Re: Flight Models for newbee's

> Frankly I have looked a jsbsim and it uses a lot of envelope
> preprocessing. It isn't so much computing the aircraft as doing lookups
> under a multitude of circumstances and giving you the response for the
> one you are in.

"Envelope pre-processing"?

Let me expand a little bit on this. JSBSim uses the coefficient buildup
method of calculating forces and moments. Within the aerodynamics section of
the JSBSim aircraft definition for a given aircraft, lookup tables,
mathematical functions, or real values can be used to define aerodynamic
forces and moments. I haven't tried this, yet, but I suspect it may even be
possible to define functions that approximate the X-Plane approach, where
individual contributions due to various "parts" are computed and summed. The
function capability in JSBSim is very powerful, but has not really been
exercised to the best of its capabilities, yet.

The approach that the "big iron" sims take is very similar (if not exactly
similar) to the approach taken by JSBSim. I've worked on several large
simulation projects over the past twenty years. Two that stand out in my
mind are the USAF F-16 simulator and the space shuttle (and spacecraft)
training and engineering simulators I've enjoyed working on, supporting our
manned space program. Those simulators (and other large simulators of which
I am aware) also use the coefficient buildup method to do aerodynamic
modeling. At this time it's just the best way to get accurate and
predictable modeling. The major difference between JSBSim models and models
from these simulators is the data that drives them. The space shuttle
simulator is modeled with lookup tables for which the data fits in two large
printed volumes, each is three inches thick. The lookup tables do
interpolation using smoothing functions.

> Xplane actually breaks down each airfoil piece and computes the forces
> on it as if it were a separate aerodynamic body

... which is a big approximation that breaks down in some circumstances...

> and then sums them all while doing a separate fuselage
> estimation and some downdraft calcualtion attempts.

"Estimation" is the key word here. When flight test data is available, or
where the time is taken to derive and compute aero coefficients and
derivatives, the coefficient buildup method theoretically results in better
performance matching in the envelope where most flying is done, in the
"normal" regime. But, the time needs to be taken in obtaining and entering
the data accurately.

> Sounds like larcsim just does very approximate lookups.

LaRCSim does *exact* lookups, but modeling is only accurate for where data
is provided.

> None of these are good for planes that haven't been stuck in a wind
> tunnel and "precomputed"

I'm not quite sure what you mean here, but this is exactly what JSBSim is
good for: if you have actual flight test or wind tunnel data, you can use
that to drive JSBSim. That's one thing that wind tunnel data is used for.

Jon