Well,

it's WIP, and it's a BETA, so, we must give it the benefit of doubt....

An interesting thread is developing at the official forum where various weirdinesses are identified and documented when possible.

Feedback for CFD simulation on C172 - Sim Update 9 BETA / Aircraft & Systems - Microsoft Flight Simulator Forums

It's worth reading, and the tests should probably be made with the C172 only, which  appears to be the aircraft where some new fine tunning parameters for the new CFD were introduced and tweaked for this SU9 Beta. So, while the visualization can be enabled for other aircraft, I believe the absence of those new parameters can cause even more weirdiness on the outcome of your tests.

Even the C172 presents some really strange behaviours. The first I noticed was a totally irrealistic lack of the need for right rudder on takeoffs and full power steep climbs.

As some of the posts mention at that thread there is actually a tendency for rolling / yawing right at high power and AoA, which is opposite RL. I. e. this post is a good reference in that thread: https://forums.flightsimulator.com/t/feedback-for-cfd-simulation-on-c172/508234/51

Spins feel strange, as do simple stalls.

It's probably the lack of more fine tunning, and it's a beta, so, let's wait and see what comes for SU9 final.

Edited April 3, 20224 yr by jcomm

Quote

Even this C172 presents some really strange behaviours. The first I noticed was a totally irrealistic lack of the need for right rudder on takeoffs and full power steep climbs.

 

Its always been like this, Asobo has been told many times but done nothing to address this, perhaps they simply dont understand.

 

they need to hire A2A to do it properly.

i do want a decent 172 in the sim.

i find the 152 much better done, default.

Excellent post by Alexis on that Official Forum thread I mentioned above:

https://forums.flightsimulator.com/t/feedback-for-cfd-simulation-on-c172/508234/78

There's still a Long way to go for ASOBO to really get it going with thst new approach...

Edited April 3, 20224 yr by jcomm

As is the case with every other home-based simulator, the result is only as good as the data input.

Many sims proclaim superior methods (one in particular making a big deal of it), but in my experience that's all guff. As clever as any 'fluid computation' wants to be, it still relies on data tables and algorithms to actually see something in the sim. 

The new CFD logic is only active on the G1000 172, so it can only be tested there.

I think it's great they're pushing this forward and still we shouldn't be expecting miracles in the first beta.

they are testing it out, so only on one aircraft, to see how it goes.  i say that is smart.

On 4/3/2022 at 3:49 PM, jcomm said:

Well,

it's WIP, and it's a BETA, so, we must give it the benefit of doubt....

An interesting thread is developing at the official forum where various weirdinesses are identified and documented when possible.

Feedback for CFD simulation on C172 - Sim Update 9 BETA / Aircraft & Systems - Microsoft Flight Simulator Forums

It's worth reading, and the tests should probably be made with the C172 only, which  appears to be the aircraft where some new fine tunning parameters for the new CFD were introduced and tweaked for this SU9 Beta. So, while the visualization can be enabled for other aircraft, I believe the absence of those new parameters can cause even more weirdiness on the outcome of your tests.

Even the C172 presents some really strange behaviours. The first I noticed was a totally irrealistic lack of the need for right rudder on takeoffs and full power steep climbs.

As some of the posts mention at that thread there is actually a tendency for rolling / yawing right at high power and AoA, which is opposite RL. I. e. this post is a good reference in that thread: https://forums.flightsimulator.com/t/feedback-for-cfd-simulation-on-c172/508234/51

Spins feel strange, as do simple stalls.

It's probably the lack of more fine tunning, and it's a beta, so, let's wait and see what comes for SU9 final.

SU9 beta was updated today with fixes to the CFD for the 172.  Please test the new beta changes and give feedback to Asobo ASAP while there is still time for one more possible beta patch.

Thx @abrams_tank  👍

they also released a very useful explanatory document:

https://devsupport.flightsimulator.com/storage/attachments/2425-debug-aircraft-cfd.pdf

I will surely give it a try when possible.

At least they're imaginative / entrepreneur in their aim at building a new, more powerful, flight dynamics engine for MFS.

I always have to make me remember they said it's a 10 yr Project, and I have to givi it time / momentum...

Edited April 15, 20224 yr by jcomm

On 4/3/2022 at 10:01 AM, Pathfinder633 said:

 

Its always been like this, Asobo has been told many times but done nothing to address this, perhaps they simply dont understand.

