE][/COLOR]

[quote name='Dawnstar']My MK1 pods always keep falling when they touchdown in water. The pod, the instruments, and the parachute(s) just keep on going until they hit the bottom of the sea.

A read of the forums indicates that a few other people have the same issue and that most people appear not to. What I haven't yet read is what might be causing this bug and how to fix it. Any thoughts?[/QUOTE]

Mods? Is the debug menu acting up? Does your mk1Pod.cfg have CenterOfBuoyancy, CenterOfDisplacement, buoyancy, and buoyancyUseSine values?

I think the underwater parachute thing is normal when they're attached to non-buoyant parts, because parachutes don't auto-cut until their parent stops moving


I actually had the sinking command pod problem several KSP versions ago, but in my case it was something I did poking around in part.cfgs... improper module indexing or something I think.
Reinstalling tends to clear that sort of stuff up. Edited November 27, 2015 by little square dot

[NathanKell]

What's your descent angle?

[Wolf Baginski]

On 22/11/2015 at 4:09 AM, NathanKell said:

CoM = CoL = CoP = Center of displacement (i.e. how we know how much is submerged) = center of buoyancy (where the buoyancy force is applied) = part origin, unless overridden (only a few parts override one or more of those).
There's also a buoyancy scalar--again, only a few parts set it to something other than 1.0.

 (Sine curve explanation snipped: used for essentially cylindrical stuff horizontal in the water)

I am thinking here of a picture or two I have seen of the Gemini capsule, more-or-less axis-horizontal in the water, with added flotation bags and hatches open. There are also pictures of Apollo capsules floating heatshield down, very high in the water, with spherical flotation bags deployed from the nose.

I don't have a figure for the displacement of the Apollo CM, but the total pressurised volume was 10.4 cu. m. with a total mass of 5.56 tonnes. Call it 5 tonnes excess buoyancy.

I haven't found any figure of Gemini, but a simple cone of the same diameter and height is about 4.7 cu. m. and the capsule massed 1.99 tonnes. Mercury had about the same ratio of displacement to mass, while floating heatshield down like Apollo.

I would suppose that, for a capsule designed to float horizontal, or for a ship, there could be an offset from the axis for the CoM and center of buoyancy. But that would make problems in-flight. My recollection is that the Apollo CM used an offset CoM to slightly angle the heat-shield during re-entry, allowing a roll to steer the capsule. Up until CM/SM separation, a slightly offset SM CoM could compensate.

But which direction should you move the CoM or center of buoyancy in?. Which way, within the coordinate system, is up for a horizontally-floating capsule?

 

 

 

 

 

[Wolf Baginski]

On 26/11/2015 at 9:34 PM, bonyetty said:

Also to move an object floating requires moving the mass of the object and the water it displaces, which would be about the mass of the object again each time it travels its own length.

Remember that seaplanes, from the earliest days, used a planing hull-form for the floats. They were not operating as displacement hulls for the actual take-off and landing.

There's an outline of the problems here, in a NASA history.

[NecroBones]

2 hours ago, Wolf Baginski said:

But which direction should you move the CoM or center of buoyancy in?. Which way, within the coordinate system, is up for a horizontally-floating capsule?
 

I'd have to double-check, but I think you want -Z direction for the center of buoyancy for it to float horizontally, with the window/hatch side out of the water. +Y would give you "heat shield down" like the stock pods, -Y is "nose first", etc.

Edited December 19, 2015 by NecroBones

[MaxwellsDemon]

On 11/26/2015 at 6:55 PM, Dawnstar said:

My MK1 pods always keep falling when they touchdown in water. The pod, the instruments, and the parachute(s) just keep on going until they hit the bottom of the sea.

A read of the forums indicates that a few other people have the same issue and that most people appear not to. What I haven't yet read is what might be causing this bug and how to fix it. Any thoughts?

Do you have the Kopernicus mod installed?  I believe the developer said that they haven't managed that bug/"feature" yet.