Aerodynamics in KSP2?

[Frozen_Heart]

So the aero system in KSP leaves something to be desired (even with how much better it is now than it once was), and one thing which I'm hoping will change for KSP 2 is a complete overhaul of the aerodynamics system. For example rather than wings acting as flat plates, it would be nice for them to act as airfoils. The difference seems pretty academic until you try building an autogyro in the stock game and find that things don't work how they should.

Do people think it's likely that the aero system will get overhauled? Or act in the same way as it does in KSP 1?

[K^2]

The difference between a flat plane and cambered wing is primarily the critical AoA and sharpness of stall onset. In that aspect, the KSP wings are already simulated as cambered, albeit, as if the camber is symmetric.

The asymmetric camber is almost entirely equivalent to increasing effective AoA while keeping the same chord and aircraft pitch. You can get that in KSP by simply angling the wings. It looks a little ugly, and getting parts that have built-in asymmetric camber for a bit of extra lift at zero AoA would be great, but it's honestly a component/design fix, rather than any change in aerodynamics.

The only bit of real physics that's lost on an autogyro blade is the fact that camber/AoA change throughout the blade, so you end up with a driving and driven region. You should be able to achieve this by building your blades in sections. Again, it's not really something you can fix with better aerodynamics. Games that focus on helicopter/autogyro physics simply simulate their blade in multiple sections. And maybe having stock blades that can be simulated in sections would be nice, but it's not really an aerodynamics change.

[Incarnation of Chaos]

Developers confirmed they're still using the same horrible system of drag cubes for KSP2, so if you want a decent approximation of aerodynamics that works for anything that flies in a direction other than straight up then you're going to be waiting for mods.

[Blaf]

That's something i don't get. They said they won't be using anything like FAR, as it would be more complex for newcomers, but i actually find FAR to be more intuitive than the stock model

[Frozen_Heart]

oof sad to hear it will still be the stock drag model. It's definitely not intuitive at all.

I swear it doesn't even do stalling at all.

[Incarnation of Chaos]

1 hour ago, Blaf said:

That's something i don't get. They said they won't be using anything like FAR, as it would be more complex for newcomers, but i actually find FAR to be more intuitive than the stock model

FAR was more intuitive for me, because as a kid i flew gliders, paper airplanes and read about all kinds of aircraft in my downtime. Then in my teens i got into simulators and the like, so KSP felt so strange to me because i knew that planes just don't do that.

But for someone who doesn't have that kind of background, and just wants to make something that'll fly. I could see how KSP's stock aero might strike them as more intuitive.

Does that mean i think they're justified? Not really, they could have a "Advanced" aerodynamics mode that would toggle FAR-like behavior instead of the stock system. But at this point I'd imagine they're far more worried about just shipping a game than anything else.

58 minutes ago, Frozen_Heart said:

oof sad to hear it will still be the stock drag model. It's definitely not intuitive at all.

I swear it doesn't even do stalling at all.

It's plenty possible to stall in stock aero, but you almost have to FORCE it to happen due to how much excess lift everything has compared to FAR.

[Guest]

Yeah this is one of the decisions I just don’t understand. Stock aero is just weird, and gets incomprehensibly weird if you try to do certain somewhat complex things, like shielding things from the airflow using anything other than a fairing or a cargo bay. I found FAR much more intuitive in every way — it’s just about as easy to do easy things (pointy, nose-heavy rockets, simple aeroplanes), and while harder things like hypersonic aircraft are still hard, they’re hard in a logical way. 

I think the difficulty with a more realistic aero model is that it doesn’t mesh all that well with the ”Lego” aspect of the game. You really ought to be able to sculpt an aerodynamic shape for efficiency, and you can’t do that if you just bolt bits together. Even so it’s not a showstopper IMO, FAR works really well and if you want to make things a bit easier that can be done by tweaking the physics.

So as a reformed flight-sim geek I would really appreciate just slightly more realistic aerodynamics. FAR demonstrates that it doesn’t make things any harder for beginners, but it does make challenges more interesting and ”fairer” for more advanced tinkers. 

So team if you’re listening, please reconsider putting in better aerodynamics. Pretty please?

