On Mon, 17 Nov 2003 11:33:35 -0500 David Megginson <[EMAIL PROTECTED]> wrote:
Again, I'm wondering if this is an aerodynamic problem (aside from the bouncing-around-sitting-still thing). Because of its lifting
surfaces, a plane is certainly more vulnerable to the wind than a car,
even when it is sitting on the ground; however, the coefficients we
use in JSBSim are designed to deal with a relative wind near or above
the plane's stall speed, coming straight onto the nose +- about 20 deg
vertically or horizontally. They probably do a pretty crappy job of
modelling (say) a 15 kt gust hitting the plane from 90 deg when it's
sitting on the ground. I expect that the same applies to the
assumptions made by YASim's solver.
Yes, I have considered that. In fact, the realization was the driving idea for the creation of the modified qbar values, qbarUV and qbarUW, which should be used in aircraft aero coefficient definitions instead of simply qbar. For instance, say we are sitting on the runway and have an 80 degree crosswind. We had better be darn sure that CL_alpha is not calculated based on the qbar from straight ahead, because if we have a high (80 or 90 degree) beta, there won't be any appreciable lift - because there would be no wind velocity (thus qbar) over the lifting surface of the wing, parallel to the chord. Also, it emphasizes the fact that our Cn_beta curves ought to cover the range from -180 to 0 to +180. Then, a high crosswind won't adversely (and incorrectly) affect the forces and moments that the gear will be expected to counter.
Good observation, and I am aware of the phenomena you describe. As I have laid out, there is a way to get around that by properly defining aircraft aero coefficients throughout the entire range of the flight envelope - EVEN if the off-nominal portion (high values) is a guess, it will be better than not defining it.
I think there are also ways to hold the aircraft steady on the ground, change the winds, and record the stabilty derivative (and/or force/moment) that results, in order to "map" the performance and accuracy of the FDM on the runway. This will help for debugging. This involves the property browser and setting of the "dt" to 0.0 during the process.
Jon
_______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
