[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[monstah]

 

16 hours ago, Sol Invictus said:

@tetryds What else would be this thread for then? Only bug reports and release announcements?

No, but these are perfectly valid examples:

14 hours ago, paul23 said:

I am just wondering what the cl-alpha and cl-cd graphs of the wings FAR uses are 

17 hours ago, Nnimrod said:

You know you could just make a mod that changes the mass of various aero parts to realistic values are see how far feels then.

I agree with @tetryds's point. If the discussion were "how to better design airplanes that work well on FAR", that's one thing; but a discussion about whether or not real airplanes are designed around the ground effect should have a place of its own.

And in reply to the second quote, there's a mod called SMURFF, I think. It's designed around stock parts (and mod parts balanced to stock parts), to make them lighter (supposedly closer to real-life counterparts) to use in RSS/RO/mods that make the stock system large.

[Sol Invictus]

Can we just all stop complaining about stuff and enjoy each others contribution? Personally I find discussion about ground effect highly interesting and relevant. If by any chance it turns out that ferram doesn't approve it, then I'll understand it, but until then all that you achieve by complaining about it is adding unnecessary cluster and negativity.

[DoctorDavinci]

3 minutes ago, Sol Invictus said:

Can we just all stop complaining about stuff and enjoy each others contribution? Personally I find discussion about ground effect highly interesting and relevant. If by any chance it turns out that ferram doesn't approve it, then I'll understand it, but until then all that you achieve by complaining about it is adding unnecessary cluster and negativity.

There have been well respected members of the community as well as moderators coming in here and all have been saying the same thing which boils down to this: 

Mod threads are for providing help to users, posting mod updates and discussing the mod in question

Ground effect is not relevant to this mod nor any mod or even KSP itself since it is not coded into the game and it's one thing that the phrase 'there is a mod for that' does not apply ... this conversation deserves its own thread not only due to the interest in the subject, but also out of respect for ferram4 (intentionally not pinged) and FAR users everywhere

Don't get me wrong, the discussion is an interesting read, however I agree with certain members of this community by saying give the subject a place of its own and leave the FAR thread to things directly related to FAR

That is all ...

[Sol Invictus]

50 minutes ago, DoctorDavinci said:

Ground effect is not relevant to this mod nor any mod or even KSP itself since it is not coded into the game and it's one thing that the phrase 'there is a mod for that' does not apply ... this conversation deserves its own thread not only due to the interest in the subject, but also out of respect for ferram4 (intentionally not pinged) and FAR users everywhere

Well, I think it is very much relevant to this mod, as it's discussed in the context of making FAR even better than it already is. FAR stands for "Ferram Aerospace Research", and people who are discussing ground effect are doing exactly that - aerospace research, unlike some other people who think they're some authorities in the field of what is allowed to discuss here. If you truly care about respecting ferram, then you would leave to him deciding what is relevant and what is not. All that is accomplished otherwise is that new users will be less likely to contribute and interesting posts will be lost in an avalanche of uselessness.

Edited December 15, 2016 by Sol Invictus

[DoctorDavinci]

16 minutes ago, Sol Invictus said:

Well, I think it is very much relevant to this mod, as it's discussed in the context of making FAR even better than it already is. FAR stands for "Ferram Aerospace Research", and people who are discussing ground effect are doing exactly that - aerospace research, unlike some other people who think they're some authorities in the field of what is allowed to discuss here. If you truly care about respecting ferram, then you would leave to him deciding what is relevant and what is not. All that is accomplished otherwise is that new users will be less likely to contribute and interesting posts will be lost in an avalanche of uselessness.

Wow, I remember not so long ago when ....

On Sunday, October 23, 2016 at 2:23 PM, Sol Invictus said:

I really don't want to be a douche, but isn't there a like button for that? When I'm getting a notification on new content in this thread I'm expecting info on aerodynamics and/or updates on FAR, not this kind of nonsense.

I'll just leave this here .....

On Wednesday, November 16, 2016 at 6:57 PM, ferram4 said:

<snip>

Also, since it seems necessary, I fully endorse all of tetryds' post, as well as @DoctorDavinci's.  Follow their example, let's not have a repeat of the leadup to FAR for 1.0.