 

I rather believe that game development is a complex task and moving one thing will cause something else to break. So they have to be very careful implementing changes. They have a lot of aerodynamics experts consulting them, so I am pretty sure most issues have already been talked about. We just need patience and some 3PD that work around the shortcomings - as they've always done in any sim.

I mean, we had high quality addons in FSX - no need to panic. On top, the PMDG is just around the corner.

@mods: ups seems like I had an old, unused account setup once and accidentally retrieved the password for that after I forgot the password for my actual account - tweekz. Please don't punish me. 😄

Edited April 15, 20224 yr by tweekzter

"I moving one thing will cause something else to break"

Yep, I think we all noticed that one

4 hours ago, jcomm said:

https://devsupport.flightsimulator.com/storage/attachments/2425-debug-aircraft-cfd.pdf

Great insight and explanation of what "CFD" means in the MSFS world and its interactions with other improvements such as prop physics and body rigidity simulations... so very likely as we suspected they are doing real-time modelling of a *subset* (what they call "custom") of a full CFD simulation which would take huge computational power.. assuming some subset of the https://en.wikipedia.org/wiki/Navier–Stokes_equations 

Optimum end-user benefits of this is going to come down to aircraft developers using this feature and associated debugging tools to keep iterating over their aircraft parameters since they factor in a more impactful way looks like (i.e. propellor/rotors config and parameters more important since with CFD enabled the airflow around the aircraft is influenced by the rotors as it should be):
"Rotors: When enabled, the airflow is reinjected for rotors and the air passing by the rotor will no longer move at the airspeed of the aircraft, but at the airspeed accelerated by the rotor. This is more realistic but will affect the performance of the rotors and may require an update of the rotor parameters. Note that this option requires that the propellers have been setup to use the new surface based propeller mode (see prop_mod_use_modern for more details)."

Presumably when they talk about "non-CFD aerodynamic simulation" they probably mean the current flight model physics: https://docs.flightsimulator.com/html/Samples_And_Tutorials/Primers/Flight_Model_Physics.htm

Also I wonder if the way aircraft geometry is defined and the specification of other parameters will be expanded/improved to allow more fine-grained details and a wider variety, as that will make this CFD simulation even more realistic.

 

Edited April 15, 20224 yr by lwt1971

“I’m not a pilot, but….”, “it seems like it should…”, “wow this is great, the plane feels so alive!”

Ugh, I’m not trying to discourage people from giving feedback but, hopefully Asobo has some way to vet feedback on something more that prior sim experience or forum groupthink. Even someone with a solid grasp on the science but no significant experience flying the plane is limited. I’m not so arrogant to say that I could properly evaluate the CFD implementation on the C172, but with 1000+ hours flying/instructing in that plane, I wish I had the time to participate in the beta 😞

Initial feedback on the latest SU9 tweaks for the new CFD modelling on the C172 seems to be fairly positive at https://forums.flightsimulator.com/t/feedback-for-cfd-simulation-on-c172/508234/318 ... hopefully it's refined even more and other default and 3rd party aircraft can start to make use of it soon.

In addition to the earlier doc jcomm mentioned, there's also more prelim docs on the CFD parameters in the .cfg file: https://devsupport.flightsimulator.com/questions/6977/su9-question-about-cfd.html ... Asobo's custom version of the Navier-Stokes differential equations being solved is currently set at a  20 voxel cube sampling resolution but it can be set higher, something to try for those with powerful CPUs :)

Edited April 16, 20224 yr by lwt1971

20 minutes ago, lwt1971 said:

Initial feedback on the latest SU9 tweaks for the new CFD modelling on the C172 seems to be fairly positive at https://forums.flightsimulator.com/t/feedback-for-cfd-simulation-on-c172/508234/318 ... hopefully it's refined even more and other default and 3rd party aircraft can start to make use of it soon.

In addition to the earlier doc jcomm mentioned, there's also more prelim docs on the CFD parameters in the .cfg file: https://devsupport.flightsimulator.com/questions/6977/su9-question-about-cfd.html ... Asobo's custom version of the Navier-Stokes differential equations being solved is currently set at a  20 voxel cube sampling resolution but it can be set higher, something to try for those with powerful CPUs 🙂

Yeah, I think this will be the final changes to the CFD for the 172 that go into SU9.  SU9 is targeted for a release on April 26.  There isn't much time for Asobo to code the  changes, do their own internal testing, and then issue another beta patch for the public opt-in beta testers.

Edited April 16, 20224 yr by abrams_tank