What Aerodynamic Features would you like to see added?

[Northstar1989]

DISCLAIMER TO MODERATORS: This thread is NOT an attempt to suggest "realistic aerodynamics", as has been prohibited on the already-suggested list. Rather this is an attempt to solicit player opinions on *which* improved aerodynamics features would be a good fit for the update aerodynamics model SQUAD has hinted might be coming in the future. I don't believe SQUAD has seriously begun work on this (if at all), so it should not be too late to perhaps have them notice this thread, and to let if influence the direction of their development.

There have been whisperings that SQUAD might start working on improving the aerodynamics model soon. Apparently some of the devs have even hinted on it on Twitter of something...

So, the question is, what aerodynamic features would you like to see added to the model? Personally, I'm a big fan of realism and there are quite a few features I would like to see added, but some of you may not agree with me on specific features. Thus, I wanted to break this down into a feature-by-feature discussion of what YOU want to see added that might be enlightening to SQUAD.

Note that this is NOT the place to just shout "Full Realism!", or "I don't want to see ANYTHING added!". Instead, you should discuss *particular features*, and why you want to see them added...

So, to start off this discussion...

Ballistic/Aerodynamic Shielding: This may not be the correct term for this, but basically the idea is that part will not experience lift or drag when they are inside of other parts, such as Cargo Bays (which will be released in 0.25) or fairings (which SHOULD get added whenever SQUAD updates the aerodynamics model, IMHO- besides fairings just LOOK COOL when they break apart to expose a payload...) This will add to both realism and fun, and I can't possibly see a downside to adding this, aside from players being unable to hide wings inside cargo bays to illegitimately improve the lift of their spaceplanes... (most players consider this a "cheat" anyways...)

Cross-Section Based Drag: This is really not as big or scary a change as it sounds like. Basically it means that a rocket two fuel tanks high will only experience *SLIGHTLY* more drag than a rocket one fuel tank high (most drag is based on cross-sectional area in the prograde direction, rather than laminar drag), when flying in the prograde direction. When spinning about randomly, the rocket/plane will experience much more drag- but this will actually make the game EASIER by tending to slow down such spins, and isn't exactly the normal state of such things anyways... (if you're in a death-spin to begin with, why does it matter to you if you're experiencing extra drag?) This feature is also necessary for features enabling nosecones to have functionality. (without it, how will you know that they are between the rocket and the airflow?)

Skin Friction Drag: This is mainly a balancing factor to complement cross-sectional drag. Without it, a rocket two fuel tanks high will experience *EXACTLY* the same atmospheric drag as a rocket one fuel tank high when flying in the prograde direction. This creates a bit of extra drag based on the increased exposed surface area, but does NOTHING to destabilize a rocket. Basically, it helps prevent the game from becoming "too easy" based on changing the basis of drag to ray-casting, while simultaneously increasing realism and making the game more intuitive (most people would EXPECT a taller rocket to still experience more drag, in absolute terms). Drag based on just these first three features will still be MUCH less than in stock.

Shape-Based Drag: This is the BIG one that would really change how rockets are designed in KSP. It would cause nosecones to have function by rewarding more gradual transitions in cross-sectional area, while simultaneously working to destabilize/penalize pancake-rockets. It will make the game a bit "harder", but not unreasonably so- and with just this feature alone you can still get pancake rockets to orbit. This is the reason nosecones are useful in real life! It is also one of the reasons fairings are useful (and fairings just look COOL!)

Mach Effects on Absolute Drag: In isolation, and without combination with some of the "harder" aerodynamic features to follow, this doesn't actually make the game significantly harder. In fact, in many ways, it makes it "easier"- rockets and planes capable of breaking the sound barrier can travel at MUCH faster speeds with the same TWR than in a drag model lacking this feature. What it means is, in a nutsehell, rockets/planes tend to gain very little velocity for additional thrust in the "transonic" regime (from about Mach 0.85 to Mach 1.2), but gain a LOT MORE velocity for their thrust in the "hypersonic" regime of speeds greater than this. It's the *other* mach effects (to follow) that have actually scare many players off of mods like FAR... Should be configurable/able to be disabled on the Difficulty Levels screen at starting a new game, and probably only enabled by default on the higher difficulty levels...

Mach Effects on Stability: This is the mach effect that scares many players away from mods like FAR. What it basically means is that, when crossing the sound barrier (the "transonic" regime between roughly Mach 0.85 and Mach 1.2), planes tend to experience wobble and some loss of control authority, that can potentially lead them to crash. Since these effects are directly related to exposed surface area and shape, they don't tend to affect rockets very strongly. They can also be delayed (so they start closer to, say, Mach 1.00 instead of Mach 0.85) by sweeping the wings back into a "Delta Wing" configuration- and in fact the majority of planes designed in stock already tend to follow this configuration. All in all, with a bit of thought into design and perhaps some extra reaction wheels to dampen the wobble, it really isn't all that hard to manage and deal with these effects. It also would be MUCH easier to minimize or avoid them altogether if SQUAD would consider adding a FAR-like informational display that could optionally be toggled in the SPH/VAB (paraps configurable as to whether even be available as an "Advanced Option" on the Difficulty Level screen when starting a new game...), which would also be a GREAT place to display information like the available Delta-V of a plane/rocket... Should be configurable/able to be disabled on the Difficulty Levels screen at starting a new game, and probably only enabled by default on the higher difficulty levels.