 

[Sol Invictus]

I see no contradiction there. Discussion about ground effects is exactly the kind of "info on aerodynamics and/or updates on FAR " that I like to see the most here. Unlike posts consisting of a single "lol" or complaints about non-relevant content. True, I'm guilty of some of these as well, but that doesn't make it any better. Nonetheless, if ferram endorses all that you say, then I have no choice but to comply. So long, and have a nice day.

[tetryds]

@Sol Invictus the thing is, ferram already knows about the ground effect, and I already said that if it comes that will happen with/after the wing overhaul, so discussing it here does not bring any useful information to the mod or to the users of the mod, neither does discussing its impact on designs here.

The FAR Craft Repository was created exactly for that purpose (discussing and sharing designs), it is even linked on the OP. And yes it was meant to get the cluster out of here.

Overall discussions about aerodynamics don't fit here either, as you said yourself its FERRAM'S aerospace research not community's.

Best thing to do is moving to the other thread or making a new one at Science & Spaceflight.

[monstah]

Again, I agree with tetryds. Aero discussion in general is fine, but not here. This place is about the mod, how to use it, how to improve on it, how to make planes that fly pretty damn well on it, and whether such and such real-life aerodynamics considerations might or might not get implemented. It not about how such and such real life airplane performs because of real life aerodynamics.

 

Lets keep on topic, everyone.  FAR, that is. 

[paul23]

14 hours ago, Drew Kerman said:

try this, a bit dated but still valid I believe

 

Skimming through the video (why would anyone make a video about this?) and I can not see anything that looks like an expected Cl-Alpha curve (or CL-alpha). I expect something resembling:

(Given the airfoil is probably symmetrical so I expect it goes to the origin too). Well this made me interested in what the underlying theory is for FAR aerodynamic calculations.

[tetryds]

@paul23 yes, FAR has that curve, check out the static curves, it's light blue.

But because we usually swipe a much larger AoA range than we can find on most graphics, the stall point shows up kinda early on it.

If your airplane is not very aerodynamically efficient it's going to be shorter too.

And remember that it looks like that for subsonic, changing a lot when supersonic.

Edit: also, if you don't have built in AoA or a large wing to add cambering it will start at Cl~=0 for 0deg AoA.

Edited December 15, 2016 by tetryds

[cantab]

The Cl versus Angle-of-attack curve is right there on the Static page in the FAR Analysis Tools, in blue. Now it might "look odd" for a few reasons. The default AoA range is quite small and often the maximum lift point is off the chart. It's for the whole aircraft not just the wing, unless you run the analysis with a vessel that's just a wing by itself. Zero AoA is I think defined based on the vessel's root part, commonly but not always the cockpit, and that could lead to very odd-looking graphs in some cases. And if the analysis is being done for supersonic speeds then of course it'll look unusual compared to subsonic. Oh, and there was at least one FAR release with a big bug in the wing behaviour.

[Drew Kerman]

5 hours ago, paul23 said:

why would anyone make a video about this?

You're saying everyone should just naturally understand how to use the FAR graphs and stability derivatives?

[Matuchkin]

How the heck do you guys get FAR? It's not updated.

[tetryds]

4 minutes ago, Matuchkin said:

How the heck do you guys get FAR? It's not updated.

If you don't already know how to you should not get it, as you said yourself, it's not updated.

[ferram4]

Since the discussion about ground effect has come up again, I'll have to make my position on it clear again.

FAR should have ground effect and it is a relevant feature to any kind of aerodynamic system.  The problem is that the way FAR is set up it is currently impossible to implement in a way that will be consistent with vehicle scale.  That is, a Cessna 2 wingspan lengths above the ground should have similar ground effect to MEGA CESSNA, KING OF THE SKIES that is also 2 wingspan lengths above the ground, despite the fact that the normal Cessna will only be a couple meters off the ground and MEGA CESSNA will be several hundred.  At least in terms of the multipliers on lift coefficient and drag coefficient and the reduction in downwash strength.

