Modern FM...

[jcomm 6,096]

The ASOBO idea for the "Modern FM" has been around since the release of FS 2020 / MFS, but it was not until SU 10 beta, which I am testing since it was made available, that some new entries in:

• flight_model.cfg
• engines.cfg

such as:

modern_fm_only

and all of the "reinject" directives were made available.

The documentation also mentions that a good deal of Legacy parameters are no longer required. You can read about that here:

Flight Model Config Definition (flightsimulator.com)

This new approach makes use of a "CFD-based" model and derives most of the legacy parameters from the [AIRPLANE_GEOMETRY] section in "flight_model.cfg". It actually looks to me closer in some aspects to FLIGHT GEAR's YASim than to other approaches available and well known...

ASOBO also made available a nifty ".xlsx" file that any one developing an aircraft model for MFS can use, and described in the documentations as:

NOTE: To help with the configuration of the Flight Model (and the Engine Config Definition) we have included an *.xlsx file with the documentation that can be used to generate the required values for many of the parameters based on a small number of inputs (these inputs are marked in blue in the file): PlanePerformance.xlsx

Altogether these new features really make me wonder how a new "geometry-based" model of an aircraft will work in MFS. It looks to me like a really promising approach / idea, and although I am sure there's still a lot to do in terms of fine tunning and giving here & there further control do aircraft developers with very specific modelling needs, the end results will be rewarding, the default C172 being already a nice proof of concept, handling pretty satisfactorily IMO.

Wonder if any of you know of freeware / payware aircraft already built based *only* on the new "modern FM" ? If you do, please use this thread to link them and / or simply express your ideas about this new approach in MFS.

Edited August 12, 2022 by jcomm

[lwt1971 7,709]

On 8/12/2022 at 6:39 AM, jcomm said:

Altogether these new features really make me wonder how a new "geometry-based" model of an aircraft will work in MFS. It looks to me like a really promising approach / idea, and although I am sure there's still a lot to do in terms of fine tunning and giving here & there further control do aircraft developers with very specific modelling needs, the end results will be rewarding, the default C172 being already a nice proof of concept, handling pretty satisfactorily IMO.

Wonder if any of you know of freeware / payware aircraft already built based *only* on the new "modern FM" ? If you do, please use this thread to link them and / or simply express your ideas about this new approach in MFS.

The Milviz C310 and Sting S4 certainly make use of the modern FM along with the prop physics and CFD advancements that came in SU8 and SU9. Not sure if they're *exclusively* on the new FM architecture but only their devs can confirm.

 

Edited August 16, 2022 by lwt1971

[charliearon 3,108]

OK, some posts had to be removed.  Keep in mind if we have to remove an objectionable post, we also have to remove any that quote said post!

[lwt1971 7,709]

Also @jcomm if in case you haven't seen this section of the SDK docs, some good insights: 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.

It should be noted that while the formulas used in this simulation are well known and proven, the implementation in Microsoft Flight Simulator is still experimental. Also note that while we display the absolute speed, pressure, and temperature values, as well as the deviation measurements, this data may not be exact and may require improvement over coming iterations of the debug window.

..........

The final option is Reinject Data Into Flight Model. When checked, this option will affect the interaction (collision calculation) between the fluid simulation and the rigid body aircraft simulation (note that if the aircraft flight model CFG file has the CFD_ReinjectBody parameter set to 1 (TRUE) then this option will already be checked and cannot be unchecked). In theory the force applied on the fluid by the airframe is always exactly opposite from the force applied on the airframe by the fluid.

When the option is unchecked: In this mode - when the interaction between fluid and rigid body is calculated - the fluid will receive the exact body to fluid force. However, the body will not receive the exact opposite force. The force received by the body will be overwritten by the aerodynamic forces as computed by the previously existing, non-CFD, calculations (which are run in parallel). What this means is that the CFD simulation is aware of the aircraft, but the aircraft will be processed by the non-CFD aerodynamic simulation and ignore the CFD simulation. Note that in this mode, Newton's third law is not respected, and the simulation will run as if there was no CFD calculated, and the CFD can be visualized but has no effect on the airframe.

When the option is checked: In this mode the aircraft body will receive the exact opposite force to that which the fluid receives, and Newton's third law is respected. The aircraft will be flying fully on the results of the CFD simulation, and the results of the previously existing non-CFD aerodynamic simulation are discarded. This will also supply you with further options that can enable/disable reinjection for some specific surfaces in order to make the port to the CFD system easier and to help eliminate bugs and problems:

Definitely interested to find out more about the interplay (or not) between the CFD / geometry-based FMs and table-lookup-based FMs.

