Vortex modeling?

[Fearless Son]

Okay, this is a rather technical question that is probably best answered by a Squad programmer, but if anyone else knows feel free to chime in.  

I have been thinking a lot about aerodynamics lately, and in particular the vortices that form in the wake of atmospheric compression.  Obviously these are essential for lift, but controlling them is essential for managing drag.  I wanted to know how closely the Kerbal Space Program aerodynamic physics engine models these vorticies.  For example, are there advantages to putting small strakes to break up airflow before it builds to a larger vortex behind the craft?  Will long swept wings generate smaller vorticies than short square ones?  Is there any lift advantage for adding bulkier perturbations in the topography of the airframe on top of it compared to on the bottom?

I know the answers to these as concerns of real-engineering, but I want to know how closely they apply in KSP, since they would impact my design choices.  

Edited December 1, 2016 by Fearless Son

[Gaarst]

I might be wrong, but I don't think KSP models any form of air flow besides the drag defined by the drag cubes. Planes don't stall, radial parts that are clipped still generate drag, radial occlusion or detached shock are not a thing...

[bewing]

Detached shockwaves are a thing, but Gaarst is right about all the rest. Air in KSP is stationary: no wind, no vortices. Swept wings have no advantage over square wings. The drag values of all the wings (except for one bad one) are precisely proportional to their mass, area, and lift.

4 hours ago, Fearless Son said:

 Is there any lift advantage for adding bulkier perturbations in the topography of the airframe on top of it compared to on the bottom?

As far as I can tell, the answer is yes but the effect is small. KSP models "occlusion" for both drag and heating for most parts (but not all of them), and it's strictly modeled based on direction of motion. So, since a plane generally flies with a positive AoA, the stuff on the top of the plane is going to get occluded and the stuff on the bottom won't.

[Fearless Son]

4 hours ago, Gaarst said:

I might be wrong, but I don't think KSP models any form of air flow besides the drag defined by the drag cubes. Planes don't stall, radial parts that are clipped still generate drag, radial occlusion or detached shock are not a thing...

 

1 minute ago, bewing said:

Detached shockwaves are a thing

Yep, I was just about to say @NathanKell established that KSP does model detached shockwaves here:

 

2 minutes ago, bewing said:

Detached shockwaves are a thing, but Gaarst is right about all the rest. Air in KSP is stationary: no wind, no vortices. Swept wings have no advantage over square wings. The drag values of all the wings (except for one bad one) are precisely proportional to their mass, area, and lift.

As far as I can tell, the answer is yes but the effect is small. KSP models "occlusion" for both drag and heating for most parts (but not all of them), and it's strictly modeled based on direction of motion. So, since a plane generally flies with a positive AoA, the stuff on the top of the plane is going to get occluded and the stuff on the bottom won't.

Come to think of it, knowing that KSP does model detached shockwaves might actually make strakes serve their aerodynamic purpose.  If that strake creates a shockwave, it means that the parts immediately behind it are a little more occluded and thus the drag is a little lessened on them.  I am not sure how practical that would be for the purposes of minimizing drag though, since the purpose is to prevent a vortex behind the body, but the game does not seem to care much about what the air does.  Might be helpful for breaking up the drag just ahead of some unusual geometry, though the things most likely to benefit from it probably have null drag as far as the game is concerned.  

[Gaarst]

@bewing@Fearless Son

I now know that detached shockwaves are a thing in terms of heating, but I'm not sure that anyone has demonstrated the effects of detached shockwaves for drag.

In the thread I wrote for drag tests in 1.2 (link somewhere in my sig), there were discussions about whether or not detached shockwaves exist. While I was obviously wrong initially in assuming they didn't exist at all, I have not seen any conclusive proof of their effects in terms of drag. Though I've never done proper tests specifically to give a definite answer. I'm not really sure I could think of a situation where it would be the dominant effect to test it, but I'd be willing to run the tests if I did.

 

Also, I'm not 100% sure that keeping high AoA shields parts on top from drag. I know from experience it does shields them from heat, but the two appear to behave differently in the game.

[bewing]

Truly, I doubt if shockwaves are handled specially in terms of drag at all. I think it just uses a non-linear function of airspeed multiplied by the drag cube multiplied by the density.

As far as AoA goes: I took a base plane with a wheesley engine to 1km, full throttle, level flight, and ran 3 tests. I made the planes be the same mass.

Base plane, top speed = 644 m/s (yes, mach 2). Added lots of crap to the top surface: top speed = 344.2 m/s. Moved the stuff from the top surface to the bottom surface: top speed = 343.6 m/s. So if there is an effect for drag occlusion on the upper surfaces of planes, it's small.

