Compressibility modelling
 

When diving under power aircraft speeds should be limited by compressibility effects. If not then aircraft that utilize vertical tactics like BnZ will get an unfair advantage over aircraft that employ angles tactics.

If airspeeds in the dive are not limited by compressibility, then the energy retention will be too high and an aircraft doing a BnZ attack will end up at to high altitude following a dive and zoom.

I don't know how/if this is modelled in the sim but I think this could be tested implicitly by comparing airspeeds from a test dive in the sim with the attached C++ simulation chart. The chart contains two graphs:

One is for the Spitfire Mk1 at +6.25 boost with compressibility modelled, both in terms of compressibility drag rise and also reduction in propeller efficiency due to Mach effects. The other has no compressibility correction whatsoever, either on drag rise or propeller efficiency. Both start from 23,000 ft altitude and 336 mph TAS.

In IL2 there was something called "devicelink" where one could extract this type of info and do a comparison. Don't know if this or something similar exists here. An alternative is of course simply to test dive and read off the speed at 2.5 Km altitude 61 s into the dive and see how this compares with the sim. If there is no way to read out TAS in the sim then I suppose one could convert the IAS cockpit reading if this is reliable enough.

I don't have the hardware to run this sim so I cannot test myself but I'm curious to know how the sim handles high speed dives.

OH MY GOD! Please stop to talk about unfair advantages in this game!!!

Aircrafts don't employ angle tactics: the pilots do. Nothing prohibits pilots to BnZ in an Hurricane... except their own noobness

It's just too difficult to track planes under your ship and too easy to spot aircrafts over you head... and this is really helpful for guys who don't care about tactics...
But NO... lets put in disadvantage the ones who actually do fly in the smart way!!!



Anyway I agree about compressibility modelling, but it's not a priority at all.

But when it comes to unfair advantages, I guess you want the same for angles fighters as well right? Otherwize there could be a Spitfire with a turn time of 21 s at 20,000 ft if Crumpp has his way :)

The Spit dive is just an example: The Me109 C++ simulation looks just the same. If no compressibility then too high speed in the dive and I assume everyone wants as close to historical performance as possible right?

I understand that there are more pressing mods to be done. While I have not simulated the Hurricane, the Spitfire top speed numbers looks way to low. I get 280 mph for the Spitfire for +6.25 boost and about 475-480 Km/h for the Me109 at SL 1.3 ata and I guess that would be a higher priority to fix.

First, I want to point out that my rant was referred to the "aircraft that employ angles tactics" statement. I go nuts when I read something similar...

Without a realistic environment around the planes it's dangerous to model some aspects that would totally ruin the combat aspect of the simulator.

Since, you know, there are guys that actually try to simulate combat tactics... if these are not practicable then what's the reason to fly this sim? For its "not really complex" engine management? To enjoy the historical speed of the plane at altitude ("Look I'm reaching the same speed of the real plane, best moment in my life!!!")?

It's just bad enough that fatigue is not modelled and every virtual pilot can sustain high G accelerations for dozens of minutes...

If the guys who follow the main rule of warbird dogfighting ("Altitude Is Life") can't even gain advantage from that (BnZ) then be sure the game will die very quickly. 1C don't even need to model altitude over the 10K feet since the "BnZ is not allowed" rule found in some IL2 servers would not be forced by those loser admins, but by the game itself.

Without a realistic world around the virtual pilot the single aircraft performance becomes far more important than the player own skill... infact in these weeks there are too much "X is too slow by 20km/h; we can't fight!!!" threads for my taste.

Can't wait the day a WW2 sim could actually simulate a realistic aircraft/engine dynamic quality (wearing)...

I'm not arguing to introduce something to upset the balance of gameplay.
I used to fly IL2 a lot back when I had more time and in that sim there were things that were off that really affected gameplay. I think the elevator authority on the Me109 was one example. IIRC this was off in IL2 in the sense that it seemed more connected to TAS than IAS.

I still think it would be good to get everything on the table (also the stuff you mentioned) but compressibility is one component there as well. As I said, I guess there are more pressing fixes to do to get the balance in gameplay right but that does not lessen my curiosity as to how the currently modelled planes behave in high speed dives :)

I want most things modelled too. Compressibility is one of these, of course.