Aerodynamic Failures: This probably shouldn't occur nearly as easily as in FAR or real-life, but explosions and rockets breaking apart are !FUN! in KSP, and there should be a couple ways to cause them, based one aerodynamics. I'll leave the exact details of which types of aerodynamic failure (high G-loads based on lift/drag, localized pressure spots, localized overheating, etc.) should and should not be included up to debate, and in my opinion this is is something SQUAD could best determine through play-testing and user feedback in Experimentals (which team I *STILL* wan to join... sigh). This one should *DEFINITELY* be configurable/able to be disabled through the Difficulty Level screen on starting a new game (and, should only be enabled by default on the higher difficulty levels...)

Body Lift: This feature gets an unfairly bad rep, and honestly I don't know why. What it basically means is that rockets and the fuselages of planes generate a certain amount of lift. With a properly-designed rocket (one that is wider at the bottom than at the top) this tends to actually *STABILIZE* rockets (making flying them easier), and this only makes flying planes EASIER as the lift generated by the plane becomes greater and slightly more intuitive to the shape (you don't have to distinguish between what is and is-not a "wing" part- it all generates lift based on shape...) The ONLY downside of this feature is that it makes it significantly more difficult to lift large, draggy payloads (such as a section of a space station) to orbit without a payload fairing- which should be added to the game in whatever update improves aerodynamics anyways, if only for appearances... (fairings just *look* COOL) It absolutely SHOULD be part of the default difficulty level if made part of the game, and perhaps it is even not worth making it tweakable in Difficulty Levels at all.

Re-Entry Heat: This isn't really a traditional aerodynamic feature, per-se, but it is so closely associated with aerodynamics that I felt it was worth including. Basically, the question here is whether the update that adds aerodynamics should also add re-entry heating. I waver on this question myself- and would be perfectly happy to see SQUAD delay it to an even later update... (so they can do a better job with focusing on just aerodynamics, especially from the coding/stability standpoint)

Note that I have INTENTIONALLY left out features many players of FAR, or real-life aerospace engineers, might be familiar with such as aerodynamic instability leading to things like "Dutch Roll" and stalls. This is because I think such features are a little too advanced for most players (even experienced ones like myself tend to struggle with them occasionally), and should intentionally be left out of any initial aerodynamics update, perhaps to infinity (I would much rather see SQUAD focus on things like adding a difficulty option to upscale Kerbin 2-3x, or finally adding a basic ISRU system- either of which players could choose *NOT* to play with if they didn't like them, but ought to be part of scope-completion IMHO...)

Regards,

Northstar

EDIT: I re-named "Ray-Cast Drag" to "Cross-Section Based Drag", and "Laminar Drag" to "Skin Friction Drag" to more accurately capture the concepts I was trying to describe with the terminology. Also, thank you for enlightening me that they are really caused by the same thing, but I'm leaving "Mach Effects" and "Re-Entry Heating" as separate features, as it IS possible to simulate one but not the other... Some of you aerospace types are so picky about your terminology.

Edited October 8, 2014 by Northstar1989

[Claw]

DISCLAIMER TO MODERATORS:

This has been a hot topic lately, so in light of that...If people can keep the discussion on topic and relevant without degrading into demands, pure negativity, and/or personal attacks then it's fine. There's no need to discuss what angers you about what's been done or failed to be done if the topic of the thread is to discuss the virtues and characteristics of specific aerodynamic features.

Personally with the introduction of difficulty settings, I've been moving toward the camp of having a checkbox for "simplified aerodynamics" vs. something more "fully featured." Casual gamers can work off a simplified construct that doesn't require in-depth knowledge of things like mach effects and maybe even aerodynamic shielding. While those who have mastered more of the game can up the challenge level a bit and give themselves more nuanced things to deal with.

That sort of thing seems more in-line with the kerbal philosophy of slapping things together, and the general open-endedness feel of the game while still providing some progression.

Cheers,

~Claw

[Twreed87]

Honestly, the only thing I care about regarding aerodynamics is the need for payload fairings. Rockets look really weird without fairings when they're carrying odd-shaped satellites with tons of bits hanging off. The same goes for radial fuel tanks without nosecones. But fairings and nosecones are pointless in the current model. I don't want to have to sacrifice efficiency of my rocket for a realistic appearance.

[Sean Mirrsen]

I've said earlier, in a different thread, that more or less the best way for KSP aerodynamics to change, would be to get the updated drag function (not dependent on mass), and to implement some basic airflow shielding so that parts not exposed to the airflow don't contribute to drag (and so air intakes not in the airflow don't function). It keeps the aerodynamics basic enough that beginner players won't need to keep an aircraft building handbook open every time they're in the SPH, and at the same time is a vast improvement over stock aerodynamics and allows for things like fairings and cargo bays.

[ferram4]

This is gonna be a long post, responding to most things.

Ballistic/Aerodynamic Shielding:

Yes, and this is not too difficult to implement. The main trick is making sure that the cargo bay and fairing parts are shaped sanely so that they can easily detect the parts inside them.