One standard way of modelling ground effect in the real world is to either assume the plane is wings level and add a simple multiplier based on height / wingspan (obviously overly simple, doesn't account for heavy dihedral / anhedral effects, flying banked, and more pedestrian concerns like defining the proper wingspan considering this is all abstracted away in FAR currently).  Another more sophisticated method is to take the same bound vortex system used to model the wing lift and mirror it under the ground, forcing the ground to be a streamline of the flow.  That last one is what I want to do, but unfortunately that vortex system is currently abstracted too far to directly do anything like that (and requires a code overhaul) and also assumes that the ground under the length of the plane is, if not level, at least flat... around hills at very low altitude it is possible that you could get the bound vortex on the wing and the imaginary mirrored vortex to intersect if using a simple mirror right below the plane, and I have no idea what could come out of that model.  Probably nothing good, considering the bound vortices are already lines of divide-by-zero that are carefully avoided and cancelled to get useful data out of the rest of the flowfield and I expect that any implementation will have many detours into NaN-space before that's working.

So the short answer is that FAR doesn't have working ground effect because I have no good, non-hack, proper scaling options to implement it.  If you're absolutely dying to get that implemented and want to help, the thing to do is to look into making a PR to replace the current wing code with a good numerical lifting line approximation (including mach effects) and also work on a way to put together entire wing shapes from many intersecting / nearly intersecting polygons, which is where I'm currently stumped on the wing overhaul.  Then, once that's done, ground effect will be trivial to implement.

 

Edit:

Good news everyone!  I believe that I have fixed the irritating multiplying buttons after scene load, the shader shenanigans, and I'm somewhat sure that there are no other new bugs in FAR itself or in RealChuteLite.  Those who know where to go to get the dev build, please go get it, get everything in the GameData directory, especially GameData/Shaders/farshaders.ksp, make sure you're running MM 1.7.5 and MFI 1.2.3 and please test it to see if it works.  In particular, I'd like some Linux and Mac users to try this to make sure that shaders play well with their stuff and just on Windows.

Also, generally try to break things.  There are some things I want to test to be sure are correct (water drag mostly) but due to that I don't know if I'll have the time for other testing before I try to get a release out, so please, play with it and make sure that no entrances to NaN space or 0-drag land have opened while I wasn't looking.

Edited December 19, 2016 by ferram4

[Maeyanie]

7 hours ago, ferram4 said:

That is, a Cessna 2 wingspan lengths above the ground should have similar ground effect to MEGA CESSNA, KING OF THE SKIES that is also 2 wingspan lengths above the ground, despite the fact that the normal Cessna will only be a couple meters off the ground and MEGA CESSNA will be several hundred.  At least in terms of the multipliers on lift coefficient and drag coefficient and the reduction in downwash strength.

One standard way of modelling ground effect in the real world is to either assume the plane is wings level and add a simple multiplier based on height / wingspan (obviously overly simple, doesn't account for heavy dihedral / anhedral effects, flying banked, and more pedestrian concerns like defining the proper wingspan considering this is all abstracted away in FAR currently).

I certainly don't pretend to be an expert (I know barely enough to be amazed at what FAR accomplishes already), but couldn't you do something like use the width of the voxel-model for wingspan, then modify it based on Vessel.srfRelRotation for bank? Though that would still be a rough hack, and wouldn't account for sloping terrain in the slightest. You could use a sliding window of the height over a couple of frames for current slope, but that wouldn't do a thing for looking ahead.

Dihedral/anhedral would be a problem too... a simple approach like comparing the height of the farthest-out voxels vs. the center voxel might work for simple wings, but would be severely messed up by winglets and similar fairly-common features. Even worse for some sort of P-38-like craft.

Sometimes it's amazing how easy something like "identifying the start and end of wings" can be to humans, but how hard it can be for computers... I suppose we can live without ground effect modelling, and be glad for what we do have instead.

 

Edit: I spent some time thinking about this, and how to go about programming it... the more I think, the more problems I find, and the less I think it would work. So, never mind, please disregard me.

 

Edit 2: Also read a mathematical document on ground effect, and while I guess I sort of knew this I hadn't had it spelled out... there's really 2 parts to ground effect, chord-dominated (increased lift due to pressure under the wing) and span-dominated (decreased drag due to wingtip vortex flattening). The latter is the one that's really hard to model, the former could probably be done with FAR the way it is now by iterating through the voxel model and calculating the cord of each segment, then increasing the lift based on that. The formula they got for a NACA-M15 (fairly flat-bottomed) airfoil was: Lift(ground) = Lift * (height / chord) ^ -0.11

Though, that wasn't accounting for mach effects, since they were looking at it for WiGE boats which weren't expected to exceed mach 0.1. Of course, anyone flying at high speeds within their wing's chord of the ground probably has other problems to worry about.