And I don't want an arbitrary game balance: do I have to fly in a crap plane? I do it and I'll try to fly it in the historical way. (for example I16 vs 109).

The thing I don't want is implementing something that heavily affects the combat simulation over more important things.

Because in real life "Good Tactic" >>>>> "Raw Performance"...

Anyway I don't think the elevator authority issue was about TAS or IAS: it was modelled like a loss of control surface effectiveness instead of a simple stick heavyness... in the game it was impossible to operate the stick with the strength of 2 arms like the real pilots did... they roughly modelled a pilot's strength related aspect without taking in account the pilot's stamina/fatigue modelling.

I hope it will be possible in CloD's future.

Holtz. I don't know exactly if your saying Airspeed should be limited by compressibility, or if your saying at which Compressibility actually becomes noticeable to causing a runaway dive / increasing speed / decreasing control.

Or are you meaning loss in engine performance? Cause your right, its going to decrease performance, but in a dive, your still going to increasing speed from the simple physics of gravity, only acceleration would be affected. But it won't act like a "Dive Break," that's how I see you trying to explain it.

Thats pretty much correct. TAS / IAS really isn't the issue. From my understanding (8 years real pilot, however no real experience in this area other than study, and never will because I'm not a test pilot and testing this can be suicide) TAS / IAS compressibility affects don't slow down the aircraft / reducing speed of dive capabilities.

The real key to compressibility stall is the atmospheric conditions via what altitude your at. Higher alt= less dense air. I say this from the fact that most stories I've read about WW2 compressibility stalls, they always start out with the pilots typically above 25,000 ft. They get into a dive, and as they go faster the laminar flow of the high alt / less dense air around the wings and control areas is so reduced to the point that the pilot looses control / has a runaway dive. At which point the only thing they can do is chop power, but your speed is still going to increase because your diving.

Your only able to regain control of the aircraft when it decreases down to denser air density returning laminar flow to the aerodynamics of the aircraft. However by this time your either A) Going so fast and theres no return of laminar flow that controlability doesnt return. B) The aircraft begins or already has broken up from structure failure of approaching the speed of sound.

There is one story I remember reading about Robin Olds, in a P-38. Started a dive at bomber escort alt (think 30,000 range) Got into a compressibility stall and literally didn't pull out of it till a few 100 ft above the ground. The P-38 is one of the most notorious WW2 aircraft with this aerodynamic design problem.


But seeing as most aircraft in CLOD at the moment are not modeled correctly above 18,000, I see it as only being a miniscule problem till this is fixed, then worry about compressibility FM problems.

The compressibility issue the simulation shows is limited to the speed aspect and controllability is not modelled. The lower curve takes the compressibility drag rise and propeller efficiency loss due to speed and Mach effects into account while the higher curve assumes constant propeller efficiency and a "flat" Cdo which does not start to rise as the speed goes up. The problem is if compressibility is not modellled then any aircraft in the sim will build up too much speed in a dive meaning it will zoom too high in the pullout. This would work to the advantage of so-called BnZ tactics in the sim since energy retention would be higher than IRL.

Mach effects on compressibility are an interesting but complex subject but the C++ simulation posted does not model any impact on controllability due to Mach effects which is beyond the capabilities of the code which is limited to performance comparisons only. As you point out the P-38 was especially susceptible to this phenomena but the Mach "tuck" effect did affect the other types as well and it was more of a question at which Mach the problems started to appear. However, I think modelling this is even more complex and I think it would be a significant step in the right direction just to get the flight performance aspect of compressibility in to begin with.

While it was some time ago I looked into it, I still believe that the Me109 elevator control modelling in IL2 was off because it modelled loss of control authority more due to TAS than IAS: At low alt you were OK at certain IAS but at high alt you lost authority at the same IAS even though the Mach number was quite low. Since I have not flown the Me109 here in CloD I have no idea how this is modelled. Hopefully it has improved.

This is a good thread.