Ray-Cast Drag:

This is a terrible method of calculating lift and drag. It is relatively "accurate" for the limiting effect of Mach number -> infinity, but it is very expensive performance-wise and it is incredibly difficult to make a stable vehicle for this type of flight model. Any attempts to reduce the expense only make it more twitchy as you have a lower number of raycasts and a smaller resolution to work with.

Laminar Drag:

This is not "laminar" in any way; you are talking about skin friction drag instead, which (for most of the situations we're looking at) will likely be turbulent rather than laminar.

Shape-Based Drag:

Using what model? After all, your earlier "Ray-Cast Drag" would also fall under this category, so either the first one was redundant, or you're thinking of some other model.

Mach Effects on Absolute Drag:

For rockets, these effects are almost unnoticeable. For planes, if KSP keeps its overpowered jet engines, this will also be rather unnoticeable. This is, however, very important for aerobraking and reentry; leaving this feature out inevitably makes any addition of reentry heat much nastier, because the lower drag from the lack of considering supersonic effects will make reentry heat nastier. More thoughts on that later.

Mach Effects on Stability:

I will note here that FAR assumes that most planes designed in KSP have very thin wings and have well-designed control surfaces to lessen the loss of control authority through the transonic regime; it is almost unnoticeable to be honest. Further, as a result of that, the brief instability near Mach 1 is only really noticeable for planes with Cessna-like straight wings; swept wings don't even suffer this effect. I will also note that for moderate supersonic flight (Mach 1.5 - 4, or ~450 m/s - 1200 m/s) the aerodynamic center (what you would call center of lift, but accounting for the effects of drag as it should) shifts backwards slightly, making the plane more stable.

Aerodynamic Failures:

This is probably something that can be ignored completely. FAR is getting its failure criteria made a little more lenient to be closer to the reality of military aircraft (because going with transports is too low for most people), and since the stock version will inevitably be no more difficult, even on the hardest setting, leaving this out completely is probably wisest.

Body Lift:

Body lift can't be said to stabilize or destabilize in general, because it depends on the configuration, just like how wings cannot be said to stabilize or destabilize. Great big lightweight spire on the front of your rocket for no reason? Body lift will destabilize it. Spherical section on the bottom of a command pod? Body lift will stabilize it.

Re-Entry Heat:

They should be related; the only reason FAR does not have aerodynamic heating is due to a quirk of history that meant that the original reentry heat plugin (before Deadly Reentry appeared) appeared before I could take on the project. I should note that reentry heat is a Mach effect, and so should be disabled with Mach effects being turned off as well. Similarly, the aerodynamic graphic effects should also be disabled if those things are not implemented.

Note that I have INTENTIONALLY left out features many players of FAR, or real-life aerospace engineers, might be familiar with such as aerodynamic instability leading to things like "Dutch Roll" and stalls. This is because I think such features are a little too advanced for most players (even experienced ones like myself tend to struggle with them occasionally), and should intentionally be left out of any initial aerodynamics update, perhaps to infinity (I would much rather see SQUAD focus on things like adding a difficulty option to upscale Kerbin 2-3x, or finally adding a basic ISRU system- either of which players could choose *NOT* to play with if they didn't like them, but ought to be part of scope-completion IMHO...)

This shows a blatant misunderstanding of how KSP is set up and how the parts interact. Did you know that dutch roll already exists in stock KSP? Do you know why? It arises naturally from the physics simulation when you simulate the aerodynamics of a plane and distribute the forces across the vehicle and account for small relative velocities. There's no way to leave those motions out at all, and it's not like they're hard-coded into FAR or NEAR. Talking as if they are just makes things more confusing for people that don't understand how KSP works or how aerodynamics work.

Please, if you're going to advocate for Squad to change aerodynamics, know what you're talking about. Encouraging them to look at raycasting for drag or trying to leave out dutch roll is just going to lead them down a wild goose chase and possibly make them think twice about changing aero.

[Northstar1989]

First of all, Ferram, I just wanted to say I highly respect your mod, and I'm glad my thread attracted your attention. Have you considered approaching SQUAD directly to see if you might be able to work with them to improve their aerodynamics model, since FAR has already been so successful?

This is a terrible method of calculating lift and drag. It is relatively "accurate" for the limiting effect of Mach number -> infinity, but it is very expensive performance-wise and it is incredibly difficult to make a stable vehicle for this type of flight model. Any attempts to reduce the expense only make it more twitchy as you have a lower number of raycasts and a smaller resolution to work with.

It was just a first idea of a simple and easy-to-understand (for most players and SQUAD) way to calculate life and drag. Thanks for pointing out its drawbacks. How would you suggest SQUAD implements it if they improve the aerodynamics model?

This is not "laminar" in any way; you are talking about skin friction drag instead, which (for most of the situations we're looking at) will likely be turbulent rather than laminar.

