2 minutes ago, SAS443 said:

You need every "friend" you can get for ground effect, tip/root loss corrections, dynamic stall models, momentum theory and all those gentlemen with weird names and their equations.

So again, what's your final point? Nobody is questioning this. Is this really merely a foot stomp against marketing that really isn't that big of a deal?

You've clearly forgotten quite a bit or are indeed lost in the translations. As far as I can see, there is zero specification for where the force coefficients come from. In other words, one can adjust the lift, drag, and moments to account for e.g. unsteady effects or compressibility or all the other factors you list there and the whole time still be strictly adhering to BET. Seriously... still waiting for a legitimate answer to that which isn't obviously proven wrong by the very words describing it or that isn't laced with cowardly condescension.

Also: "Weird names and their equations"? Such respect for the great minds. Showing true colors there.

So this  whole pointless argument, was once a discussion, comes down too, There are two ways of doing this, BET and CFD, Both, like all theories have there plusses and minuses. Advantage of BET is it needs less computational resources, but in the end they both need tweaks to get the required results. So I think that concludes this argument.

Errors corrected for thosde who cant stand normal people.

Edited December 2, 20241 yr by mjrhealth

One can dream, at least.

That's mostly correct. Properly high-quality/resolution CFD doesn't really need tweaks.

1 hour ago, blingthinger said:

One can dream, at least.

That's mostly correct. Properly high-quality/resolution CFD doesn't really need tweaks.

The higher the quality the more data the more processing power Id assume.

2 hours ago, mjrhealth said:

So this  whole pointless argument, was once a discussion, comes down too, There are two ways of doing this, BET and CFD, Both, like all theories have there plusses and minuses. Advantage of BET is it needs less computational resources, but in the end they both need tweaks to get the required results.

It was not pointless; I didn't know how CFD and BET differed until I started digging into the topic recently, and I'm sure many others didn't know the difference either.  I thought it was a good discussion and somewhat informative.

I think I have a better understanding now though of the 2 techniques. Basic summary.
Blade Element:

• Fast computation (which is probably why it was used at the start by Austin)
• Less accuracy and limited when compared to CFD

CFD:

• Computationally intensive and requires experts
• highly accurate for both fixed wings and rotary wings, and can capture complex flow phenomena like turbulence, wake vortices, and dynamic stall.

The good news regarding the speed is that what can be accomplished in Realtime is quickly changing "thanks to the integration of machine learning(ML) and deep learning(DL)". I saw that AIRBUS was one of the clients using CFD with AI to speed up design and development. So, with the advancements, it could be promising if it can be applied to flight simulation.  For example, why build a generic flight model trying to do it all if you can use CFD/ML to come up with a highly accurate flight characteristics specific to each aircraft.

1 hour ago, brinx said:

t was not pointless;

It became pointless once it became an argument over who was right, Before that it was interesting, than the heads started butting, Must be rutting  season.

It is a very interesting subject and one I've tried to learn more about, specially after FS 2020's release, but also when Totorito released the CL 650, because he did use CFD to model some of the characteristics of the real aircraft trying to approach the XP model the closest possible to RW numbers.

But both in XP12 and in FS 2024 BeT and CFD fail miserably in even basic aspects.

To give you just two examples you can take the reaction to relative wind of prop blades of a failed engine whose prop has been feathered. In FS 2024 you can sideslip as much as you can and yet the rotor will not start turning, assuming there was no mechanical damage to it's engine, and specially on a free-running turboprop equipped for instance with the PT-6.

I am aware that ASOBO has fine tuned their model and I did already notice changes, for the better feel, in rotary wing and not only, but still it's far from perfect.

In XP12 you also can design an aircraft with exact rw data for the various elements - wings, fuselage, aerodynamic surfaces, ... and yet it can fail to reproduce aspects of it's RW counterpart such as pitching moment due to flap retraction / deployment, that have to be baked in order to come closer to reality.

Hidden surfaces are sometimes used to create drag and / or lift in order to fine tune some performance numbers / characteristics, and some more sophisticated developers have asked for dynamic wing profile to overcome the limitations of the flight model...

Then there are bugs and limitations. Both FS 2024 and XP21 have them, some very evident like the new "negative torque roll" in XP12 not yet addressed by LR... or the kite-like behaviour of all aircraft in FS 2024 when flying through unstable air...

There is no winner IMHO.

Edited December 2, 20241 yr by jcomm

4 hours ago, mjrhealth said:

The higher the quality the more data the more processing power Id assume.

Exactly. 

3 hours ago, brinx said:

I saw that AIRBUS was one of the clients using CFD with AI to speed up design and development

It absolutely can and does. But going back your car example, you cannot predict wildly different geometries that were not in the training set. You cannot train the AI model on a 1000 Cessna-like airframes and 1000 Airbus-like airframes and then expect it to accurately predict a Gripen. That's not how it works. Or, say the Cessna database doesn't have wing struts. If you toss a design at it that does have struts, it will get confused.

That car prediction was made with a model trained from 1000 sedan-like shapes. We can tolerate some goofiness in images of puppies but ask any programmer about the computer code coming from chatgpt and it doesn't take long to find the hallucinations. A flight model would have no tolerance for this. And to train a model sufficiently to eliminate all of those quirky results would take so much CFD effort that at the end, you might as well build an excellent table model and have certainty that you're precise.