Manu, I don't believe that holtzauge was implying that using BnZ is unfair; only that aircraft whose performance was limited by compressibility effects would see an unfair advantage, I.e more speed and "e retention" than they would have historically had.

This is why some 1946 servers banned the 190D9, because it was possible to fly it and a few other aircraft through the sound barrier, since compressibility was not modeled.

Holtzauge, it would also be helpful to be able to see the source code for this c++ solver so we can audit it. Otherwise, and no offense intended, from our perspective we have only the word of an anonymous forum poster that the simulation is accurate. Ihope that makes sense.

We do not even know if compressibility is modelled or not. We do not have real world test figures for correct high speed dive behaviour, and a random poster's colored graphs based on cluster of guesswork figures and labeled '109' or 'Spitifre' hardly make a comparison basis.

It makes an interesting theoretical discussion but the whole thread is absolutely useless for checking the FM's validity. Unless someone would suggest the developer's should rely on hobbysts guesstimations of performance.

He didn't say that, of course and I agree the lack of compressibility gives an unfair advantage to EVERY diving plane... my reply was about the unfair advantage against "angle fighters".

My dream is to dive with the sun at my six and approaching slowly the enemy: in IL2 you need to go so fast since he can see you in 2 second (if the pilot is smart) and so you have almost no time to attack his plane (if you can spot and track it under your plane, of course).

With the implementation of a realistic visibility function (realistic scanning time and reviewed dot system) and the abolition of the damned engine radar then the compressibility issue can be limited, since we won't need to dive at high speed.

Because of this I said it's not only a issue of low priority, but it can damage the entire combat simulation.

That is correct. I am only raising the issue because I know it was missing from IL2 and I was curious how it was modelled in CloD.

As to posting my C++ source code I will be a bit stingy and keep that to myself. I'm actually working on a book project so I need it for my own selfish purposes. I do understand your scepticism though. I would not buy a pig in a poke either. I guess that means you need to apply a healthy dose of scepticism to my posts ;)

Would be nice if you could post some results though. If you have an idea about an alternative dive scenario, either Me109E or Spit Mk1 just let me know and I could model that as well. What I need is boost, weight, dive angle, initial altitude and TAS speed. Based on this I could generate another chart like the one I posted.

Elevator response is not quite so brick-like as it was in 1946 but then again these are emils we are flying, and not the G or K series.

Sad to hear. I am a big proponent of open source software but to each their own. Would it hurt your profits from the book?

Perhaps, once we get a good idea of how the atmosphere and aerodynamics are modeled in the sim, we could run some analyses using a free tool like OpenFOAM which I've used in the past. I may PM you in the future to pick your brain about simulating this stuff, Holtzauge

I will once I'm able, though I doubt my ability to hold a constant dive angle as precisely as a simulation

Agree with all this however I suspect this will be a hundred page thread or near enough... :rolleyes:

Ok Holtz, that makes more sense for me via your reply. Wasn't 100% what you were getting at, at first. TY

It occurred to me that Ace of Aces was conducting "e retention tests" a few years back using devicelink and a devicelink autopilot script, using a dive-zoom profile.

While I disagree with the idea of "energy retention" or as some call it "energy bleed", I seem to recall that he had some WWII era tests that used the same profile.

Might be good to use that same profile so we can use any existing historical tests to validate.

Here is one of the threads on the topic: http://forums.ubi.com/showthread.php...Status-(BETA)?

If I were doing the book stuff for money I would most likely starve to death anyway. In addition, I'm more of a system engineer these days so I do not want to to be embarrased by a professional programmer pointing out my poor syntax :)

Jokes aside: You are welocome to PM me though and I could tell you a bit about how I've gone about it.

Concerning the validity of the C++ modelling, I just remembered we did a comparison of the C++ compressibility modelling and what was implemented in IL2 at the time (4.09m). As can be seen there is good agreement in the models as long as the speed does not start to reach into the compressibility range. The figure also shows that the lag between the compressibility and non-compressibility models following the pullout. As can be seen, the non-compressibility model means the P-51 will end up higher than it should. I think the effects should be less here in CloD though, since the Me109E and Spitfire Mk1 were not as fast, not as powerful and had a higher Cdo/weight factor so it took them longer to reach compressibility speeds.

