Deeper accuracy in the flight model

[GoranM 3,764]

This didn't really get a proper mention in here, and I thought it deserves it's own thread.   This looks very fascinating to me and appears to make the X-Plane flight model even more sophisticated. I did some asking around and from the responses I got, I don't think something like wing sweep physics and Critical Mach Number has been done before in a home market flight simulator.

If it's true that wing sweep and critical mach number hasn't been done before in a home market flight simulator (based on the responses I got when I asked around), this is really ground breaking flight model technology for X-Plane. 
 

 

Edited January 1, 2023 by GoranM

[jcomm 5,929]

Hi Goran,

yes, it has been done in at least:

FlightGear

DCS World

IL2 Sturmovik

Aerowinx PSX

now also, if I'm not wrong, in Aerofly FS

BTW: I think Austin didn't correctly code the drag side of the sweep-back into his latest update, so he is taking the exact same cosine reduction in effective flow when in fact for the drag he should consider direct stream... 

I think he did nonetheless take into consideration as per Alexis and Murmur suggestions, using the square root of the cosine in order to more properly cope with the fact that there's a difference between a theoretically infinite wing and a finite one, as some articles sent to him suggest.

Edited January 1, 2023 by cagarini

[GoranM 3,764]

Interesting. Do you have any links where I can look into this to verify for myself?

[jcomm 5,929]

It's been a long time since I participated in DCS World and IL2 forum threads about the theme, and I do no longer even use DCS, but as you may guess DCS has been modelling modern fighter jets since quite a while, and the Professional FMs have long coped with those effects.

I just found some videos although not that technical, but where even the effects of asymmetric wing sweep on aircraft with variable wing geometry are modelled:

Questioned: Can A Jet Fly Higher With Straight Or Swept Wings? | DCS WORLD - YouTube

F-14B Tomcat: Operations With Asymmetric Wing Sweep | DCS WORLD - YouTube

Questioned: Is "Mach Tuck" Modelled In DCS WORLD? - YouTube

In FlightGear, which I also seldom use, I remember a very detailled JSBSim-based TOMCAT model:

Grumman F-14 Tomcat - FlightGear wiki

read the wiki to find mentions to those effects being modelled.

In Aerowinx PSX Hardy Heinlin even models reversal effects due to elasticity in the control surfaces - ailerons. 

Aerofly FS 4 has recently released their Concorde and while I am not absolutely sure about where I read it, developers did mention it, and if you try to dig into the flight dynamics wiki you might find it there too...

There are many references to the cosine factor for infinite wings, like in here: High Speed Aerodynamics, swept back wings

It's in more technical documents that you will be able to find additional modelling details suggesting the use of the square root, or even less, of the cosine to be taken as a correction factor to cope with finite swept back wings, as well as how to model the effects regarding Cd (linear in that case).

 

Edited January 1, 2023 by cagarini

[GoranM 3,764]

@cagarini I will definitely look deeper into those flight models.  I really want to see if this has been done to the extent you say it has.  I had a look into the F14 in FG, and although I see a mention of "Full aerodynamic affects of wing sweep are modelled.", it doesn't really say, in detail, to what extent.  "Full" can mean anything.  And I can't see any mention of Critical Mach Number.  Unless I missed it somewhere.  

What I love about Austins video is he goes into every detail of the subject.  But I'll dig a little deeper, anyway.

Edited January 1, 2023 by GoranM

[jcomm 5,929]

Well,

one thing is for sure, X-Plane has been modeling, in the area of general purpose civil flight simulation, details that no other simulator, with some remarkable exceptions for some aspects modelled by AEFS and FlightGear, do.

For instance, aircraft with multiple wings, tails, or even asymmetric geometry can me modeled in XP and not, for instance, im FSX and derivates or MFS.

Modelling of propwash effects, modelling of failures, etc.., are yet other aspects that put XP, IMO, ahead of other general purpose civil flight simulators.

This particular aspect of swept-back wing aerodynamics still requires further tuning IMO, and that is problematic because if Austin changes calculations the developers will have to re-adapt their creations to it, but in the end we all win...

[GoranM 3,764]

1 minute ago, cagarini said:

This particular aspect of swept-back wing aerodynamics still requires further tuning IMO

You and Austin are almost best buddies now, after all your emails to him.  I'm sure he would welcome your input on that.

 

[jcomm 5,929]

6 minutes ago, GoranM said:

You and Austin are almost best buddies now, after all your emails to him.  I'm sure he would welcome your input on that.

 

Well, Austin is Unique in the way he always cares to answer user's emails / suggestions, even if the replies we get aren't always the best we wanted it to be 🙂

From what I gather regarding the "slack team" forum, this is also true regarding interaction between the whole LR team and active developers, and is fundamental to build an even better platform by continuous development. XP has always been a dynamic platform, and each new version grows towards "perfection" all along it's run. XP12 IMO is probably the one which started higher on most fronts compared to previous versions, so from here on, it's just a question of fine tuning some aspects.

[blingthinger 649]

7 hours ago, cagarini said:

IL2 Sturmovik

Aerowinx PSX

Do you know if these are distributed point force models? 

I think I saw an explanation by murmur a while back on .org that DCS can be either. Sounded like a JSBsim concept that allows the user to specify the final cumulative forces however they desire (distributed or single point).