I'm talking about friction along a smooth solid surface (like the walls of a rocket), where there are no irregularities to disturb it or cause mixing. I don't have a perfect understanding of fluid dynamics (after all, I'm a biologist, not an aerospace engineer), but isn't laminar flow the type of floe that will be experienced in this situation?

http://en.wikipedia.org/wiki/Laminar_flow

More broadly, though, I was using my terms a little loosely. I was referring to drag along the exposed surfaces of a rocket running parallel to the direction of movement, rather than based on the cross-sectional area exposed to drag. Skin friction drag might be the correct term for it...

Using what model? After all, your earlier "Ray-Cast Drag" would also fall under this category, so either the first one was redundant, or you're thinking of some other model.

I was sub-dividing the aerodynamics model into smaller sub-features. After all, it's possible to have an aerodynamics model which penalizes exposed surface area and cross-sectional area, but doesn't create any penalties for abrupt transitions in diameter... I'm aware that most *complete* aerodynamics models account for both considerations... (shape and exposed area)

For rockets, these effects are almost unnoticeable.

I think we're both in agreement about that.

For planes, if KSP keeps its overpowered jet engines, this will also be rather unnoticeable.

[Red Iron Crown]

A shape-based drag model, plus atmospheric frictional heating.

I don't have a great understanding of aerodynamic principles, but for me the game would be greatly improved if a new aero model makes streamlining important and reentry somewhat dangerous. Anything beyond that (or required to make those things happen) is just gravy.

[John FX]

For me a new aero model should pretty much add FAR and DRE (realheat). I`d like for nosecones and fairings and heatshields to be needed.

[RocketPilot573]

I would like to simply see something similar to NEAR and DRE, and Ferram did have some good points.

[ferram4]

First of all, Ferram, I just wanted to say I highly respect your mod, and I'm glad my thread attracted your attention. Have you considered approaching SQUAD directly to see if you might be able to work with them to improve their aerodynamics model, since FAR has already been so successful?

It's in my mind to poke them and remind them that I'm here if they want my help.

It was just a first idea of a simple and easy-to-understand (for most players and SQUAD) way to calculate life and drag. Thanks for pointing out its drawbacks. How would you suggest SQUAD implements it if they improve the aerodynamics model?

The first pass shouldn't implement it at all. The only reason to implement it is if you account for Mach number effects, and you need to be really careful; imagine a plane where the horizontal tail is directly behind the wing in the airstream. In real life, at low Mach numbers (M < ~2 or so) this would have only minor effects, mostly from the wake from the wing, and it would fly fine. At the higher Mach numbers this simulates, the tail wouldn't do anything until it popped out from behind the wing, and so the plane would be unstable for that small angle of attack region. While it would be nice to simulate that, those Mach numbers don't occur commonly in KSP at the densities where such a thing would be noticeable, and so it can be left out, I think. In any case, implementing it would be best done by using the light cast for the aerodynamic effects even at the lowest graphics settings.

I'm talking about friction along a smooth solid surface (like the walls of a rocket), where there are no irregularities to disturb it or cause mixing. I don't have a perfect understanding of fluid dynamics (after all, I'm a biologist, not an aerospace engineer), but isn't laminar flow the type of floe that will be experienced in this situation?

http://en.wikipedia.org/wiki/Laminar_flow

More broadly, though, I was using my terms a little loosely. I was referring to drag along the exposed surfaces of a rocket running parallel to the direction of movement, rather than based on the cross-sectional area exposed to drag. Skin friction drag might be the correct term for it...

While yes, laminar flow would exist if the surface were absolutely perfectly smooth (and there were no adverse pressure gradient to destabilize the laminar boundary layer), it's likely that turbulent flow would be tripped by the nose of the rocket or aircraft, and once the flow is turbulent there's not much you can do to make it laminar again downstream. While it's likely that at the lower densities well out of the worst of the atmosphere that laminar flow will be dominant, the dynamic pressure will also be low enough that ignoring that is a decent choice as well.

I was sub-dividing the aerodynamics model into smaller sub-features. After all, it's possible to have an aerodynamics model which penalizes exposed surface area and cross-sectional area, but doesn't create any penalties for abrupt transitions in diameter... I'm aware that most *complete* aerodynamics models account for both considerations... (shape and exposed area)

Any type of system that accounts for exposed cross-sectional area is going to penalize abrupt changes in area, unless it's not based on reality at all. The problem with designing models not based in reality is that there isn't muhc way to figure out how to balance them to get the behavior you want.

I would disagree with that. I've had significant difficult breaking the sound barrier with my single-jet engine aircraft (due to their velocity curves, they produce relatively little thrust by Mach 1.0 or so...), for instance. But yes, the effects aren't enormous with the currently OP'd jet engines (which probably aren't going to get nerfed or made more realistic), which is why I believe it shouldn't be too great of a difficulty for most players to adjsut to...

After FAR nerfed the jets because they were too powerful? The standard jets blow through Mach 1 like there's no problem at all.

That's a good point I didn't think of when writing this, but it certainly would be important... Why exactly is this, though? I'm not 100% I, or other players, get the reasons for this... Definitely a reason to include this before re-entry heat, though, to make the game easier for new players...

Higher Mach number == greater density changes in the flow == higher pressure differences on front and rear surfaces == greater drag. For the most part. Not including Mach effects, incidentally, is equivalent to declaring air to be incompressible, which (as a side note), is fairly accurate for flight below Mach 0.3 (< 113 m/s at SL).

