Is anybody else excited about the new 20 KM CFD in SU11?

[Greazer 1,321]

17 hours ago, jcomm said:

Has  been done in X-Plane for some time, I should point out.

Yea but XP only models the box around the aircraft. Not environment flows.  If I fly over the parking lot at Walmart I would expect some uplift (from heated concrete). Not going to happen in XP.  It doesn't come to Austin's attention, that whole 'environment' thing.

[BobFS88 635]

13 minutes ago, Greazer said:

Yea but XP only models the box around the aircraft. Not environment flows.  If I fly over the parking lot at Walmart I would expect some uplift (from heated concrete).

And maybe they have not got around to that yet or revealed how they want to represent it. But whatever they do will be on first principle.

The time the beta has been out is still young with only month in. There is still more work to be done and were just on beta 11. 

 

15 minutes ago, Greazer said:

Not going to happen in XP.

You do know what you don't know. 

[abrams_tank 7,423]

5 hours ago, ha5mvo said:

At this point it's nothing more than just a bunch of letters and some colorful squiggly lines. All the rest is mere speculation based on very little.

Once it's out, I guess we'll see what it's really all about and how much it adds to or subtracts off the product.

There is like 2 different threads on the new 20 KM CFD in the official MSFS forum with people that have tried it already in SU 11.

There really isn't anything to speculate about.  Just go through those 2 threads and there is feedback on already on how it works from various MSFS users that downloaded the SU 11 beta (of course the 20 KM CFD isn't perfect and needs further tweaking, as per the feedback in those 2 threads).  And if you can't wait for it to be officially released, you can try the SU 11 beta for yourself.

 

Edited November 5, 2022 by abrams_tank

[abrams_tank 7,423]

20 hours ago, lwt1971 said:

This is a great first attempt in SU11 by MS/Asobo to realistically simulate atmospheric airflow, thermals, turbulence of various kinds, vortices, etc and the aerodynamics/physics interaction of said airflow with the aircraft (thus the 20km cocoon around aircraft). Hopefully they continue to build on it in SU12 and beyond.

A lot great information and findings are in this thread on the official forums https://forums.flightsimulator.com/t/testing-the-new-airflow-simulation/549454 for those wanting to learn more about what's coming. The updated weather debug window provides a lot of info now on all the factors involved and what's being simulated. Also this thread https://forums.flightsimulator.com/t/cfd-is-amazing/549365 on both aircraft CFD and atmospheric CFD is worth a read. Especially cool how turbulence now is actually based on interactions of various airflows/thermals/etc, such as rotor turbulence in the screencap below from the first thread.

Thermals themselves are based on a multitude of factors as stated in that thread:

• Position of the sun (Day of year, time of day, location)
• Temperature
• Wind speed
• Clouds
• Albedo (Colors) of the satellite imagery
• Trees
• Water
• Scenery
• Trees
• Terrain slope

Thanks for posting this! The 20 KM CFD looks very promising.  Based on what I read in the official MSFS forums from the people that have tested it in the SU 11 beta, the 20 KM CFD largely works well, but it still needs some further tweaking.  I think once it has been tweaked over the next several Sim Updates (so several Sim Updates may take an entire year, at the current rate Sim Updates come out), it should be a pretty good system.

[jcomm 5,939]

1 hour ago, Greazer said:

Yea but XP only models the box around the aircraft. Not environment flows.  If I fly over the parking lot at Walmart I would expect some uplift (from heated concrete). Not going to happen in XP.  It doesn't come to Austin's attention, that whole 'environment' thing.

Nor convergence, which appears to be present in the MFS 20km cube model, and probably other forms of flow interaction too, which will be interesting to try with gliders...

Oh, if only gliders felt realistic 😕  They're so miserably modeled in MFS SU10 and XP11/12 ...

Will have to test - I didn't have the chance to properly test SU11 beta...

Edited November 5, 2022 by jcomm

[B777ER 2,486]

If it helps me to stop bouncing the Kodiak on landing then yes. 

[Pathfinder633 223]

offensive initials not allowed is a Voxel?

[jcomm 5,939]

1 hour ago, Pathfinder633 said:

offensive initials not allowed is a Voxel?

https://en.wikipedia.org/wiki/Voxel

 

[Krakin 5,770]

I did some research and from what I'm seeing, X-Plane does not use a CFD at all. Blade element theory is a completely different thing.

[abrams_tank 7,423]

6 hours ago, Krakin said:

I did some research and from what I'm seeing, X-Plane does not use a CFD at all. Blade element theory is a completely different thing.

Thanks for reporting your findings with your research!

The 20 KM CFD for MSFS is definitely interesting as I would think that any type of environment simulation for air flow would probably have to use a similar method to CFD - that is, an environment simulation for air flow should be simulating how the air flows around mountains, hills, trees, buildings, objects, etc.  Maybe MSFS will be the first home market flight simulator to implement this and then the other flight simulators, even the combat flight simulators, may implement a similar technology in the future to simulate air flow in the environment, including a somewhat longer distance from the plane (20 KM in the case of MSFS), in their flight model.