[Bjoern 1,157]

6 hours ago, cagarini said:

Well, Austin is Unique in the way he always cares to answer user's emails / suggestions, even if the replies we get aren't always the best we wanted it to be 🙂

His "ok" reply (literally the only content) to my preceding "Yeah, your solution worked because [yadda yadda]" e-mail was my personal crowning moment of Austin.

[jcomm 5,929]

1 hour ago, Bjoern said:

His "ok" reply (literally the only content) to my preceding "Yeah, your solution worked because [yadda yadda]" e-mail was my personal crowning moment of Austin.

🤣

[blingthinger 649]

 

On 1/1/2023 at 3:24 AM, GoranM said:

If it's true that wing sweep and critical mach number hasn't been done before in a home market flight simulator (based on the responses I got when I asked around), this is really ground breaking flight model technology for X-Plane.

I was sifting through some of the F14 docs buried in the links cagarini posted above. The flightgear model is 'hangared' at this page:

https://zaretto.com/f-14

Even though JBSsim can ostensibly handle a multi point force model, they are using single point force windtunnel data:

"There are 63 distinct detailed aerodynamic coefficients"

Certainly my favorite quote on that page. Most people only consider lift, drag, moments. In those url links there are tons of wind tunnel data. For example, describing yaw rate when the inner spoilers are deployed (one coefficient!) and another plot showing yaw rate when outer spoilers are deployed (two!, two coefficients!) and on. There's a ton of (surprisingly) public data for that airframe. Richard Harrison has done an amazing job of creating the F14 in flightgear.

 

BUT that's not 'simulating' wing sweep and mcrit. It's replaying the overall behaviors of the airframe centered on a single point in space at a very detailed level but not actually simulating the physics behaviors on the wings themselves. 

Like you point out, XP does actually simulate the physics and there's a ton of airframe motion that inherently falls out of the model because of it.

For example, say the flaps fail on one wing: because the left wing forces are independent of the right (with up to 40 point forces on each, contrary to the slander/lies propagated by competing 'force element' documentation), the pitch/yaw/roll moments induced on the airframe are an inherent by-product of the simulation. Austin likes to use the term 'emergent'. No need for pages and pages of aerodynamic tables.

There's no other real-time sim that does what XP is doing.

Edited January 2, 2023 by blingthinger

[jcomm 5,929]

1 hour ago, blingthinger said:

 

I was sifting through some of the F14 docs buried in the links cagarini posted above. The flightgear model is 'hangared' at this page:

https://zaretto.com/f-14

Even though JBSsim can ostensibly handle a multi point force model, they are using single point force windtunnel data:

"There are 63 distinct detailed aerodynamic coefficients"

Certainly my favorite quote on that page. Most people only consider lift, drag, moments. In those url links, there is tons of wind tunnel data. For example, describing yaw rate when the inner spoilers are deployed (one coefficient!) and another plot showing yaw rate when outer spoilers are deployed (two!, two coefficients!) and on. There's a ton of (surprisingly) public data for that airframe. Richard Harrison has done an amazing job of creating the F14 in flightgear.

 

BUT that's not 'simulating' wing sweep and mcrit. It's replaying the overall behaviors of the airframe centered on a single point in space at a very detailed level bu not actually simulating the physics behaviors on the wings themselves. 

Like you point out, XP does actually simulate the physics and there's a ton of airframe motion that inherently falls out of the model because of it.

For example, say the flaps fail on one wing: because the left wing forces are independent of the right (with up to 40 point forces on each, contrary to the slander/lies propagated by competing 'force element' documentation), the pitch/yaw/roll moments induced on the airframe are an inherent by-product of the simulation. Austin likes to use the term 'emergent'. No need for pages and pages of aerodynamic tables.

There's no other real-time sim that does what XP is doing.

You're right ( and Goran was too ) regarding that F-14 FG model, and actually although I didn't have much time to read the latest JSBSim docs, it appears as definitely not being natively modelled. I know YaSim has also been used to model some swept-back wing airliners, but it's still basically the same approach as MFS. I'd say MFS uses something pretty similar to YaSim: JSBSim vs YASim - FlightGear wiki

I haven't yet found DCS references to how it is implemented there.

I don't know how Aerowinx PSX works, but I'd bet it's based on the same model used in FG.

Bummer, XP12 does it like no other!!! 🙂

They still have to fix those drag calcs though...

Edited January 2, 2023 by cagarini

[snglecoil 1,828]

I’ve got to say, I don’t understand all of the flight modeling calculations you guys are discussing, but as someone transitioning to my first jet job and digging into swept wing theory, I appreciate the effort to see it modeled with a higher level of fidelity in a (largely) home based sim.

[jcomm 5,929]

40 minutes ago, snglecoil said:

I’ve got to say, I don’t understand all of the flight modeling calculations you guys are discussing, but as someone transitioning to my first jet job and digging into swept wing theory, I appreciate the effort to see it modeled with a higher level of fidelity in a (largely) home based sim.

A very intuitive text here: Why Do Fast Aircraft Have Swept Wings? | Boldmethod

and here too: Swept wing - Wikipedia

or here: Sweep Angle and Supersonic Flight | AeroToolbox ( refers to the "cosine factor"... )

 

Edited January 2, 2023 by cagarini