Asobo in 2024 seems to have put the batteries something that was not difficult because the model was terrible, that yes this will depend as always on the developers.

This benefits us users who enjoy this hobby and have no problem with any software.

My simulator of choice is still X-Plane while we have gems like Hot Start, Torquesim, Toliss, FF making these addon of honor but I also have 2020 with Leonardo, A2A, I have installed BMS Falcon and of course DCS.

I'm not talking about other simulators like run8 or simrail what I lack is time for everything !!

Out of curiosity, I'd like to see a scientific comparison between X-Plane's BET-based flight model and MSFS' CFD-based flight model for a common aircraft like the C172, using the exact same target data (POH or flight test) and starting from scratch, covering the following aspects:

• Largest possible performance envelope that can be attained with a 99 % accuracy compared to real world values (flight test and POH). This also tracks the used work flow (including workarounds for flight model deficiencies) and the work hours spent to get there. Done by two independent and competitive teams that are firmly rooted in and very familiar with their simulator. Obviously requires physical presence to avoid cheating during work time tracking.
• Measuring CPU time required for a flight model loop.
• Review by one or more licensed and very experienced C172 instructors, also covering subjective elements like the handling "feel" (on the exact same hardware and in the exact same simulator flight conditions).

This should point out each approach's strength and weakness pretty well.

 

On 12/1/2024 at 5:09 PM, blingthinger said:

🙂  🙂 Let them have some hope!

Never! 🙃

2 hours ago, Bjoern said:

Review by one or more licensed and very experienced C172 instructors, also covering subjective elements like the handling "feel"

Absolutely not. Or at least not by itself. A 6DOF accelerometer and GPS is nearly dirt cheap these days. I am very very surprised that of the many people floating these and other sim forums, with all manner of official-looking license lettering and certification in their account Signature below, no one has once done a detailed data acquisition study with any airframe. Put it on github: data and all the details needed to recreate it in the sim. How has this not happened yet? Not even with a good-sized R/C aircraft even? Maybe it does exist and has been lost in the wind?

33 minutes ago, blingthinger said:

Absolutely not. Or at least not by itself. A 6DOF accelerometer and GPS is nearly dirt cheap these days. I am very very surprised that of the many people floating these and other sim forums, with all manner of official-looking license lettering and certification in their account Signature below, no one has once done a detailed data acquisition study with any airframe. Put it on github: data and all the details needed to recreate it in the sim. How has this not happened yet? Not even with a good-sized R/C aircraft even? Maybe it does exist and has been lost in the wind?

I once had a very funny conversation with such a guy claiming that he had those data and claimed that XP‘s flighmodel was bad in many aspects and MSFS (2020) better after measuring them against those data. He said this was objective because of the measured data and so could be seen as „facts“. He also said he was a test-pilot and could judge it better than many professional pilots (who btw. all - as I told him - came to the conclusion than XP‘s flightmodel was the best for airliners after comparing both sims). As I asked him to show these data he had allegedly measured, he argued with the usual „I don‘t have to show it to you because I am not here to educate you“. You can guess in which forum on avsim that happened…

Edited December 2, 20241 yr by Franz007

3 hours ago, blingthinger said:

Absolutely not. Or at least not by itself. A 6DOF accelerometer and GPS is nearly dirt cheap these days. I am very very surprised that of the many people floating these and other sim forums, with all manner of official-looking license lettering and certification in their account Signature below, no one has once done a detailed data acquisition study with any airframe. Put it on github: data and all the details needed to recreate it in the sim. How has this not happened yet? Not even with a good-sized R/C aircraft even? Maybe it does exist and has been lost in the wind?

Laminar does it fairly regularly as needed, one of their biggest updates was for the helicopters in XP11.10, the others were completed a long time ago.

https://www.x-plane.com/kb/x-plane-11-10-release-notes/#Bug_Fixes

Quote

More tuning of the propwash (and thus hover power requirement) and vortex ring state and effective translational lift based on momentum conservation and actual helicopter performance data. Definitely more accuracy there now, based on a lot of testing against very precise Bell-206 performance data in ground effect, out of ground affect, and at various flight speeds as well, and also in various decent rates including settling with power and power-off vortex-ring-state.

It is however a constant battle with new updates to fix one thing breaking others, in fact Austin wont make any changes to the flight model unless you prove its wrong with real world data, and man has he made a lot of changes over the years.

AFIAK getting a simulator configuration FFA approved requires proving it matches the aircraft getting certified, hence xplane having all the tools embedded in the developer menu to do so out of the box.

Edited December 3, 20241 yr by mSparks

1 hour ago, Franz007 said:

He said this was objective because of the measured data and so could be seen as „facts“.

hahaha Yes, the 'trust me because I am objective because I said so' line. Precisely why IRL, peer review of data like these is a requirement before acceptance by the community.

6 hours ago, blingthinger said:

I am very very surprised that of the many people floating these and other sim forums,

What happened to your own quest for data some years ago. The user singlecoil -instructor on the C172 - even offered you a free ride to go up with him. (Not many would do this, a golden opportunity indeed)

Did that just evaporate?