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

[Alpine Scenery 1,609]

Navier stokes, but no fixed textures folks...

 

[edpatino 1,980]

Very good this fluid dynamics simulation is coming to MSFS in SU11. I would expect it'll be improving over time after it first release.

[Bartul 46]

Man am I excited! My electrons are so excited I glow in the dark!

[DAD 2,662]

3 hours ago, haskell said:

I KNEW it couldn’t be Cheyenne Frontier Days. 

Not to mistake with CTD. Wait, a 20km simulated CTD area around all aircraft would be fun, wouldn’t it 🙂

[abrams_tank 7,477]

6 hours ago, jcomm said:

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

You can actually see the vertical components of the airflow around an aircraft in that simulator using the available on-screen output, and X-Plane has modelled since XP10 the deflection of air currents due to orography, as well as wake turbulence, shear and even microbursts.

The difference IMO is that ASOBO is taking a slightly more elaborate approach to the modelling of some of these currents, including the way convection is represented, the sources of energy that feed it, and the combination of various air currents into more complex patterns, as far as I could tell from some of the videos ASOBO has published.

My main focus goes to gliders and their new functionalities and to rotary wing, but this new "cube" around the aircraft is surely going to be food for my tests.

Presently thermals aren't very consistently modelled, and it's evident from the way you catch them in simulated gliders and the kind of flight path you're able to follow in order to try to stay within their "cone of action", so, I hope this got fine tuned for the upcoming soaring weather in SU11.

But, as a comparison from previous versions of MS SF, WOW!!! that's surely a HUGE step ahead...

Hi Jcomm. I don't think a 20 KM CFD around the plane in MSFS has been done by XP before as @lwt1971  pointed out, at least not on the scale of 20 KM around the plane, with the air movement in the environment being simulated for 20 KM around the plane (or something comparable to that scale).

The 20 KM CFD means that air movement from as far as 20 KM away from the plane, could eventually affect the plane as the plane approaches that air flow.  For example, air that bounces off and moves over a hill or tree from 20 KM in front of the plane, if the plane continues moving in the direction of that air flow movement and gets closer to it, that air that bounced over the tree or hill can eventually affect the plane if it eventually collides with the plane.  I have not seen any video or photographic demonstration of XP where air flow from as far as 20 KM away, can affect the plane.

You'll have to provide video or photographic evidence showing the air flow movement lines in XP being simulated for 20 KM (or something comparable to that scale) around the plane if this is what you claim.

 

Edited November 5, 2022 by abrams_tank

[spearmint_flyer 580]

7 hours ago, mspencer said:

For everyone else who had to Google this, it's computational fluid dynamics.

You’re asking and answering the correct questions. You genius you. 

[jcomm 6,020]

1 hour ago, abrams_tank said:

Hi Jcomm. I don't think a 20 KM CFD around the plane in MSFS has been done by XP before as @lwt1971  pointed out, at least not on the scale of 20 KM around the plane, with the air movement in the environment being simulated for 20 KM around the plane (or something comparable to that scale).

The 20 KM CFD means that air movement from as far as 20 KM away from the plane, could eventually affect the plane as the plane approaches that air flow.  For example, air that bounces off and moves over a hill or tree from 20 KM in front of the plane, if the plane continues moving in the direction of that air flow movement and gets closer to it, that air that bounced over the tree or hill can eventually affect the plane if it eventually collides with the plane.  I have not seen any video or photographic demonstration of XP where air flow from as far as 20 KM away, can affect the plane.

You'll have to provide video or photographic evidence showing the air flow movement lines in XP being simulated for 20 KM (or something comparable to that scale) around the plane if this is what you claim.

 

You're right abrams_tank, indeed nothing comparable to that scale.

In XP, using the FM visualization mode ( CTRL-M ) we can observe a "cube" around the aircraft, proportiional to it's dimensions, but never that big.

Also in XP we do not have the kind of detail I see in SU11 regarding the combination / interaction of different air masses including an image I "suspect" can be an hint of something like "convergence" being simulated AND! if that's the case it'll be a first time ever since I use flight simulators, including soaring flight simulators. 

So, I am certainly expecting to really get even more excited whe I start using SU11 ! Specially the rotaries, more even than the gliders 🙂

[abrams_tank 7,477]

3 minutes ago, jcomm said:

You're right abrams_tank, indeed nothing comparable to that scale.

In XP, using the FM visualization mode ( CTRL-M ) we can observe a "cube" around the aircraft, proportiional to it's dimensions, but never that big.