I've nearly lost control of several aircraft in the transonic regime (they had very large wing areas, and thus were strongly affected by trans-sonic instability), so I wouldn't exactly call this a minor effect. It significantly limits aircraft design choices, by pushing them strongly towards craft with strongly swept wings and low overall wing area...

High angle of attack through the transonic regime, I'm guessing? Because that will certainly make controlling it through there difficult.

The craft I nearly lost control of had very strongly swept wings, but HUGE wing area. I think you're probably too used to playing with craft with smaller wing areas, and have forgotten how devestating these affects can be for craft with extremely large wings... (yes, I was aware that was a danger I was facing when designing those planes- which is why I strongly swept back their wings: as a result they COULD make it through the sound barrier if they had to, as when in a dive, even though they were designed mainly for subsonic conditions...)

Not really. Building oversized rockets and planes is my thing, and transonic instability is nowhere near as strong as it used to be or even should be for some of the designs people create. Ir probably won't be added to a stock model though, because it's "not fun" or something like that.

As for the shift of the Center of Lift towards the tail, that's an effect that occurs in FAR, but might not necessarily occur in a revised stock aerodynamics model. That's another important feature for the devs to consider including, however...

Not including that tends to make supersonic flight much more difficult, because drag becomes stronger relative to lift, which tends to make planes less stable if the aerodynamic center doesn't shift as it should.

I agree with your move to rate up structural margins to military levels in FAR, of course, as any spaceplane would probably be built to those kinds of margins as well... (there is a direct correlation between structural margins and reusability) But I'm surprised to see you don't think a revised stock aerodynamics model should include them, considering you're the creator of the mod that's most famous for them...

Nah, this is horribly unrealistic. SSTOs wouldn't be able to get to orbit if they were designed to take 9 gs without failure; they would have too much dry mass to have the dV necessary. I also note that the idea of failures is "Kerbal" and "awesome" for the community when it happens to other people, but is "broken" and "hard" and "unfair" when it happens to them. So this is just me accepting that there's no way that would ever be accepted as a feature, because too many people would complain.

Fair enough. I agree that generally common-sense rocket designs won't suffer too strongly from body lift (except when carrying unshielded payloads), and think that it helps players more than it hurts them, if anything. I would very much like to see it included in a revised aerodynamics model myself. So I take that as a "YEA" in favor of SQUAD putting that into a revised aerodynamics model if it were up to you???

My opinion is simply that if Squad includes something better than FAR, I will be happy. If they include something less than NEAR, I will be sorely, sorely disappointed. Something in-between but that I can easily modify, that's workable.

Uhhh, OK... That's nice to know... So FAR might have included a re-entry feature if DRE's predecessor hadn't already been around first. I think most players *might* appreciate the separating out of these two features so they can be independently enabled/disabled in the Difficulty Levels settings, however...

I, personally, think that the physics should be set up to be consistent, or else it will be confusing to players reading up on things. Which means that if Mach effects aren't there, reentry heat can't be there, because aerodynamic heating is (primarily) a result of compression effects, which only happen if M > 0.

This might result from a mixing up of terminology. It might not have bee *Dutch* roll I was thinking of, but the other type of instability roll along that axis. What was it called again?

The only other one is wing rock, which is an inevitable result of accounting for wing sweep. That motion is also not hard-coded into FAR, it just arises from the simulation.

The key idea I was trying to get across is that those principles that can be avoided, should. If there are relatively simple tweaks to the aerodynamics simulation that might prevent the introduction of new forms of instability while still introducing some of the features I mentioned earlier, then they probably should be implemented.

NNNNNNNNOOOOOOOOOOOOOOOOOOOOOOOOO that is a terrible idea! The only "simple tweaks" that you can add are tantamount to magical torques applied to the vehicle to keep it pointed int he correct direction; the problem is that those same torques will result in lots of completely confusing and unrealistic effects for players that encounter them. Worse, it will encourage infiniglider-like exploitation of those forces, replacing an already bad model with something not much better. Finally, such kinds of "fixes" will inevitably hinder attempts to build crazy designs, not because they're impossible according to the real world, but because the tweaks declared "I think you're doing something wrong, LET ME HELP YOU."

The very worst part of this is that it will throw physics consistency (which is a very important thing for a newbie, who is learning the game, to have) out the window.

Though on second thought, maybe it's best not to baby the players *too much*. If a type of instability naturally arises from an aerodynamics model for the same reasons as in real life, then maybe it *should* be left in the game for players to "discover". I really don't know. What I *DO* know, is I want to see a MUCH improved aerodynamics model in stock, with as much realism as possible, without including some of the scarier features like aerodynamics instability that I think might overly frustrate some of the new players.

The open secret is that once you disconnect aerodynamic forces from mass you can have aerodynamic instability. The same forces, however, can cause aerodynamic stability. I don't think new players will be frustrated; they'll learn aerodynamics fine, just as they learn orbital mechanics now. The people who will really complain will be the experienced players who dislike becoming newbies again, because everything they have learned up until now is gone.

