What Aerodynamic Features would you like to see added?

[Northstar1989]

Yeah, actually, that's a much better solution. Change it so a difficulty option is "parts do/don't explode on > max temperature) and then you don't have to worry about toggling mach effects.

I like that idea. That captures what I was trying to describe perfectly.

Glad to see such a vibrant discussion about this thread. Maybe the devs will take notice...

Regards,

Northstar

[Northstar1989]

Mach Effects on Stability: - yes, should be implemented in.

This is the mach effect that scares many players away from mods like FAR.- no, it's not. Aerodynamic failures are.

I think you mis-interpreted what I wrote by not reading it carefully and precisely. I said this is the Mach effect that scares many players away from FAR. Aerodynamic Failures, while they interact with and can be caused by Mach Effects, are not Mach effects... (and can easily occur even at subsonic speeds)

Aerodynamic Failures: - shouldn't be part of a stock game, just like random part failures aren't.

There's nothing "random" about Aerodynamic Failures. It's cool if you don't want them, but I feel the need to say this:

If you try to fly your rocket with a 45-degree Angle of Attack, you're going to get them. If you follow sane/rational principles of flying your rocket, you probably won't have to deal with them at all. Sure, they'll abolish the advice "Straight up until you hit 10 km, and then bank hard to 45 degrees", but IMHO that's a GOOD thing...

In short, Aerodynamic Failures are something you only have to deal with if you do something wrong. And, after all, explosions are !FUN!

Regards,

Northstar

Edited October 8, 2014 by Northstar1989

[TheGatesofLogic]

If only performance wasn't an issue then I would love to see an implementation of blade element theory... But as it is that would be a very bad idea...

Regarding Aerodynamic failures: comparing them to random part failures is exceedingly ignorant. It's the same principle as solar panels falling off while flying through atmosphere, or parts of rockets being crushed by the ground because you're going too fast. If you think about it this way it's the only thing that makes sense.

Edited October 8, 2014 by TheGatesofLogic

[wizzlebippi]

I think you mis-interpreted what I wrote by not reading it carefully and precisely. I said this is the Mach effect that scares many players away from FAR. Aerodynamic Failures, while they interact with and can be caused by Mach Effects, are not Mach effects... (and can easily occur even at subsonic speeds)

There's nothing "random" about Aerodynamic Failures. It's cool if you don't want them, but I feel the need to say this:

If you try to fly your rocket with a 45-degree Angle of Attack, you're going to get them. If you follow sane/rational principles of flying your rocket, you probably won't have to deal with them at all. Sure, they'll abolish the advice "Straight up until you hit 10 km, and then bank hard to 45 degrees", but IMHO that's a GOOD thing...

In short, Aerodynamic Failures are something you only have to deal with if you do something wrong. And, after all, explosions are !FUN!

Regards,

Northstar

1. Mach effects on stability - Ever heard of Mach Tuck? Basically, as you accelerate through the transonic region, the center of pressure shifts aft. This increases pitch control forces and as quite a few Leer drivers discovered in the 70's, it can quickly overpower the pilot. I never want to see such a thing implemented in game.

2. Aerodynamic Failures - There are those who fear flutter because they don't understand it and those who fear it because they do. Again, something I never want to see implemented in game and a key reason why I hate FAR.

3. Aside from actual shape/profile based drag and something like deadly re-entry, I don't think KSP needs a healthy dose of reality. More realistic tendencies would be nice, but absolute realism is unatainab

[Basilo1146]

See, the only thing I dislike about FAR is also in NEAR. I think that the drag model needs serious overall and that mach effects are relevant to the game, but at the same time I don't want to change lift and similar things so that aircraft that fly in the current aerodynamic system would be notably more unstable in the newer one. KSP, while a good simulator, should not become too difficult with regards to plane stability. On a similar note, it seems far harder to make a stable plane in FAR and NEAR than it does IRL, but that is just what I think.

[Sky_walker]

I think you mis-interpreted what I wrote by not reading it carefully and precisely. I said this is the Mach effect that scares many players away from FAR. Aerodynamic Failures, while they interact with and can be caused by Mach Effects, are not Mach effects... (and can easily occur even at subsonic speeds)

I understand just fine what you said, and I still think that what scares people off most are aerodynamic failures.

There's nothing "random" about Aerodynamic Failures.

