Abstract : A general, neural network based algorithm has been developed and applied to the problem of helicopter rotor smoothing. This approach provides non-parametric mappings between the spaces of rotor adjustment and vibration measurements, which are derived directly from empirical data, and permits to relax the usually used linearity assumption. Additionally, the rotor smoothing solutions are optimized to minimize not only the predicted vibration levels and track split but also the number of required adjustment moves. The neural network rotor smoothing system is a part of the VMEP (Vibration Management Enhancement Program) PC Ground Base Station program and has been successfully applied to the AH-64 Apache and UH-60 Blackhawk helicopters. Applications to other types of helicopters are under development.
[1]
Heekuck Oh,et al.
Neural Networks for Pattern Recognition
,
1993,
Adv. Comput..
[2]
John Berry,et al.
Automated Helicopter Vibration Diagnostics for the US Army and National Guard
,
2001
.
[3]
D. Wroblewski,et al.
Neural network system for helicopter rotor smoothing
,
2000,
2000 IEEE Aerospace Conference. Proceedings (Cat. No.00TH8484).
[4]
Aviv Rosen,et al.
MATHEMATICAL MODELLING OF A HELICOPTER ROTOR TRACK AND BALANCE: THEORY
,
1997
.
[5]
Thomas P. Goodman,et al.
A Least-Squares Method for Computing Balance Corrections
,
1964
.