"Mary Shafer" wrote in message
...
On Sun, 23 Nov 2003 12:35:11 -0000, "Simon Robbins"
wrote:

"Corey C. Jordan" wrote in
message
.. .
However, can anyone tell me what data is used to generate the
mathmatical
model coded for the previously mentioned simulators?

No, but the procedure for simulation is approached from a different
angle.
For example, a "true" simulation of an aircraft will model the
atmosphere
and airframe so that the resultant parameters determine its performance,
i.e. the exit parameters will cause (for example) a Spitfire to roll at
say
45 degrees a second, whereas a PC simulator will use a previously
defined
input parameter to dictate the aircraft rolls at that rate. It's not a
simulation of the airframe, but of its known performance. Not the same
thing. You're not going to be able to design an aircraft and predict
it's
flight envelope using MS-FS because you have to start by knowing the
performance envelope you're wanting to simulate. A militray simulator
will
take the design of the aircraft and inform you of its likely real-world
envelope.

No, no, no. Only at the very beginning do we use predictions based on
the design. Those predictions come from the wind tunnel, CFD, and
prior experience. The instant we get such data from flight we update
the simulation with the actual, not the predicted, numbers.

Only a select few simulators actually produce a correct mathematical
modeling of the aircraft's characteristics, but in many cases a simulator
has a different purpose than modeling. In the case of the F-18, there was
no mathematically correct simulator until HARV was built in '86.

Simulators are not tools used to predict the actual flight dynamics at
all. This is exactly backward. Simulators mimic the real thing. The
flight envelope is usually defined by _design_ limits, like load
factor and qbar.

During the development of flight controls the simulator is where flight
dynamics are modeled. The YF-22 deviated from this formula and it's fate is
directly tied to skipping a step. Much the same as skipping full scale
development does not necessarily eliminate the work.

Read the paper I mentioned. You'll find a very complete explanation
of how the actual data is used.

The PC games use a very generic model, not a detailed model of the
specific vehicle. They're not predicting anything, either.

The need for predictor correctors has been addressed with faster machines
these days. I did see an instance with TCAS III flight test where I thought
a rate based stabilization algorythem might have made the system work, but
it is a lot of processor overhead.