I do not master devicelink so Wurkeri helped me with the IL2 data which he sent me in Excel format so I could get that into the graph. One graph shows the raw data (red) and the other (black) where I translated the data so as to be aligned with the C++ data. I think this figure gives a good illustration of the problem: The aircraft should not in a dive speed up to a "point" but rather have the top truncated by compressibility effetcs.

Doggles: If you do run a test in CloD and can get the Spitfire or Me109 data into an Excel sheet then that would be good because then I can include that together with the C++ results. However, If it is difficult to get continuous data out of CloD then I think a simple dive test down to 2.5 Km alt and reading off the speed there ( or a number of dives and averaging the result would be even better) could do to begin with to get a ballpark estimate.

You will only be trading one set of assumptions for another set of inaccuracies.

Compressibility effects are accounted for by a tabular velocity correction in subsonic incompressible flow theory. The formulation for that correction is an approximation that is considered valid one up to ~Mach 2.3.

Who will determine the onset of compressibility and who will determine what theory we use? Each method accounts for the effects in varying amounts because each one gives different levels of agreement within small sections of the flight realm. Think of them like mathmatical microscopes, each is designed to exam with greater agreement a specific area of flight.

Now, that being said, there have been some real advances in this area. Our ability to express compressibility has greatly improved and I believe in our childrens lifetime, it will no longer be an issue.

Which area's are we going to trade one unrealistic set of assumptions for another unrealistic set of assumptions in order to be more realistics in say....

Mach .95 to Mach 2.2?? Nothing in WWII flew in this realm....

Or Mach .7 to Mach .9?? Nothing in WWII spent any significant time in this realm......

Or subsonic incompressible flow....

Zero airspeed to ~Mach .65 to .7, where ever you put the onset of compressibilty??? Hey, this is the realm WWII airplanes spend almost all of their time flying in!!

IMHO, this is an enviromental issue. The subsonic incompressible flow theory used in the game is considered valid and appropriate. The inaccuracies of it's tabular compressibility corrections still return a valid assumption and equally effects all the FM's.

The relative performance is in intact and all aircraft gain equally.

In short, you will not be making anything more accurate. You will only be trading one set of assumptions for another set of inaccuracies wrapped in a much more complicated package.

BTW, If you can dive from 7Km to 2.5KM at +6.25lbs in the Spitfire Mk I.....

If you mean you dove the aircraft at Emergency power.....

There is much bigger fish to fry in this "sim" than how compressibility is handled in transonic flight.

Personally I would like to see more realistic operation of the aircraft, a stability and control model, and atmospheric modeling.

If they get the atmospheric model down, nobody will be using high rpms and overboosting at high density alitudes.

Your point?

It is written in english. Is that not your language?

All aircraft will see a reduction in drag in the transsonic realm, true. But not all aircraft can get to that realm, and so see no benefit.

In 1946 it was possible to fly a handful of planes fast enough to just break Mach 1 or so, if you dove just right. I am sure that a guy working with a group that would restore Fw 190s would agree that D-9s were not capable of supersonic flight.

The simple fact is that all aircraft respond to the atmospheric model the same way, or at least we hope so, but not all aircraft can exploit flaws in said atmosphere.

The faster aircraft, whose performance was severely limited by wave drag, etc. in reality, gain a significant speed advantage over others.

It should be plain to see that the relative performance is not intact in this case.

Wow....that is not a good simultaion at all, LOL. I did not play IL2 1946 except on very rare occasion with a few of our members.

I do play some of IL2:COD. It is not the same game.

Given that the original sim was designed only to simulate the Sturmovik, I think it has held up quite well over the years. Looking at the engine in that light, their choice not to include some kind of linearized compressibility correction (Prandtl-Glauert or otherwise) is justified in my opinion.

Designing a game engine is just like engineering anything else. You select the use cases you want to have, and make simplifications without worrying too much about how they affect your outside cases.

Yes, getting at least some modelling of subsonic compressibility based on the Prandtl factor could have been a way to go to get at least something in there even though that would have affect all airplanes the same but if you look at wind tunnel data and results from flight tests then the different aircraft have significantly different drag characteristics in the M=0.5-0.85 range with some like the P-51 and Spitfire being better than others like the late war Me109's which suffered from an earlier onset of drag creep.

