Airflow Code: How does it work?

[Saxman]

Could someone provide an explanation on how the game's current airflow model works? Preferably in a manner accessible to someone not familiar with programming or aeronautical engineering?

I ask specifically because of the recent complaints about the F4U's turning performance. There's a LOT of accusations that it's the airflow coding itself (either directly or by inferrence from all the "After the airflow code was redone...") that favors the Corsair flight model and flaps. However, I believe it was posted in one discussion by Pyro or HTC that aircraft are NOT floating in a vacuum and that there IS an atmosphere at work on the airframe. So my main question is: if the airflow code is set up to mimic the actual fluid effects of air moving over the airframe (essentially, a virtual wind-tunnel) how could airflow ALONE be responsible for any overmodelling in the flight model (or under-modelling of other aircraft)? Shouldn't this then be more a matter of undermodelled torque or some other factor?

[Krusty]

Only HTC can tell you, and I doubt they ever will tell you.

As for my comment about the airflow code: The f4us were not remodeled when they got their super performance boost. The update that did it was when HTC released the patch that was described tersely as "we redid the airflow code" [paraphrase]. The planes' FM code were apparently untouched.

Keep in mind this airflow recode fixed many bugs and problems with the 109s and 190s and many other planes. P-47s started flying like the true 15,000lb plane they were, p-51s a little more so.

There are several points of contention. I believe WideWing has posted about the issues with P-51 flaps, and then there's the F4u stall/stability/flaps/lift issue, and there are several other issues as well that folks have brought up.

All these showed up after the airflow recode. ALL planes handled different to one extent or another, after this update. The recode was probably the most important update (IMO) since AH1 changed to AH2, and I would not go back to the old system ever, but it seems that a few planes have certain side effects that need tweaking under the new system, that did not show up under the old one.

[Ghosth]

I think the key is in the flaps. Flight models for all planes were not redone at that point to the best of my knowledge. However FLAP effect WAS redone, and now we have F4u's that can hang with a spit, and a Pony that feels more like a 109.

So if the flight model for an aircraft gained XXX more lift, stability, at X degrees of flaps that can have a large overall impact on the flight model.

[trotter]

Originally posted by Ghosth
and a Pony that feels more like a 109.

You mean 190, correct? 109's tend to maul 51's in flaps deployed fights from my experience.

Also, it's difficult to conclude with certainty that airflow code tweaking alone caused the changes in FM. There have been times when HT has made updates in patches that were not specifically mentioned in the changelog. It's not completely out of the realm of possibility that FM's with regards to flap performance were tweaked separate of airflow recoding.

Of course, I agree that the airflow coding alone is the most likely suspect. Just pointing out that there's a chance everyone could be barking up the wrong tree. Changelog notes don't always release everything. Would be nice to get to get some clarifying remarks from HT in this thread, and the one about F4U flaps specifically, so we have a little more information to go on.

[Kweassa]

In the end, nobody knows how it works.

 There could be a "universal FM" regarding general aerodynamical forces, coupled with individual "tweaks" for each planes.... or on the contrary, everything could be "chart-coded". There could even be virtual fluid-dynamics or whatever necessary to create "true virtual air"...

 It's something HTC has never officially disclosed IIRC, so its anyone's guess on why some planes are behaving in a particular way.

[Krusty]

It's not just flaps.

190s couldn't pull any Gs above 200mph without finding themselves inverted a micro second later. I could not fly 109s effectively. The bugs in the 109s from AH1 were lessened to the point that my computer could fly them, but I would be flying level, bank 40 degrees for a non-combat turn, and find myself in an accelerated stall where first one wing totally stalls and dips, then the other, and I would rock back and forth stalling both wings alternately.

It was some buggy sh**

All that was with flaps UP. If all they changed was just the flaps, we'd still have those problems. So it's not just flaps. There have been major changes to lift/flight in general.

MOST are good.

Some are just.... "iffy"

[Murdr]

To my knowledge the last major flight model change was v2.07, and affected the thrust/drag model.  That will affect sustained turn rate and radius because the turn you can sustain will be at the point where your total drag is using all of your thrust.  Since we are talking about a point of equalibrium where we are sustaining a turn indefinately with little altitude or speed change, we are also talking about relatively slow speeds.  So this includes a speed range where flap usage can be a factor, and the lift/drag of those flaps will have an inpact.

by Pyro
Made several changes to the flight model affecting static thrust, propeller slipstream effects,
ground effect, flaps and slats.  This affects all planes to one degree or another.  Due to this change, the
table that controls combat trim had to be redone for all planes.

These changes did not benefit all planes equally.  For instance the P-51 appears to have lost stability, while the 109s and F4Us gained stability.  Why?  Hard to say.  Was it a specific item that was changed (for example, the decimal point typo on the full flaps for the P-38 in the initial 2.07 release)?  Or was it something inherent in the affected planes existing model that did not manifest itself until the changes were applied? 

I don't think the question really is "how does it work?", but rather "are the results correct?" 
I believe the answer is that there are some tweeks that need to be made.  I also believe tweeks will be made, but they have a full plate and it may be quite awile until they get around to splitting hairs with the flight model.

But to the original question, here are some semi related topics
link
link
link

[bozon]

So my main question is: if the airflow code is set up to mimic the actual fluid effects of air moving over the airframe (essentially, a virtual wind-tunnel) how could airflow ALONE be responsible for any overmodelling in the flight model (or under-modelling of other aircraft)?

Let me tell you something about flow models at high Reynolds numbers - they are crap. It is not HTC's fault, NOBODY has managed to give a reasonable solution to full flow equations set. The non-linearity f***s up numerical codes to the point that no computer today can save you. Deviate even slightly from ideal conditions and you get nonsense as results.

Theory is in even worse situations. There is no good explanation to why a flow through a tube turns turbulent at a certain Reynolds number. Theory says it is "linearly stable" and all disturbances should decay exponentially. Astrophysics still has problems explaining how a gas in a disc, at extremely high Reynolds numbers around a compact object is accreted. It shouldn't - both numerically and analytically.

Practical numerical codes are not pure physics. The algorithm is "guided" to converge into the desired kind of solution or some behavior is being enforced. The higher the Reynolds numbers, the worse it gets. Modeling the airflow around a wing at the onset of non-steady turbulance (stall) is near hopeless. Modeling the airflow at low angle of attack is bad enough, but workable.

HTCs code, like any other is an approximation that looses its accuracy fast when approaching extreme conditions - flaps out, near stall flight, are just those conditions.