The 20 KM CFD is definitely a nice step forward, and enhances the flight model in MSFS.  There is still a lot of work from the MSFS team to work on the flight model, but IMO, this is a huge step and huge advancement in the flight model in MSFS.

Edited November 6, 2022 by abrams_tank

[SAS443 1,723]

Any word on the actual CFD model quality/resolution? I mean a fine vs coarse model is night & day regarding computational complexity. It's not like we have an abundance of CPU overhead as it is. 

edit: or is the heavy lifting done on Azure  side?😁

Edited November 6, 2022 by SAS443

[jcomm 5,939]

2 hours ago, SAS443 said:

Any word on the actual CFD model quality/resolution? I mean a fine vs coarse model is night & day regarding computational complexity. It's not like we have an abundance of CPU overhead as it is. 

edit: or is the heavy lifting done on Azure  side?😁

Surely very coarse, otherwise couldn't be usable.

Even the professional desktop tools for aerodynamics simulation using CFD approaches take minutes in the most powerful desktops to simply study the airflow over the wing / fuselage / tail of a model aircraft, thrust factors apart...

Simulating Stall of the DarkAero 1 Using SimScale Cloud-Based CFD - YouTube

ANSYS Fluent Tutorial I Turbulent Air Flow Over Car - YouTube

I am still unable to appreciate the new CFD-approach when, for instance, I take the default fully-CFD-enabled C172 G1000 for a ride, specially if I take it to closer to stall / post stall situations... 

I think that among the various default GA prop in MFS Standard Edition, the Cessna 152 is the most convincing.

Then if I take the supposedly best Spitfire for MFS - Flying Iron Spitfire Mk IXc - it feels far from what I can experience in DCS, IL-2 and even in some selected Spitfire models for X-Plane. X-Plane ground physics for taildaggers are still highly bugged too, unless fancy tricks with hidden structures are used.... DCS and even better IL-2, model taildraggers amazingly!

I believe there is a lot of potential in MFS's "CFD" approach, but while the whole idea is great, the end results still leave a lot to be desired...

What we like - well I do - is the feel of the aircraft inflight compared to previous versions of MS FS. Nowadays that feel even makes XP look like on-rails at times, but that is mostly due to the new "active" weather effects. The amazing scenery, the Weather effects and Rendering, the functionality of the various default features that previously had to be bought and made us spend hundreds if not thousands of dollars/euros to bring our previous MS FS or XP sims to an acceptbale level of detail according to our preferences... These all make us somehow forget about how limited some aspects of the flight and systems modelling still are.

Spinning the default C172 results in a very questionable set of rather unplausible situations... I would expect a lot better from the "CFD" approach.

Testing prop dynamics also brings a few evident limitations, like no influence of sideslip in prop bahaviour on a free-running turbine turboprop with it's engine shutdown inflight and the prop feathered ( or not )...

These and other "basic" aspects of flight simulation are still rather limited or unplausiby moedeled in MFS, while simulators like DCS and IL-2 have been doing it in a rather detailled / plausible way since ever.

My first test flight when I get SU11 40th Aniversary installed will be directed towards one of the helis. I really look fwd to test it ! Then one of the gliders. So far all of the gliders I have tested in MFS feel poorly modeled, I'd say not better and not worse than the gliders in XP... And, I would think, gliders should be a lot easier to model acceptably since they don't even have to deal with any moments due to thrust, prop effects, etc...

As I was saying in a PM to one of you, MFS is great, I'd say 2nd to none, when I use it to proceduraly simulate an airline flight. I use Fenix the most for that, the default ATC ( which has many limitations but is a lot better than what it was in previous versions of MS FS ) the default Online AI raffic, and I really feel great ( although the Fenix still wants to rock the winds at times during the bellow 30 ft AGL blend from Normal to "Ground" / "Direct" law...).

Edited November 6, 2022 by jcomm

[KERNEL32 500]

I do not understand, do we need to always turn on the weather CFD via the dev mode weather debug menu every time we want to use the weather CFD? Or is it "always on"? I'm asking about the actual modelling/injection not the visualization..

Edited November 6, 2022 by KERNEL32

[Krakin 5,770]

3 hours ago, KERNEL32 said:

I do not understand, do we need to always turn on the weather CFD via the dev mode weather debug menu every time we want to use the weather CFD? Or is it "always on"? I'm asking about the actual modelling/injection not the visualization..

It is always on. If you want to see the visual representation of it with the lines, you go into dev mode and turn some stuff on

Edited November 6, 2022 by Krakin

[KERNEL32 500]

1 hour ago, Krakin said:

It is always on. If you want to see the visual representation of it with the lines, you go into dev mosde and turn some stuff on

OK thank you! It was a bit confusing to me because if we go to dev mode then the weather debug menu - the options to have it on and the injection check box both are un-checked.