[Frozen_Heart]

Wing sweep doing nothing at all is another jarring aspect to it. Apart from changing the CoL, having completely straight wings or heavily swept wings doesn't effect how the plane flies at all.

[Kerbalwerks]

This is what bugs me the most: " incomprehensibly weird if you try to do certain somewhat complex things, like shielding things from the airflow using anything other than a fairing or a cargo bay. "

[Incarnation of Chaos]

9 hours ago, Kerbalwerks said:

This is what bugs me the most: " incomprehensibly weird if you try to do certain somewhat complex things, like shielding things from the airflow using anything other than a fairing or a cargo bay. "

Which mind you just is a byproduct of the drag cube model, you have to have magical shielded parts.

Likewise in FAR... it's just a byproduct of the approximation actually taking the shape into account. It doesn't have to have magical shielded parts, because it can determine if something is shielded from the geometry.

[K^2]

20 hours ago, Frozen_Heart said:

Wing sweep doing nothing at all is another jarring aspect to it. Apart from changing the CoL, having completely straight wings or heavily swept wings doesn't effect how the plane flies at all.

Swept wing only makes an impact if the airflow gets close to speed of sound. Since KSP doesn't simulate anything relating to speed of sound, swept wings make no difference.

Making a simulation that faithfully simulates impact of swept wings is actually pretty hard, as it has to simulate the way air flows around the wing, but like most things, so long as you don't need it to be precise enough for engineering work, it can be faked rather convincingly. But it is a packaged deal. If you want swept wings to be meaningful at transonic speeds, you'll have to deal with a whole host of problems if you plan to actually go supersonic. Wave drag, shock cones, loss of control due to separation... Personally, I'd welcome the added challenge, but I can see how it can be intimidating to a lot of people and why Intercept might be hesitant to go that route.

[Frozen_Heart]

1 hour ago, K^2 said:

Swept wing only makes an impact if the airflow gets close to speed of sound. Since KSP doesn't simulate anything relating to speed of sound, swept wings make no difference.

Making a simulation that faithfully simulates impact of swept wings is actually pretty hard, as it has to simulate the way air flows around the wing, but like most things, so long as you don't need it to be precise enough for engineering work, it can be faked rather convincingly. But it is a packaged deal. If you want swept wings to be meaningful at transonic speeds, you'll have to deal with a whole host of problems if you plan to actually go supersonic. Wave drag, shock cones, loss of control due to separation... Personally, I'd welcome the added challenge, but I can see how it can be intimidating to a lot of people and why Intercept might be hesitant to go that route.

I mean FAR manages a decent enough rough approximation. If the game was designed around building aircraft which will never reach the speed of sound I could understand handwaving it. But any spaceplane in KSP will be dealing with transonic and hypersonic flight, so it would be nice to see some considerations made for designing around it. Hypersonic straight winged biplanes is pretty jarring in the stock game.

I'm glad we don't have the souposphere still though. Back when adding nosecones made the craft more draggy.

[K^2]

17 minutes ago, Frozen_Heart said:

I'm glad we don't have the souposphere still though. Back when adding nosecones made the craft more draggy.

Oh yeah. The drag model might be primitive, but at least it makes some consideration for occlusion now.

[Bej Kerman]

On 1/16/2021 at 12:20 PM, Incarnation of Chaos said:

Does that mean i think they're justified? Not really, they could have a "Advanced" aerodynamics mode that would toggle FAR-like behavior instead of the stock system. But at this point I'd imagine they're far more worried about just shipping a game than anything else.

One of the reasons why I hate this development model of 'pump everything out at once behind curtains', unlike KSP 1's 'follow development intimately and help guide the developers while they get time to add whatever needs adding'. There's plenty of good reasons to not rush, but then the game is being run by corporate overlords that will cut any number of features if their silly hype-retention techniques lost any more footing than it already has. /tangent/ Besides, KSP 2 has lost quite a bit of its apparent hype over the past few years, but KSP 1 wasn't being kept like it's some sort of secret. There wouldn't be a very exciting fanbase if Squad just kept it up until 2015. If KSP wasn't dawdling with load times that rival Sonic '06, with both games having loading times between redundant scenes you see for 2 seconds, and bugs that would make Fallout 76 blush, then I wouldn't be waiting for the replacement.