Even without including the span-dominated ground effect (which would be WAY harder to program), this would help. The most serious problem I see is it would be quite computationally-expensive and have to be done every frame. Which unfortunately is a problem.

Edited December 19, 2016 by Maeyanie

[Nansuchao]

Having ground effect would be great, but most of the times, the issues are just that the Runway it's too short and the landing wheels are kinda broken. If those two things were working as they should, we will complain much less about the ground effect.

[kcs123]

16 hours ago, ferram4 said:

Those who know where to go to get the dev build, please go get it, get everything in the GameData directory, especially GameData/Shaders/farshaders.ksp, make sure you're running MM 1.7.5 and MFI 1.2.3 and please test it to see if it works.  In particular, I'd like some Linux and Mac users to try this to make sure that shaders play well with their stuff and just on Windows.

Also, generally try to break things.  There are some things I want to test to be sure are correct (water drag mostly)

Does github is updated or it shows obsolete files only for me ?
It shows that latest commit is ea259af... from October 11. I doubt that is intended last update. There is nothing new under releases either. Only last official update 0.15.7.2. is shown there.

I can try shaders on both, windows and linux, but I don't thonk that GitHub shows proper files for me. farshaders.ksp is 10 month old on GitHub.

EDIT:

Disregard question, Akira_R already answered.

Edited December 19, 2016 by kcs123

[Akira_R]

7 minutes ago, kcs123 said:

Does github is updated or it shows obsolete files only for me ?
It shows that latest commit is ea259af... from October 11. I doubt that is intended last update. There is nothing new under releases either. Only last official update 0.15.7.2. is shown there.

I can try shaders on both, windows and linux, but I don't thonk that GitHub shows proper files for me. farshaders.ksp is 10 month old on GitHub.

You aren't looking at the correct branch.

[Flashblade]

1 hour ago, kcs123 said:

Does github is updated or it shows obsolete files only for me ?
It shows that latest commit is ea259af... from October 11. I doubt that is intended last update. There is nothing new under releases either. Only last official update 0.15.7.2. is shown there.

I can try shaders on both, windows and linux, but I don't thonk that GitHub shows proper files for me. farshaders.ksp is 10 month old on GitHub.

There is no update per se. You have to look elsewhere. Since nobody is suppossed to talk about where to go I feel I can't direct you any further than that and ferram4 specifically said he would like the people who already know where to get it to grab that.

BTW @ferram4 happy to oblige. Can't give you anything on the shader thing since I only have a Windows system but I will look at the other stuff. Mayhaps it is time for me to build my first sea-plane

Edit I will post a preliminary report in the relevant discussion on github.

Edited December 19, 2016 by Flashblade
Preliminary report

[Freds]

WOW! FAR 'Update'!! 
I guess my Factorio base expansion plans will have to Wait so i can play KSP

I have a spaceplane near Duna, and another prototype meant for an Eve land and return mission. The prototype is not ready for Eve, but flyes well in Kerbin! If i see any problems i post here. 

[soulsource]

Just to let you know: I still get multiple FAR buttons. Apart from that it seems to run fine up to now.

[kcs123]

1 hour ago, Akira_R said:

You aren't looking at the correct branch.

Thanks, I never go in detail how ferram organize his stuff. That is more than enough info.

Is it OK to post any findings here, or is it restricted to github ?

[Flashblade]

Just now, kcs123 said:

Thanks, I never go in detail how ferram organize his stuff. That is more than enough info.

Is it OK to post any findings here, or is it restricted to github ?

As ferram4 puts it: "Please consider being awesome and post your issues on github", but he never said that it is restricted to there.

[Matuchkin]

On 12/18/2016 at 10:52 PM, tetryds said:

If you don't already know how to you should not get it, as you said yourself, it's not updated.

I want to get it, and just because I do not know how doesn't mean I shouldn't. Can I have a serious reply? I mean, there's a new version released (all of you have it), and for some reason I can't find it.

Edited December 20, 2016 by Matuchkin