To quote you: "I think you mis-interpreted what I wrote by not reading it carefully and precisely." - I never said anything about aerodynamic failures being random. What I was referring to were random failures. As in "dang it" mod.

1. Mach effects on stability - Ever heard of Mach Tuck? Basically, as you accelerate through the transonic region, the center of pressure shifts aft. This increases pitch control forces and as quite a few Leer drivers discovered in the 70's, it can quickly overpower the pilot. I never want to see such a thing implemented in game.

2. Aerodynamic Failures - There are those who fear flutter because they don't understand it and those who fear it because they do. Again, something I never want to see implemented in game and a key reason why I hate FAR.

3. Aside from actual shape/profile based drag and something like deadly re-entry, I don't think KSP needs a healthy dose of reality. More realistic tendencies would be nice, but absolute realism is unatainabe

1. This would be solved by SAS

2. See Northstar1989? Told you that what scares people off from FAR are aerodynamic failures.

3. "absolute realism is unatainabe" - agreed. And lucky enough - noone asks for it. But there's much more that can be done to improve KSP while at the same time adding to the realism - see "how to improve KSP" in my signature.

Edited October 10, 2014 by Sky_walker

[wizzlebippi]

1. No properly designed supersonic aircraft should exhibit mach tuck, therefore it really has no place in ksp. Also, you need to quit thinking of SAS as a stability band aid. It's a simplistic attitude hold autopilot because it removes the need to actually fly your craft.

2. Unless you tell players at what dynamic pressure their ship will fly apart at, aerodynamic failures are just punishment for players. And while we're on the realism side of things, not many aircraft can achieve their max dive speed in level flight. Yet in ksp, I've yet to build one that can't while climbing using FAR.

3. If you want to make ksp more realistic, you need to figure out why it's broken in the first place. Basing all aero forces off of airspeed/mach would be a good start.

[Guest]

A nice and simple improvement I'd love to see would be wind noise based on speed and angle of attack.

Would it be possible to have a visual effect for the turulence? Starting for example when your drag exceed your lift (high aoe or high speed or both), and whose shape is based on its center of lift?

[klgraham1013]

This thread reminds me how smart this community is. I love it. On topic: I play with NEAR and DRE and think re-entry heat is vital for all but the easiest difficulty settings. I still believe KSP should be a space flight simulation and planning for re-entry is a vital piece to that puzzle.

[sherkaner]

I tend to think that there's not much point in us debating implementation details, since Squad (and perhaps a precious few modders like Ferram) are about the only ones with the basis to consider that with any confidence. In terms of what a player actually experiences, I think the focus should be on an aerodynamics model that adds fun through a sense of reality.

Note that "reality" is not necessarily the same as having "realistic" aerodynamics. But I think people love that when something in KSP is hard or counterintuitive (such as orbital dynamics), once the hurdle is overcome there is a feeling of having mastered something "real" (you can read about how the Gemini astronauts had difficulty with the same thing!). However when the best way of doing something is more about exploiting or avoiding quirks in the game itself (such as taking nose cones off rockets because they actually add drag), the player generally feels let down. I think that's what's ultimately behind why so many people here get passionate about this aerodynamics topic.

[Las-pen]

A boom when going supersonic.

[Northstar1989]

I tend to think that there's not much point in us debating implementation details, since Squad (and perhaps a precious few modders like Ferram) are about the only ones with the basis to consider that with any confidence. In terms of what a player actually experiences, I think the focus should be on an aerodynamics model that adds fun through a sense of reality.

Squad listens to the forums now and then, so it's not necessarily a bad idea to discuss things here, hoping they'll pay attention...

Note that "reality" is not necessarily the same as having "realistic" aerodynamics. But I think people love that when something in KSP is hard or counterintuitive (such as orbital dynamics), once the hurdle is overcome there is a feeling of having mastered something "real" (you can read about how the Gemini astronauts had difficulty with the same thing!). However when the best way of doing something is more about exploiting or avoiding quirks in the game itself (such as taking nose cones off rockets because they actually add drag), the player generally feels let down. I think that's what's ultimately behind why so many people here get passionate about this aerodynamics topic.

