Date of Graduation

1997

Document Type

Dissertation/Thesis

Abstract

This research is focused on parameter identification for the NASA F/A-18 HARV. The HARV is currently used in the high alpha research program at the NASA Dryden Flight Research Center. In this study the longitudinal and lateral-directional stability derivatives are estimated from flight data using the Maximum Likelihood method coupled with a Newton-Raphson minimization technique. The estimated aerodynamic model describes the aircraft dynamics over a range of angle of attack from 10{dollar}\\sp\\circ{dollar} to 60{dollar}\\sp\\circ{dollar}. The mathematical model is built using the traditional static and dynamic derivative buildup. Flight data examined in this analysis are from a variety of maneuvers including large amplitude multiple doublets, optimal inputs, pilot pitch stick inputs, and pilot stick and rudder inputs. Estimated trends are discussed and compared with available wind tunnel data. The resulting aerodynamic model from this study was used to create a full 6 degree of freedom F/A-18 HARV flight software simulation supporting pilot stick, rudder, and throttle inputs. This simulation is a central tool for a second study which examines the feasibility of employing Neural Networks to act as aircraft total normal force coefficient generators. A preliminary investigation is also made into the application of this technique to actual flight data collected during the F/A-18 HARV flight testing activities. These Neural Networks are trained with the Extended Back-Propagation Algorithm to predict aircraft total normal force coefficients based upon known control surface positions and appropriate aircraft states. Overall, these studies indicate the ability of Neural Networks to successfully model nonlinear aerodynamic functions as well as generalize when presented with flight telemetry never before encountered during the training process.

Download

Share

COinS