Maybe there's an easier solution to this involving the Difficulty Levels feature, or maybe I'm under-estimating the new players (and they'll be perfectly capable of dealing with realistic instability issues), I just don't know. Keep in mind that when I first started playing KSP, I made it to orbit with maybe my 2nd or 3rd launch (once I got a good feel for the rate at which the engines burned fuel) after reading through maybe one or two very basic tutorials on the Wiki, and the Mun *very* soon after that (with only a couple failed attempts- this time while I got a feel for how early/late I needed to start my final landing burn). So I didn't go through the frustration of struggling just to get a rocket into orbit many new players do, and as such am not really sure how much a new player might be willing to take before they give up... I might tend to under-estimate new players' determination as a result.

My first orbit actually was inside the atmosphere as a result of how tall it is compared to the planet; my sense of scale was wrong. Then again, I also played before we had landing legs, and so my first Mun landing was trying to put a rocket on there Buck Rogers-style. Let me tell you, that was frustrating.

Considering that proper aerodynamics will inevitably result in a somewhat more rewarding rocket launch (you spend a lot more time getting speed and a lot less time waiting to get high enough) I think it will be more enjoyable for players, even with the failures.

I'm *hoping* that SQUAD will spend more time listening to guys like you who know how to actually implement revised aero, and less time listening to guys like me who just have some experience playing with it.

I'd be highly appreciative if you actually posted a suggestion/discussion thread yourself about what features should/should-not be included in a revised aero model. I only created this thread myself because I saw a need, and nobody seemed to have yet done it. Your name/reputation would probably also draw in more players to the discussion that anything I could start...

Ummm... what I'd like in one is kind of obvious. I mean, I'm only developing and supporting one, maybe I'd like what's in that? I realize that a simpler version is probably what we'll get though, so something like NEAR would be a close 3rd (with 2nd being something in between FAR and NEAR).

[NathanKell]

It's worth noting that if you implement reentry heat *without* mach effects (although as ferram points out they are the same thing) you will have much *harder* reentries than you should.

I mean, reentry is *already* harder than it should be, because parts generally only have 40% of the surface area they should for their mass (consider the Mk1 pod: masses what Mercury does, but 64% of the width, height, and depth, so .64*.64 the surface area). If you then don't consider increased drag from Mach effects, you're going to get a "double whammie" worth of increased heating.

[Basilo1146]

I personally think just the first 4 + body lift should be added by default, as those would actually make the aerodynamics model somewhat realistic. All the other features should be added as well, but they should be optional and be selected in the difficulty tab.

[micha]

I know next to nothing about aerodynamics, so purely from a dummy's gameplay perspective I would like to see an aerodynamics model which rewards players for building streamlined rockets. That is, if I put nosecones, fairings, and cargo-bays on a rocket, I expect the rocket to require less dV than if I just plonk an unshielded mess on the top or build a pancake.

Everything else (aerodynamic stresses, mach effects, etc) I don't care about; that is, I'll learn to fly with or without them, but I don't know enough about aerodynamics to appreciate them (yet). Re-entry heating (and burnup) should be an optional difficulty item imho.

Either way, the simulation shouldn't add significantly to the games' CPU requirements (yes, I want my cake and eat it!). I currently play with NEAR and it's ok, and while I haven't re-tested it recently, I think when I did try it, FAR was a bit heavy for my computer. From a gameplay perspective, both do what I wanted and because I installed them very shortly after starting to play KSP I don't consider them to add any difficulty to the game - but I only fly rockets.. it might be different for planes.

[Sky_walker]

Ballistic/Aerodynamic Shielding - that will be partially implemented in .26

Ray-Cast Drag: - it was already proven that ray-casting doesn't work well in KSP. At least: not in any simple implementation that Unity can handle well in real time.

This feature is also necessary for features enabling nosecones to have functionality. - no, it's not. See: FAR

Laminar Drag: - no comment

Shape-Based Drag: - pretty much must-have at some point in KSP

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

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.

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

Body Lift: - that will be partially implemented in .25 for SP+ parts

Re-Entry Heat: - pretty much must-have at some point in KSP

It's worth noting that if you implement reentry heat *without* mach effects (although as ferram points out they are the same thing) you will have much *harder* reentries than you should.

I mean, reentry is *already* harder than it should be, because parts generally only have 40% of the surface area they should for their mass (consider the Mk1 pod: masses what Mercury does, but 64% of the width, height, and depth, so .64*.64 the surface area). If you then don't consider increased drag from Mach effects, you're going to get a "double whammie" worth of increased heating.

^What he said.

I'm *hoping* that SQUAD will spend more time listening to guys like you who know how to actually implement revised aero, and less time listening to guys like me

Agreed.

[arkie87]

I should note that reentry heat is a Mach effect, and so should be disabled with Mach effects being turned off as well. Similarly, the aerodynamic graphic effects should also be disabled if those things are not implemented.

I have to humbly disagree with ferram on this one. Just because they are both technically "Mach effects" doesnt mean they should both be disabled as a group-- that is like arguing that since N-body physics isnt used, the patched conics method shouldnt be either since they are both caused by gravity (ok, a bit of an exaggeration but you get my point). I think its obvious that this game has to pick and choose which physics to simulate and how accurately to simulate it, and there is no inconsistency in choosing to model Mach effects for heating from not for aero. The fact that they are technically caused by the same thing I think is irrelevant (though it would make grouping things into checkboxes in the difficulty window easier).