 

 

Edited December 2, 2016 by bewing

[foamyesque]

It is possible to stall out wings in KSP; there's a critical angle of attack where they do a hard transition from a high-lift and (relatively) low-drag mode to an almost-no-lift, high-drag mode. It's really easy to spot if you turn on aerodynamic overlays and swing back and forth between a 30 degree and a 50 degree AoA. Whether or not that matches the actual aerodynamics involved in real-world stalls I leave to others.

Edited December 2, 2016 by foamyesque

[AeroGav]

11 hours ago, foamyesque said:

It is possible to stall out wings in KSP; there's a critical angle of attack where they do a hard transition from a high-lift and (relatively) low-drag mode to an almost-no-lift, high-drag mode. It's really easy to spot if you turn on aerodynamic overlays and swing back and forth between a 30 degree and a 50 degree AoA. Whether or not that matches the actual aerodynamics involved in real-world stalls I leave to others.

I'd say that it does , more or less.   KSP tends to encourage canard designs* , which are hard to stall with a CoL behind the CoM.

It does not model "deep stall" blanketing of the tail surfaces, where the disrupted air from the stalled main wing stops your tailplane from working and being able to push the nose back down.   Some real world aircraft (T tails in particular) suffer with this.  

This all ties in to the Vortex thing,  basically KSP considers every wing part in isolation and does not consider how one wing part might affect airflow in a one next to it.  Sweep angle and wave drag aren't modelled, and all wing parts have the same lift drag ratio, using the same curve for aoa vs lift and drag.

Hence, biplanes fly exactly the same at low and high speed as a monoplane with the same wing area.   Canards don't disrupt the airflow to the main wing behind them, reducing it's efficiency slightly.    Strakes and leading edge droop devices don't enable the wing to fly at higher AoA without stalling.

This is not to say you can't produce more or less efficient aircraft with stock aero :

•    Minimising radially attached stuff, frontal area, open node drag, fuselage AoA, trim drag 
•   Optimising AoA for control surfaces and wings, and wing/control surface size
•   Using low drag adapter parts whenever transitioning from parts of different diameter in a stack

these things all make a huge difference.

[NathanKell]

The heating model doesn't affect drag, correct. Detached shockwaves (or attached ones for that matter) are relevant only in terms of the thermo model.

It is quite possible to stall, but stalls are very gentle and you can sometimes trap yourself with no downpitch moment. This is because of the trickery we use in terms of treating all lifting area as high aspect ratio unswept wings at low speed (and low AR swept wings at high speed).

Unless things have changed recently of course.

[foamyesque]

1 hour ago, NathanKell said:

It is quite possible to stall, but stalls are very gentle and you can sometimes trap yourself with no downpitch moment.

 

That's happened to me more times than I like to count. Can't get down, can't get level, when I'm up I can't get down, can't get my feet back on the ground

[omelaw]

isn't there a tip of putting a comm 16 in front of nosecone to reduce drag? is that a different effect?

[bewing]

3 hours ago, omelaw said:

isn't there a tip of putting a comm 16 in front of nosecone to reduce drag? is that a different effect?

That's to reduce heat transfer into your craft. Only the tip of the comm16 gets really hot when you go fast that way, and that part doesn't transfer heat well into the rest of your vessel. But that trick doesn't work as well in ver 1.2.2. The Comm16s tend to blow up now.

[GoSlash27]

IRT the original question, no. Stock KSP does not model vortices, turbulence, boundary layer disturbances, shear flow in swept wings, or transsonic shockwave attachments/ migrations. KSP is, by necessity, too basic to model such behavior. The general ballistic properties are approximated mathematically, but the dynamic airflow around the airframe is not modeled. Doing that would eat up a lot of processing power and only serve to frustrate new players.

 The specific questions: 1) Wingtip vortices are not modeled, so winglets have no effect. 2) The aspect ratio of the wings has no effect on the flight characteristics whatsoever. All wings are equal in all flight regimes. 3) Radially attaching perturbations do not affect lift, only adds drag. They will induce their full value of drag regardless of how much they are clipped. Furthermore, "physicsless" parts add to the total drag of their parent part at it's center of mass regardless of where they are attached.

 HTHs,
-Slashy

Edited December 15, 2016 by GoSlash27

[Palaceviking]

Kerbal atmospheric flight program. 

[fourfa]

On 12/15/2016 at 8:51 AM, bewing said:

But that trick doesn't work as well in ver 1.2.2. The Comm16s tend to blow up now.

The Place-Anywhere linear RCS thrusters work as well - and they have a higher innate temp spec.  Can be rotated and offset bell-in for more fashion value, or bell-out as a functional retrothruster.

[foamyesque]

3 hours ago, fourfa said:

The Place-Anywhere linear RCS thrusters work as well - and they have a higher innate temp spec.  Can be rotated and offset bell-in for more fashion value, or bell-out as a functional retrothruster.

 

I use a pair of 0.625 nosecones, myself. It makes a nice niche for nose-mount RCS units to slot into and doesn't look as dumb as the antenna thing.