Also in XP we do not have the kind of detail I see in SU11 regarding the combination / interaction of different air masses including an image I "suspect" can be an hint of something like "convergence" being simulated AND! if that's the case it'll be a first time ever since I use flight simulators, including soaring flight simulators. 

So, I am certainly expecting to really get even more excited whe I start using SU11 ! Specially the rotaries, more even than the gliders 🙂

Ahh thanks, I thought so!

If the cube around the aircraft in XP is not that big (proportional to the aircraft dimensions as you say), then if there are hills that are say, 50 meters away from the plane in XP, those hills are outside the reach of that cube.  This means that on landing, those hills from 50 meters away have no effect on the plane in XP.

In MSFS with this 20 KM CFD, the hills from 50 meters away, if there is wind and air flowing around/over those hills, that wind/air flowing around/over those hills can affect the plane on landing 👍

[JDWalley 354]

9 hours ago, mspencer said:

For everyone else who had to Google this, it's computational fluid dynamics.

And here I thought I was going to need to correct them and point out that the term is "CTD"... 😁

[Gerwil 179]

nevermind 😊

Edited November 5, 2022 by Gerwil

[wim123 765]

i wonder what the performance cost is going to be..

[abrams_tank 7,477]

7 minutes ago, wim123 said:

i wonder what the performance cost is going to be..

Performance seems acceptable since there are quite a few people in the SU11 beta already and I'm not hearing large complaints about reduced performance because the 20 KM CFD is being used. Of course, from my understanding, they are not doing the full CFD simulation at high resolution because that takes ages to calculate and probably requires a super computer to calculate it too.  I believe they are using a more coarse method of simulating it so that it works well on the common PC gaming computer.

[ha5mvo 261]

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.

[lwt1971 7,464]

6 hours ago, abrams_tank said:

Performance seems acceptable since there are quite a few people in the SU11 beta already and I'm not hearing large complaints about reduced performance because the 20 KM CFD is being used. Of course, from my understanding, they are not doing the full CFD simulation at high resolution because that takes ages to calculate and probably requires a super computer to calculate it too.  I believe they are using a more coarse method of simulating it so that it works well on the common PC gaming computer.

Yup so we know for sure the local aircraft CFD that came out in SU9 can be configured to be more fine-grained at the expense of computing resources... I'd assume so can the new atmospheric airflow CFD. I need to try it out again but since SU9 for the aircraft with local CFD enabled in their FMs you can tweak various things in the CFD debug window (in dev mode) in addition to hand-editing the flight_model.cfg.

https://docs.flightsimulator.com/flighting/html/Developer_Mode/Aircraft_Editor/Debug/Debug_Aircraft_CFD.htm
The CFD currently proposed uses a default of 20 x 20 x 20 cubic voxels that englobe the aircraft. When enabled, the simulation will then solve a custom version of the Navier Stokes equations that includes upwind advection solving designed to cope with high fluid velocities. The CFD is solved over each of the voxels by the CPU at a rate of 100 times per second with three global passes per iteration. The load is spread over 5 different threads for minimal cpu impact.

Looking more closely at all the CFD related parameters/config available in the FM (https://docs.flightsimulator.com/flighting/html/Content_Configuration/SimObjects/Aircraft_SimO/flight_model/aerodynamics.htm#CFD_EnableSimulation) there is: 
CFD_VoxelNbVoxels: This can be used to set the number of voxels that will be cubed to make the sample volume for the CFD simulation.
IMPORTANT! This may have a serious impact on performance if set to values greater than the default value, due to it currently having a time complexity of O(n3).
Default value is 20.0, and the value will only be used when the CFD_EnableSimulation parameter is set to 1 (TRUE).

So due to the exponential O(n^3) complexity where n is the # of voxels, increasing the CFD calculations to be more fine-grained will quickly eat up CPU cycles.. and this is even for their "custom" version of the Navier Stokes set of equations which is likely a subset of the full version which would definitely need supercomputers :)

Like you said I'm also not seeing any reports of performance degradation for the various users already testing the SU11 beta.

 

Edited November 5, 2022 by lwt1971

[MarcG 3,035]

Anything that adds to the realism of flight is always exciting for me, even though I'm a novice in the world of flying I still want to be challenging myself in flight in realistic conditions. This new (to MSFS) tech is a great move by Asobo, the physics department are clearing striving to be as close to reality as possible and they're going to great lengths to achieve this, recent videos on the choppers & gliders have shown exactly that.