Improved Dammage Modell (more parts)

[Ack-Ack]

Originally posted by hubsonfire
I saw someone in game express frustration at the damaged flap effect. The game shows the flap being shot completely off, but the flap's effect remains if the flap was deployed. I'll agree that this doesn't make a great deal of sense.

 

Yeah, it is curious how the blown off flap can still generate lift for its respective wing.


ack-ack

[Octavius]

^^^

If it became visually "stuck" or damaged, then I'd be content.  Never could fully appreciate the current visual of invisible "shot off" flaps behaving in their last extended position.  

I'd like to see some variation in the flap damage as well.  If a flap is destroyed completely, then let it visually represent its destroyed state.  If it's damaged and stuck, let it visually remain stuck and providing lift/drag.  Basically I would appreciate five possible states:  1)  Functional (undamaged), 2) Destroyed completely (gone), 3) partially destroyed, 4) jammed/stuck, and 5) a combination partially-destroyed/stuck.  Which state the flap finds itself in would depend on a few factors... location of damage is a biggy, round type, and a little randomness (much like an oil hit's duration-of-leak is random now).

Let us not stop at flaps (per wilbus' post).


BUT.  Based on the current way lift is calculated (4 points on each wing for example? ... would need a detailed explanation first! or the following is bunk)

Is it the case where a coding overhaul of the lift and vector calculations (large chunk of FM) would be in order?   For the individual damaged part that is.  Say for example half an aileron's working surface remains, or three quarters of a flap still functioning... a new lift calculation point for each damagable part... Any way to simplify the amount of data required for processing?  Would I be correct in that description?

[beauzo]

AH is either a single lift component calc per wing, or many, like that of X-Plane.  Either way, aileron effect, for example, is just added to the appropriate lift component (non-dimensional).  Aileron effect is either positive, down aileron, or negative, up aileron (depending on your coordinate system).

The method I suggested above would work if the FM is either single or multiple lift component calculations.  

However, if the FM lift component is the multiple kind, then individual component effect is important.  For example, missing a wingtip-end of an aileron is going to lower the possible moment on the roll axis.  Loosing an inboard part of the aileron will have less effect on total possible roll moment.

I'll try to keep the physics minimal... ;o)

[Octavius]

Dun worry about the physics, I follow everything (phys major) .  

Having read your suggestion, it's definitely a possibility.  It would offer the desired effect (albeit with a random factor tossed in, similar to what I had tossed out there).  But I don't think, in AH's case, that what we're looking for is as simple as modifying a few components based on the operational status of an aileron.  I think this was discussed before (maybe deja vu).

An FM 'overhaul' not too long ago saw an increase in the number of components per wing.  Used to be two, I *think* it's four now.  

Trying to do a quick search now... I know for a fact AH does not use one component per wing.

And AH is not like X-plane.  Xplane goes through a dynamic lift calculation of a sort, correct?

[Octavius]

Found it:  

We have been going through the flight engine over the past couple of weeks. The changes have been piling up with one change leading to others. So far, we’re pretty pleased with the results but it does require extensive work. The changes we’ve made result in a more accurate model and we’ve also implemented better methods and tools to verify accuracy.

One of the major changes we’ve made is how we model the forces on the plane. We wanted to increase the number of force points by about an order of magnitude or even more if necessary. For example, prior to this the wing was split up in large chunks with the applicable forces applied to each chunk. In level steady flight this is fine, but it shows its limitations when you get outside of it. Now we have it split up into a lot of small pieces. This allows us to closely replicate the stall progression characteristics of the different planes. It’s also led us to some oversights in the model. An example of this is the effects on the propeller vortex on the wings. We’re modeling the forces of the vortex and its effect, but we weren’t applying it exactly where it needed to be applied, thereby creating an inaccurate force moment.  

Per Pyro from this thread

(edit, holy cow, that was two years ago, not too recent )

[beauzo]

Unfortunately, I don't know exactly how AH has their FM setup.  They could be calculating Cl, Cd, etc., using lookup tables for all I know.  (When I was working on FMs for simulators, we would use CFD and windtunnel test results as lookup tables for various coefficients in the FM (using Mach, AoA, etc., as the dependents).)

X-Plane is indeed doing some deeper real-time computations to get coefficients, but I doubt other parts of the FM is considerably different to AH's.  IIRC, X-Plane computes coefficents using many airfoil sections along a wingspan.  In effect, X-Plane is generating the lookup table data on the fly.

Despite all that, what I mentioned in my first post above should work quite well without adding anything major to the current FM.  Instead, the system could be inserted between the lift force computation and the application to the kinetic body.