Realistic aerodynamics ARE intuitive. Many KSP players come in with an accurate understanding of physics. Some even with pilot training or experience with flight-simulators. Whereas orbital dynamics are more esoteric, and fewer players accurately know what to expect (I was the rare exception to this- since I knew about how orbit actually works, and already had a basic grasp on the Rocket Equation I performed a gravity-turn and made orbit on my first successful launch...) a lot of players come in knowing what to expect when flying in the atmosphere. Thus, the current aerodynamics system is a pain in the a$$, and LESS intuitive to rookies than a realistic system.

If the devs implement realistic rules about how airflow behaves, and the effects of Mach (speed), then ALL the other quirks of aerodynamics (from ballistic coefficients to types of instability) naturally flow from those rules. A sufficiently intelligent player should be able to reason their way through those- especially if the devs improve the tutorials available to new players to mention aerodynamics, maybe even provide some "Advanced Tutorials" that teach players actual principles of aerodynamic design in detail...

Regards,

Northstar

[FishInferno]

I just hope that body lift gets added. That would allow for lifting body reentry vehicles like Dreamchaser

[rkman]

I want body lift because i think it is kind of necessary to be able steer (even vessels without wings) for a safe trajectory if there is reentry heat, and i do want reentry heat.

[sherkaner]

Realistic aerodynamics ARE intuitive. Many KSP players come in with an accurate understanding of physics. Some even with pilot training or experience with flight-simulators. Whereas orbital dynamics are more esoteric, and fewer players accurately know what to expect (I was the rare exception to this- since I knew about how orbit actually works, and already had a basic grasp on the Rocket Equation I performed a gravity-turn and made orbit on my first successful launch...) a lot of players come in knowing what to expect when flying in the atmosphere. Thus, the current aerodynamics system is a pain in the a$$, and LESS intuitive to rookies than a realistic system.

If the devs implement realistic rules about how airflow behaves, and the effects of Mach (speed), then ALL the other quirks of aerodynamics (from ballistic coefficients to types of instability) naturally flow from those rules. A sufficiently intelligent player should be able to reason their way through those- especially if the devs improve the tutorials available to new players to mention aerodynamics, maybe even provide some "Advanced Tutorials" that teach players actual principles of aerodynamic design in detail...

I don't disagree with you, but it's not quite as simple as that. Even if Squad could simply load up the Navier-Stokes equations and have perfect aerodynamics applied, you would have problems because the craft themselves are greatly simplified from real-world craft, of necessity to make the whole modular easy-to-build Kerbal system work. In the real world, very fine details of craft design can have a huge impact on performance and that's just not practical or appropriate for KSP.

So there has to be some kind of simplified model for that reason, as well as simply for processing power and having a more forgiving version of the world for gameplay reasons. The question (I believe) is to figure out how to best model an approximation of reality, even if it is not technically speaking fully realistic. I absolutely agree that people coming in with varying levels of familiarity with physics shouldn't be disappointed by incorrect behavior -- that was very much the point of what I wrote.

[Alias72]

I don't disagree with you, but it's not quite as simple as that. Even if Squad could simply load up the Navier-Stokes equations and have perfect aerodynamics applied, you would have problems because the craft themselves are greatly simplified from real-world craft, of necessity to make the whole modular easy-to-build Kerbal system work. In the real world, very fine details of craft design can have a huge impact on performance and that's just not practical or appropriate for KSP.

So there has to be some kind of simplified model for that reason, as well as simply for processing power and having a more forgiving version of the world for gameplay reasons. The question (I believe) is to figure out how to best model an approximation of reality, even if it is not technically speaking fully realistic. I absolutely agree that people coming in with varying levels of familiarity with physics shouldn't be disappointed by incorrect behavior -- that was very much the point of what I wrote.

there are a number of problems with aerodynamics. allow me to explain.

we are trying to approximate the real phenomenon of lift, drag, and moment on an aircraft.

The Navier Stokes Equation will almost always give you the correct solution. The problem is we can't solve it yet. we can't solve it fully with a real life supercomputer, multi-million dollar research budget, and an actual set of projects that would benefit. we can solve specific examples but it is computationally expensive especially for viscous flow (read 90% of flows).