[Guest]

57 minutes ago, Bej Kerman said:

One of the reasons why I hate this development model of 'pump everything out at once behind curtains', unlike KSP 1's 'follow development intimately and help guide the developers while they get time to add whatever needs adding'. There's plenty of good reasons to not rush, but then the game is being run by corporate overlords that will cut any number of features if their silly hype-retention techniques lost any more footing than it already has. /tangent/ Besides, KSP 2 has lost quite a bit of its apparent hype over the past few years, but KSP 1 wasn't being kept like it's some sort of secret. There wouldn't be a very exciting fanbase if Squad just kept it up until 2015. If KSP wasn't dawdling with load times that rival Sonic '06, with both games having loading times between redundant scenes you see for 2 seconds, and bugs that would make Fallout 76 blush, then I wouldn't be waiting for the replacement.

All of that is the byproduct of one game being developed by a "traditional" studio, and the other being an accidental indie hit.

They also didn't plan to reveal the game 2 years before release.

 

Back to topic, has any recent statement confirmed that they still plan to put in the exact same model of KSP1?

If I'm not wrong the "confirmation" that they would use the same exact model was made back when the game was just being announced during the early Pax interviews.

I need to go back to the notes I've taken on the "magic podcast of all answers".

[OHara]

I think there reason to believe more realistic aerodynamics is possible in KSP2, enough that it is worthwhile being specific about what we would like to see.

A while ago I reverse engineered KSP1 aero somewhat, and also experimented with FAR and read through its code.  KSP1 aero is not drastically simpler than FAR; I guess it would take half as many pages to document.  The differences seem due to history: KSP1 aero was gradually improved starting with a non-serious model in a simple game about rockets, while FAR was developed by one person who knows standard aerodynamics. 

FAR estimates the drag-versus-speed for the entire craft (updating occasionally during flight) using standard aerodynamics heuristics based on the sections presented to the assumed direction of airflow.   KSP1 estimates for each part its drag and its dependence on speed, based on some analysis of its shape, for six cardinal directions of potential airflow, and then has rules (that we don't all like) for combining the parts to get drag of the craft.

So I can hope that a re-working gives opportunity for improvement over KSP1 without wasted effort.  

I remember KSP2's Creative Director saying he wanted KSP2 aero to be not too different to KSP1, for purposes of easy transition between games.   That might imply keeping the very soft stall, or the artificially high lift and drag at low speeds (so players can make craft that work over an unrealistically large range of speeds).  It might also imply that he doesn't see aero as a high priority when he wants primarily to extend the interstellar aspects of the game, and to smooth the learning curve.

Those gameplay desires can be had by adjusting parameters in a realistic model,  just as the game-play adjustments can be removed in KSP1.  Years ago on the forum I posted a MM patch to give KSP1 planes a realistic stall and spin.  Ferram4 posted a patch to remove the artificially high low-speed lift.  So I think we can hope for a realistic model, tweaked to let KSP2 have an easy transition.

On 1/15/2021 at 10:28 PM, Frozen_Heart said:

For example rather than wings acting as flat plates, it would be nice for them to act as [cambered] airfoils.

I like the symmetric airfoils in KSP, because camber is a subtle thing, and I wouldn't notice a cambered wing to know which way is up in a computer game.  Symmetric airfoils are easier to use as rocket fins.   But I think we would both like a simulation of interactions between wing-parts.  Then we could build a cambered airfoil out of two sections (then I'll know which way is up because I made it that way). And then flaps would have a realistic effect.

Edited January 17, 2021 by OHara
add links

[KerbMav]

I like "what you see is what you get" in this case.

To use open attachement nodes that are not visible in the actual sandbox and only in our tool shed still seems weird to me.

If I see a flat surface I expect it to be more draggy than any form of roundness or pointiness, and that is what the aerodynamic system should work off.

Any difficulty increasing things like mach effects etc. - optional to me, as it might introduce too much things to consider, at least on lower difficulty settings.