3 hours ago, jcomm said:

Probably one of the best opinions I have seen about MFS CFD...

MSFS - Another Aerodynamics Discussion - YouTube

"CFD is putting the cart before the horse."

His DC3 video, also linked in this thread, is also good.

He complains about the fact that Asobo doesn't use actual airframe geometry. Given that Asobo is not calculating forces on the elements, there's no reason to put the actual geometry in there. All they need are a few flat surfaces to trigger the 6 degrees of motion (translation, rotation). The final aerodynamic behaviors for a v-tail would have to be rolled up in the FSX curves. Same illness as with roll-yaw coupling.

His discussion about lift distribution is also another point against the idea that the elements calculate forces. If those are forces, you wouldn't need an Oswald efficiency factor. Oswald is indeed an option provided by Asobo.

 

6 hours ago, abrams_tank said:

CFD lines for the helicopters

I put helicopter CFD up there with the weather. It's actually a GOOD idea that I'm a bit jealous of. Reason being that airframe geometry is secondary concern. The single biggest factor in flight dynamics is the rotor disks. This low res CFD can actually be used in this case. No messy geometry concerns, just a big, flat spinny thing to solve for with very little to worry about in the middle of the air mesh.

10 hours ago, blingthinger said:

According to this pic from the SDK

Before talking about math details, you need to understand the big picture. Why do I still have to lecture you about what section in the text is about what?

The picture is located beneath the caption, which says: Now to explain how these normalization coefficients are computed for each of the 20 chosen aircraft control configurations during the initialization stage...

So you are mixing this coefficient calculation, which is done once at initialization with the runtime calculations, where the forces come only into the picture.

It is very clear from the essay, from the other SDK page from the videos and also when thinking logically (coefficients at runtime are multiplied with what?) that per surface a force is calculated and that Blinthinger for the hundreth time is just wrong.

I also don't know, how exactly they incorporate the moment per surface. This is not explained to the lowest detail in the essay. That does not nullify the part that is described clearly. If you spot a contradiction in a part which is hardly explained in the essay, it is simply mean-spirited if you conclude that Asobo lied about the other part, which is explained clearly, and explained in several sources.

3 hours ago, mrueedi said:

this coefficient calculation, which is done once at initialization

You are correct. The normalization coefficients are calculated once, at initialization.

You are also claiming that the force and moment coefficients are also used at initialization, and that they then turn into just forces (not coefficients), correct? That's another question you keep avoiding.

 

3 hours ago, mrueedi said:

I also don't know, how exactly they incorporate the moment per surface.

This is not complicated. It's a fundamental physics equation.

Moments are derived from forces (moment = force x distance). You do not solve for them simultaneously. Claiming to solve simultaneously (as the SDK very clearly states that they do in fact do) makes no sense. Unless the FM solution loop, as shown in their image below is NOT solving forces.

 

3 hours ago, mrueedi said:

simply mean-spirited

I'm sorry that you interpret fundamental physics as "mean-spirited". But lies are lies. They are saying one thing but actually doing another. I don't know how else to interpret it after so many years of marketing campaigns and press releases.

 

3 hours ago, mrueedi said:

per surface a force is calculated

There is another reason you don't want to claim this. If they are somehow defying fundamentals (solving moments and forces simultaneously), the planform geometry is where they are getting the forces from. For every single airframe that will ever go into the sim: 

4 hours ago, mrueedi said:

I also don't know, how exactly they incorporate the moment per surface.

AHHH! I can see a scenario in which I am wrong!! If so, I would indeed apologize for accusing Asobo of lying!

Here's a thought: they are using the balsa glider fuselage to calculate forces. The forces on all of the elements are based on flat plates (the balsa glider). Not airfoils and not actual fuselage geometry.

Because of that, they use the force/moment coefficients from the FSX lookup tables to squish and squeeze these forces such that the final resulting lift/drag/ rotation moments satisfy the single-point-force conservation equations.

They also don't actually model the flap and spoiler deployments. That's where the normalization coefficients enter. There's no other way to simulate a dynamic/static response from those controls.

Not sure how they would account for the aileron and elevator movements. Also likely flat plates (not airfoils). Those forces will also get scaled towards the FSX tables.

Would that make sense?