Personally, I would prefer the game to have NEAR like aero (though FAR aero with aerodynamic disassembly disabled is good too). But just because NEAR doesnt include mach effects for aerodynamics, doesnt mean it shouldnt include them for heating (though it could/should still obtain Mach number from NEAR/FAR).

Edited October 6, 2014 by arkie87

[arkie87]

While yes, laminar flow would exist if the surface were absolutely perfectly smooth (and there were no adverse pressure gradient to destabilize the laminar boundary layer), it's likely that turbulent flow would be tripped by the nose of the rocket or aircraft, and once the flow is turbulent there's not much you can do to make it laminar again downstream. While it's likely that at the lower densities well out of the worst of the atmosphere that laminar flow will be dominant, the dynamic pressure will also be low enough that ignoring that is a decent choice as well.

While i will admit that i study mostly in-compressible, laminar, internal flow (and also, this argument is mostly irrelevant), it seems to me that you will always get turbulent boundary layer far enough downstream since

Re = V/nu*x

Flow is laminar as long as Re_crit < 5e5

Once x gets long enough or V high enough, bam! Turbulence!

For a 1 m/s air stream at stp, x_crit is 7.68 m... so i think its pretty clear that it will be turbulent...

unless your rocket is less than 8m long and you plan on taking off to space at 1 m/s...

[ferram4]

Ballistic/Aerodynamic Shielding - that will be partially implemented in .26

I haven't seen this said anywhere. Source?

Body Lift: - that will be partially implemented in .25 for SP+ parts

If it's done in any way similar to the way the SP+ parts are already set up, body lift will be too powerful and the drag of the SP+ parts will be much, much, much lower than that of all of the other parts. I'll admit, I'm not really looking forward to this one being stock given how hacky the implementation is (it's just using the winglet code).

I have to humbly disagree with ferram on this one. Just because they are both technically "Mach effects" doesnt mean they should both be disabled as a group-- that is like arguing that since N-body physics isnt used, the patched conics method shouldnt be either since they are both caused by gravity (ok, a bit of an exaggeration but you get my point).

I cite NathanKell's post above you:

It's worth noting that if you implement reentry heat *without* mach effects (although as ferram points out they are the same thing) you will have much *harder* reentries than you should.

I mean, reentry is *already* harder than it should be, because parts generally only have 40% of the surface area they should for their mass (consider the Mk1 pod: masses what Mercury does, but 64% of the width, height, and depth, so .64*.64 the surface area). If you then don't consider increased drag from Mach effects, you're going to get a "double whammie" worth of increased heating.

It's fully worth disabling them as a group considering doing otherwise will make KSP reentries harder than in reality. That is not desirable in any way. I also think that the alternative (that you will have reentry heat without actually compressing the flow) is terrible simply because it reenforces the "reentry heat is from viscosity" idea. After all, what else could be causing the heating if you've removed compressive effects?

While i will admit that i study mostly in-compressible, laminar, internal flow (and also, this argument is mostly irrelevant), it seems to me that you will always get turbulent boundary layer far enough downstream since

Re = V/nu*x

Flow is laminar as long as Re_crit < 5e5

Once x gets long enough or V high enough, bam! Turbulence!

For a 1 m/s air stream at stp, x_crit is 7.68 m... so i think its pretty clear that it will be turbulent...

unless your rocket is less than 8m long and you plan on taking off to space at 1 m/s...

You're not accounting for air density changes. The airflow around most reentry vehicles is laminar due to how low the air density is, which reduces the Reynolds number quite significantly. You'll also find that Re_crit is based on empirical measurements of real-life surfaces, which are, of course, not perfectly smooth.

[Whirligig Girl]

I don't think implementation of Re-entry heating and improved Aerodynamics are going to be perfect models of reality. I think they're going to be custom-built for game balance and authenticity. After all, Ferram and the other Realism-Junkies would still be able to use FAR, presumably, and I don't think exact implementation of aerodynamics is what is planned, I think they will probably custom build a system that doesn't follow real rules behind the scenes, but still feels right to the player.

I'm sure Ferram will have something to say about this.

[arkie87]

I haven't seen this said anywhere. Source?

If it's done in any way similar to the way the SP+ parts are already set up, body lift will be too powerful and the drag of the SP+ parts will be much, much, much lower than that of all of the other parts. I'll admit, I'm not really looking forward to this one being stock given how hacky the implementation is (it's just using the winglet code).

I cite NathanKell's post above you:

It's fully worth disabling them as a group considering doing otherwise will make KSP reentries harder than in reality. That is not desirable in any way. I also think that the alternative (that you will have reentry heat without actually compressing the flow) is terrible simply because it reenforces the "reentry heat is from viscosity" idea. After all, what else could be causing the heating if you've removed compressive effects?

You're not accounting for air density changes. The airflow around most reentry vehicles is laminar due to how low the air density is, which reduces the Reynolds number quite significantly. You'll also find that Re_crit is based on empirical measurements of real-life surfaces, which are, of course, not perfectly smooth.

Re-entry heat can be scaled back/calibrated to account for reduced drag. This will have to be done anytime game physics are changed...