So what are the equations being used in real life and in FAR? well these are numerical approximations. in Aerodynamics we rarely solve a set of true equations based on physics. Instead we have created equations that are close to a real life solution, with a bunch of assumptions, and then apply corrections. Here is the problem. These equations are designed to approximate reality with a low cost. They are what is deemed absolutely necessary to get an accurate solution; You want to create an approximation OF AN APPROXIMATION and still get ACCURATE RESULTS. Do that successfully and aerospace companies will pay you hundreds of thousands of dollars, universities will give you an honorary PHD, and some obtuse physical quantity will be named after you. Don't believe me on the latter? Reynolds number is a number named after someone who said that "flow becomes turbulent about here sorta?" It is incredibly unlikely you will get anything close to accurate results with less than the aerodynamic solutions. It is like approximating patched conics. will all your orbits be circles?

"Well", you may way, "we don't NEED to accurately simulate aerodynamics right? this is just a GAME! QUIT KILLING MY FUN NERD!" ok stop a moment. I think the point is somewhat lost. What we are saying, with all this math, is that you will not get anything close to real life without those equations.

If you don't simulate mach effects then ANY aircraft flying greater than Mach 0.3 (read very slow) will behave wrong. Not wrong in an ok its just a game way. Sloped wings will be pointless. Drag won't increase rapidly so dagger shaped aircraft are bad. you would want massive aspect ratios (look at a B 52). This is not even close to what we want to represent.

Well why don't we get rid of viscous flow? Ok you're solution is accurate if you are flying either really fast or really slow (all flow is laminar at Mach 0). No turbulence is great but... well you get no lift. without viscosity the air leaves the way it came, merely moving around the wing, and you get zero lift. congratulations you accidentally the physics. See D'Alembert paradox. We actually do assume laminar flow in the real equations but use a correction to give us a usable solution (kutta joukowski)

Why don't we assume air is always incompressible! Less math FTW! well you could. except that a dart shaped projectile traveling at Mach 20 will most CERTAINLY compress the air. Everything above Mach 1 RELIES on compressibility! aerodynamic heating? compressible flow. Supersonic drag? Compressible flow. Supersonic Lift Compressible Flow!

The reality is that Aerodynamics is as simple as it gets. If you want a system where wings magically generate lift based on speed, and parts generate drag if there in front then you can get that. But the best designs will be nothing more then a dart. Picture everyone flying missiles with landing gear. Ok if your focus is on rockets very bad if you like actual planes.

[NWM]

ALL! Give all the effects!

[impyre]

Just my two cents here: I use NEAR. I used FAR, and while I can understand the graphs and physics behind it (at least as well as any other non-aero-engineer), I simply found that it made planes far too difficult and time-consuming to develop (not to mention costly in terms of poor test kerbals and funds). Sure it can be done (and with some care and knowledge it works extraordinarily well, especially compared to stock); however, (and this is just my opinion) I play KSP primarily for space travel... all other concerns are secondary for me. I like NEAR because it handles the bare minimums needed for rockets (and its shortcomings aren't as noticeable over the short and relatively vertical paths you usually follow through the atmosphere) without making planes too difficult. My main concern is ascent profiles and reentry trajectories. I find that NEAR + Deadly Reentry + RealChutes work well together for what I want.

Basically, I have to go through air to get to space. I prefer my time in the atmosphere to be fun AND non-soupy, but I'll take soup over ragequit.

[sherkaner]

there are a number of problems with aerodynamics. allow me to explain.

...

Uh, I'm not sure why this was a reply to me since it didn't seem to have much to do with what I was saying. I'm quite aware that we're not getting anything close to a Navier-Stokes equation solution -- that's why I said "even if". But yes, you are factually correct with everything you say. Although it seems like your conclusion is that we might as well give up because nothing short of an accurate simulation is worthwhile? That seems a bit silly. The problem to be solved is trying to create something as good as possible within constraints of processing power and there are absolutely ways of approaching that problem. Toy models for the different aerodynamic regimes and effects should be able to get you a lot closer to something satisfying than what we have. Will there be edge cases and deviations from real behavior? Sure. But hopefully these can be minimized and at least kept within relatively sane boundaries so that the craft isn't doing anything patently nonsensical.

[Kegereneku]

Time to participate in this thread since 0.90 is out and Aerodynamic ought to be improved

I have to say that although I disagree entirely with his choices on the matter, I like a Feature-driven approach similar to Northstar1989, as an opposite to what I'll call a Realism-driven approach.

To clarify.

# A Realism-Driven approach would be to consider that KSP 'must' only go toward more realism at any point, focusing on making the game "act like reality" rather that "look like reality", thus discarding simplification that make the game fun or playable at all.