I'm really looking forward to how in SU11 (40th anniv edition) they'll expand this CFD processing to a 20km cocoon around the aircraft to more realistically simulate thermals, up drafts, down drafts, vortices, and other atmospheric airflows (hopefully also enable the visualizations of that airflow which they promised for SU10 but I don't believe will come till SU11)... all of this certainly helps their core support for gliders also planned for S11 which I'm sure you're looking forward to :)

 

Edited August 16, 2022 by lwt1971

[espent 1,498]

Maybe we will see this in the future.

 

[Krakin 6,021]

11 minutes ago, lwt1971 said:

The Milviz C310 and Sting S4 certainly make use of the modern FM along with the prop physics and CFD advancements that came in SU8 and SU9. Not sure if they're *exclusively* on the new FM architecture but only their devs can confirm.

 

There's an Islander that uses it as well. AvAngel did a video on it and the pitch is much more progressive, as she put it. You can tell the plane is really interacting with the wind more.

[F737MAX 5,868]

1 hour ago, espent said:

Maybe we will see this in the future.

 

No. Some recreation of wake turbulence can make our sim flying more realistic.

Glamourising real world crashes in a civil aviation flight sim isn't ok.

[espent 1,498]

Where is the glamour?

If not for the video of this (not-fatal) accident I am sure many would not believe wake turbulence could be that strong so many seconds behind the helicopter.

 

[F737MAX 5,868]

7 minutes ago, espent said:

Where is the glamour?

If not for the video of this (not-fatal) accident I am sure many would not believe wake turbulence could be that strong so many seconds behind the helicopter.

Non-fatal or not, there has been a long-standing fetish from some flight simmers to (re)create crashes - see the outrage when 'realistic damage' wasn't available for some aircraft models in previous flight sims.

IMO, there are other areas of flight modelling that need attention before work occurs on (further) wake turbulence effects.

[espent 1,498]

Dude, this is not on the same level.

This is actually good information and a good case study.

 

You should check out The AOPA Air Safety Institute's website and their videos.

Edited August 16, 2022 by espent

[Kaboki 443]

5 minutes ago, F737MAX said:

Non-fatal or not, there has been a long-standing fetish from some flight simmers to (re)create crashes - see the outrage when 'realistic damage' wasn't available for some aircraft models in previous flight sims.

IMO, there are other areas of flight modelling that need attention before work occurs on (further) wake turbulence effects.

And im the fetishist your reffering to... I dont simulate accidents that have happend or doing a World trade center run, but what I always missed from DCS is damage modeeling, like the tires pops with hard landings, or the whole undercarriage gets broken with real hard landings. So getting a wake turb like in that video would be awsome for me. And flight simming is about learning, why prevent us from learning what a wake turbulance can really do with an aircraft and not some silly wobbling that we have now. Consequences for doing something wrong should be reflected in the sim, isnt that how we learn+ the immersion for having such things...

Why is there so many snowflakes in the civil aviation sim communety? Like we watch movies of people getting slaughtered, raped, wars and god knows what, and nobody raises an eyebrow, but having damage model in a flightsim is soooooo tabu,,, I dont get it...

[Tuskin38 6,885]

They modelled wake turbulence for the Reno Races mode, but have yet to add it anywhere else.

[jcomm 6,096]

6 hours ago, Tuskin38 said:

They modelled wake turbulence for the Reno Races mode, but have yet to add it anywhere else.

It's baked / scripted as some say in Reno Air Races, not that accurate... not FM-based / derived...

Edited August 17, 2022 by jcomm

[bobcat999 8,550]

1 hour ago, jcomm said:

It's baked / scripted as some say in Reno Air Races, not that accurate... not FM-based / derived...

I wondered about this.  One thing I will say though is how many people could tell the difference?

[jcomm 6,096]

19 minutes ago, bobcat999 said:

I wondered about this.  One thing I will say though is how many people could tell the difference?

Yes, it's well disguised 🙂

They even try to be somehow accurate in placing it down the flighpath as if actually in the downwash of the leading racer.

OFC if it wasn't "scripted" it would have been ported to the simulation in general and we should have it even out of teh Reno Air Races...

Maybe one day...