I thought skin friction is a significant component to heating (in addition to compressibility effects)?

True, i calculated x_crit at 1 kPa, and its 700 m at 1m/s velocity. But even so, velocity is about 1000 x faster at re-entry, so x_crit is still small enough...

[ferram4]

I don't think implementation of Re-entry heating and improved Aerodynamics are going to be perfect models of reality. I think they're going to be custom-built for game balance and authenticity. After all, Ferram and the other Realism-Junkies would still be able to use FAR, presumably, and I don't think exact implementation of aerodynamics is what is planned, I think they will probably custom build a system that doesn't follow real rules behind the scenes, but still feels right to the player.

I'm sure Ferram will have something to say about this.

Sounds like NEAR to me. Actually, to be honest, it sounds more like FAR. There's a lot of stuff I've left out of it because it's difficult to implement or limits designs too much (looking at you, area ruling requirements).

The thing is though, going with vague "feelings" is terrible. The average player, just starting out with KSP, "feels" that orbital mechanics is wrong. Does that mean that it should be removed? Of course not. Then again, for the average experienced player, orbital mechanics "feels" correct. Feelings are too relative to make judgments like that on and will be dominated by attempts to stick to the status quo, however bad it is, simply because it is familiar.

Re-entry heat can be scaled back/calibrated to account for reduced drag. This will have to be done anytime game physics are changed...

The problem is unintended consequences of that.

I thought skin friction is a significant component to heating (in addition to compressibility effects)?

It can contribute to the heat transfer, but the heat itself is almost all from compressibility effects. Especially since the heat drops viscosity enough that the heat produced due to viscous losses is almost nothing.

True, i calculated x_crit at 1 kPa, and its 700 m at 1m/s velocity. But even so, velocity is about 1000 x faster at re-entry, so x_crit is still small enough...

Velocity can be 1000x faster, but density can be much, much smaller and the forces will still be identical. You have to consider that most literature on reentry vehicles considers their performance in free-molecular and laminar flow, with almost no consideration given to turbulent flow.

[Red Iron Crown]

I haven't seen this said anywhere. Source?

Not to speak for Sky_Walker, but it's probably a reference to this tweet:

"Shoutout to HarvesteR who successfully completed a module that eliminates all drag for things within a cargo bay in the stock game." - Maxmaps

Not sure it'll be 0.26 when it appears, but it looks like there's at least a working prototype for the code to make cargo bays and fairings relevant in stock. No details on how it works, my bet would be that it determines which parts are enclosed and temporarily sets their Cds to 0.

[Laie]

Ballistic/Aerodynamic Shielding - that will be partially implemented in .26

I haven't seen this said anywhere. Source?

There was a tweet from maxmaps that cargo bays will shield their contents.

There was a little more that can, possibly, be understood to mean that they're actually considering a proper solution, rather than a cargo-bay-only kludge. This has only been hinted at and is by no means confirmed, though.

[Sky_walker]

I haven't seen this said anywhere. Source?

Twitter. Harvester did something to exclude cargo in the cargo bays from aerodynamic drag.

[edit:] Red Iron Crown got it.

Another twitt he posted: "Hollow part with a door into which you put other parts to release at your disclosure."

And then: "Don't expect the module in 0.25 though. The more we dig the more we find regarding aerodynamics. Promising."

Miguel will release post with plans for 0.26 after 0.25 release, so we'll know more details when this one arrives.

If it's done in any way similar to the way the SP+ parts are already set up, body lift will be too powerful and the drag of the SP+ parts will be much, much, much lower than that of all of the other parts. I'll admit, I'm not really looking forward to this one being stock given how hacky the implementation is (it's just using the winglet code).

No idea how they plan to resolve it, but AFAIK there will be come changes to the parts configuration from SP+.

Edited October 7, 2014 by Sky_walker

[ferram4]

Twitter. Harvester did something to exclude cargo in the cargo bays from aerodynamic drag.

[edit:] Red Iron Crown got it.

Another twitt he posted: "Hollow part with a door into which you put other parts to release at your disclosure."

And then: "Don't expect the module in 0.25 though. The more we dig the more we find regarding aerodynamics. Promising."

Miguel will release post with plans for 0.26 after 0.25 release, so we'll know more details when this one arrives.

I know all about the module already, I follow this stuff pretty closely. So, no source for it being in 0.26. So in that case, don't go around screaming that it will be in 0.26, you'll get lots of people hyped about stuff that isn't happening yet.

No idea how they plan to resolve it, but AFAIK there will be come changes to the parts configuration from SP+.

And yet the lifting body effect will still probably be too strong. I suspect they won't resolve that issue, because they won't recognize it as one.

[NathanKell]

Going back slightly, I realize I didn't answer this. I am less against there being mach effects without reentry heating than I am there being reentry heating without mach effects (also, poor choice of words: reentry heating *is* a mach effect, so said better, I am less concerned at one of the mach effects being left unimplemented, than *only* one being implemented). While I think in general they should be tied together, I can see why one might want Custom difficulty to allow toggling heating off (or, maybe better, simply turning off "part explodes at max temperature" rather than turning off heating and cooling).

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.