Re: [Jsbsim-devel] DATCOM+ released

[Ron F. <ron...@at...>]

----- Original Message ----- 
From: "Jon Berndt" <js...@ha...>
To: <jsb...@li...>
Sent: Wednesday, July 28, 2004 8:52 AM
Subject: RE: [Jsbsim-devel] DATCOM+ released


> >     There is an Air Force report on measured SD's for the T-37B,  up to
> > about 10 deg alpha.  The T-37 is slow enough so compressibility
shouldn't be
> > a factor.

> Compressibility is not necessarily a problem with DATCOM. Specifically, I
believe the
> transonic regime can be problematic, though. However, if you can supply
some "seed"
> values, DATCOM will fair the data and give a better output. This is all
from what I've
> read - I haven't modeled anything supersonic using DATCOM, yet.

    There is a pdf report at Mason's univ Virginia site that compares some
SD's between DATCOM, Wind Tunnel, and the LDstab results.  However, I can't
find it right now.  I think the LDstab program gave better Lateral
Derivatives.  However, that Fortran program had to be 'adjusted' to match a
747-100, and it was suggested that reasonable results would only be obtained
with similar AC.  ;)

    LDstabdoc.pdf mentions DATCOM and includes the .f source code and a list
of SD's it gives for the 747-100.

 ConfigAeroHiAlphaNotes.pdf is a useful document for one of Mason's classes.
It has info on pitching moment, Rudder Moment, etc. curves necessary to keep
an AC (typically a fighter) controllable at abnormal attitudes.

 There is also ConfigAeroHiLift.pd: typical curves on Lift augmentation
devices, lots of references.


>I'm excited about DATCOM+. One it gets put through it's paces (and once we
find a way to
> produce rudder coefficients, which DATCOM does not do), then I think it
might be possible
> to write a script or something that could somehow have Aeromatic and
DATCOM+ cooperate.
> The ultimate goal would be to craft a tool that made it as easy as
possible to specify the
> relevant geometric, mass, and planform data about an aircraft, have
DATCOM+ produce the
> aerodynamic JSBSim output, have Aeromatic produce what it does, and then
end up with a
> file that defines a high fidelity aircraft flight model.
>
> Jon

     I skimmed Bill's notes on what SD's DATCOM[+] generates.

    In MSFS, I generally set control moments by experiment.  I think the
published values would be for small deflections; but they drop in effect at
maximum deflections.  Maximum control moments are important in determining
Roll Rate, Pitch Rate and ability to stall and AC (and, recover with down
elevator).  And, finally how much Beta the rudder will create.  I generally
set Cn_dr to get 15 to 20 deg Beta.

    I set Cl_da to get an appropriate p_max.  Easy to calculate, but since
p_max also varies directly with V, it's better to adjust  experimentally.
Of course, Cl_da/Cl_p determines Roll Helix Angle, and one can estimate that
from the AC's roll performance.

    While I keep Cm_de/Cl_de  constant; based on the normalized tail moment
arm distance. Tweak Cm_de to the final value experimentally, but always keep
that ratio constant.

    I figured the same thing applies to Cn_dr and CY_dr.  Only the moment
arm is normalized to b rather than c_bar.

    I have noted that elevator vs horiz_stab lift and moments are typically
in a 1:3 ratio in MSFS.  That suggests the elevator is 1/3 of the H Stab
area.   Higher than the typical physical ratio.

    However, the total H. Stab needs to be considered as a unit.  Curves in
McCormick, etc. suggest the elevators have a greater effect than simple
consideration of their relative area gives.

    The Vertical Fin and Rudder would be similar.  Though,
d(epsilon)/d(alpha) typically reduces the horizontal tail effects more than
the corresponding effect for the vertical stab.

    Further, the vertical stab is generally asymmetrical so it adds some
roll moment, not just yaw. Regardless, CY_Beta is generally the most
important Side Force, I see I set CY_dr to only 1/6 the former in my 727.
Based on keeping Cn_dr/CY_dr at about 0.4.

    Complicating matters is how q effects the control moments.  Either due
to limits on pilot strength, purposeful limits on hydraulic pressure, and
control system scheduling.

    On top of that, all MSFS SD's can be modified by Mach Number.  However,
the effect is limited in jet transports.  Even Cd_m (Cd_p(M)) is small at
typical cruise Mach,  It's the Lift Related component that mostly determines
increase in CD and the optimum flight level at a given weight.  Of course
Cm(Mach) pitches the nose down in the transonic region, but some curves I've
seen have a fine structure that would be impossible to calculate without
CFD.

    -------------------------------------------------

    To sum up Control Moments and associated 'lifts': I've found  2/3 the
published value appears to be 'about right' as far as sensitivity and
available control goes. That suggests lower than calculated values may also
be desirable. Cn_r should generally be able to create about +/- 15 deg of
Slip.  This is required for X-Wind landings!    Multi engine AC have to have
enough rudder control for engine out conditions.

    While the control surface forces are pretty much dependent on normalized
moment arms.  Though, just what point one measures them from  isn't exactly
the CG or aerodynamic center.

     I've verified many of the MSFS SD's by calculation, so I'd imagine they
would apply to simulators with similar  flight model approaches.

    Ron