For example, considering that plane must have an aerodynamic shape or insisting to add a plethora of "realistic details" that bring nothing to the gameplay.

This is, to me, a very bad if not impossible way to take in video game design.

# A Feature-Driven approach would be to set the gameplay you want to achieve, then shape the rules. Thereby focusing on making the game "fun" rather than "tedious", and including acceptable break from reality.

For example, for all intent and purpose KSP's spaceplane requirement can be split up to these basic feature :

- Anything with enough "wing/speed" shall be able to achieve aerodynamic lift.

- Anything with enough/adapted engines shall be able to accelerate to any speed.

The ease with which a design is able to achieve this is then subject to parameter that allow to balance the game

Note that in the process one must be careful to not suggest "features" that imply dozen of feature no-one else want*.

And also that it is acceptable to let player go away with unrealistic looking plane, we love that stuff. What's important is the design process they had to overgo to make it fly.

I personally think that as far KSP go, we do not need to care about the spaceplane's shape** outside of Center of Mass/Lift/Thrust as the game is already doing. But this is NOT to say that plane shall stay like supersonic balloon with wing as they feel now, The system just have to be improved.

What seem capital to me is the number & function of parts and their relations, as in "As long as the plane have these parts it shall reach orbit". For example the key to "simulate Mach effect" wouldn't be in the shape of the spaceplane, but in the engine's specificity, e.g : A Basic jet gradually ceasing to give thrust over a certain speed/pressure. For good measure there would still be visual effect to mark the limit.

The shape would of course, stay essential for control & landing purpose, but play little in what a design can achieve.

*: On the matter, we just can't cast everything as "difficulty setting", the physic must be the same for everybody. So don't use that excuse to ask devs for your own personal feature, there's mods for that..

** : Stop dreaming, KSP is not making you a real engineer, you wouldn't actually be able to build a plane if it followed real aerodynamic.***

*** : And KSP couldn't simulate complex aerodynamic anyway.

[Red Iron Crown]

I feel that a new aerodynamic model that doesn't take vessel shape into account would hardly be an improvement. Aerodynamics are all about shape. Everyone knows that streamlined objects are better for higher speeds, it is less intuitive to new players to have shape ignored.

The comments about using engine specs to simulate Mach effects demonstrates a fundamental misunderstanding of what Mach effects actually are. (Hint: They exist even with no engine running.)

[Kegereneku]

Oh be assured I KNOW FULLY what the hell Mach effects are. Probably better than most of those who will claim to.

If you read more carefully my post I just consider that creating the illusion of flying a realistic spaceplane is more important than simulating a real aerodynamic model. Thus linking the engine to mach effect (because those DO have to be built specifically to keep working through it) was to give it an existence in term of gameplay and verisimilitude. (With very realistic reason to do so, ask if you want)

Because we mustn't overestimate our own design prowess. It would get extremely frustrating if the game punish you for not being able to do what team of engineers work to achieve for years.

An improvement to me would be to feel like we are flying a real spaceplane, not failing to fly in a more realistic environment, Big nuance here.

[lincourtl]

Thus linking the engine to mach effect (because those DO have to be built specifically to keep working through it) was to give it an existence in term of gameplay and verisimilitude. (With very realistic reason to do so, ask if you want)

KSP is a spacesim where a large, if not majority, of the time spent at transonic and hypersonic speeds is falling through the atmosphere with no engines running, or no engines at all. So linking mach effects to engines doesn't make a whole lot of sense.

[Kegereneku]

Just to clarify, visual effect linked to Mach would only be linked to speed, not engine state.

What I mean is that Engine are what would react differently, for the two following reasons :

# Airflow do act differently at speed below and above the speed of sound. Thus requiring different intake.

Today's ways of dealing with Supersonic flow are :

• Inlet cone that slow down airflow before compression (Turbojet & Ramjet)
  
• Or using the "effect" to compress air before burning fuel at subsonic speed (Ramjet)
  
• Or seeking to compress air and burn fuel at hypersonic speed (Scramjet)
  

Making it a pretty gradual way to shape player's experience (we even have the part, making me suspect that's what SQUAD plan to do).

# Gameplay speaking, I think it would be easier to make a spaceplane spec 90% dependent of its engine "design" (and number of Lift part) as it is a matter of simple combination and discrete quantity rather than its global shape which can have utterly unpredictable result.