This report has some nice info on the P-51 and was one of the sources I used as input to the drag modelling.

http://ntrs.nasa.gov/search.jsp?R=19...20the%2520drag

In the IL2/C++ P-51 20 degree dive comparison the Cdo goes up from the low Mach Cdo of around 0.018 to about 0.034 at M=0.75 which is hardly negligible and explains the truncating effect on speed and why the Mach never goes higher in the C++ simulation. Since this increase in Cdo seems not to have been present in IL2 the speed increases unabated to 960 Km/h which seems a bit on the high side ;-)

Pity that this was not included in CloD when a lot of work seems to have been done in other areas such as damage modelling and improved graphics etc. Being familiar with IL2 I was hoping that CloD would also take a significant step towards better FM as well. If I had to choose between better graphics and better flight models then it would have been the latter no contest. On the other hand from a commercial perspective I guess better graphics wins the day over hard core simmers fancies so I can understand the priorities taken from that perspective.

I agree with the fm choice, but the problem is that u need a better physic management in general and then a good environment management (seeing the contacts, the clouds, tree collision etc etc), then the graphic. About the commercial perspective, seeing the results they had until now i would not guess so much ;).
About the compressibility, despite the interest of the argument, the list of the important stuff to correct or implement is so long that now it is a "not a problem". to give u an example less used of the wrong simulation of the atmosphere, the "amazing" Cem if you check the engineswith the megnetos, does not change the rpm, and for me this is a bigger feeling killer (an trust killer, as they pushed alot on the new total sim engine management).

I am strongly disagreeing with you, last time i checked the magnetos on a Bf110 the engine lost waaay too much rpm's when switching one off, - 500 instead of - 35 rpm's!!!

well if they fixed in all the airplanes (i tried on the bf109 and wasn't working at all), also if really late (it is something that I could not understand how they didn't notice at the start) it is a good news in a drab landscape.

Right and the point I was making.

When it comes time though, the community will have to be very careful in its selection of theory.

The agreement between theory and flight is not always very good. Some are worse than others.

Depending on what your doing with the Prandtl-Glauert correction, you can vary velocity for example as much as ~75 mph TAS over measured results or even more depending on the design specifics.

The basic correction agreement between calculated and wind tunnel data for a tailess DC-9 for example shows the Prandtl-Glauert correction to be some 66 mph TAS pessimestic over measured results.

Got a link?

Ah. Not quite what I expected.

Very good link - Thx

Most of the text seems to be extracted from an abridged version of the book written by the father of modern aerodynamics at NACA/NASA in the early 50's. This book is available on Amazon at such a cheap price that's it's a shame that it's not part of the basic every engineering student shld have right on his desk. No, most of the time you have to buy expensive book written by Mr Professor or one of his collegue with only a fraction of the knowledge available inside. I have alrdy given the title somewhere.

Ok just that to say that Crump is really helping the comunity to step forwrd. He has is own style, that's for sure, but ehhh, why do you think some of us are bidding so much time writing here ?

And for all the aficianados of obscurantism that want to raise the 1940 RAF in lieu of a new religion may I remind you that this fight was all about putting an end to obscurantism . You seems to forgot that with your arguments hammered contently in every forum despite the fact tht Britain in those times was the bright and best example of a democracy at war struggling with all her might to survive - making me really thinking of - Oh funny is that - to France at the time of the Napoleon's war ! (yeah I know you just fall of your chair ;-D ) - or Israel in the early times - and Israel strategy was very much inspired by the Britain example )

However, I would remind to all that all those tearing depate for who is the real pilot and who is not, this sentence from a test pilot : None pilots will ever do a a good engineer and vice versa. You don't have to expect the same from the ones that are proud to put frwrd their"pants" when other swear only by their brains.

I think we need some kind of PETA's campaign on this forum with some kind of slogans like " Free the Crumpp" or whatever because the constant firing on him look like very much the beating of baby seals.

just my two cents.

~S to all !