Aerodynamic Model Identification of a Fighter Trainer using Partial Orthogonal Least Squares

摘要

This work presents a modification of the orthogonal least squares method to cope with the aerodynamic coefficients estimation problem, when two or more independent variables may be related. The proposed method retains and estimates the most important aerodynamic derivatives, based on their physical relevance, and then the remaining independent variables are classified according to their capability of residual fit reduction. The named partial orthogonal least squares technique is employed to estimate lift, drag and pitch moment dimensionless coefficients of a fighter trainer aircraft, showing adequate capability of identification and classification of independent variables.