Abstract:

Partial least squares regression(PLS) is a new multivariate data analysis method taking advantage of the Least Squares regression, principal component analysis(PCA), and canonical correlation analysis(CCA). In this paper, it is innovated and utilized for the complex multivariate nonlinear problem of modeling for the aerodynamic data of the axi symmetric flight vehicle with control surfaces. Based on the now available trigonometric mathematical model of the axi symmetric flight vehicle with control surfaces, the cross validation method is used to determine the number of the principal components and the PLS model with good generalization capability is constructed. Furthermore, the PLS method is used for two examples of modeling of normal force for an axi symmetric flight vehicle with aileron deflecting only and modeling of axial force for the axi symmetric flight vehicle with all the control surfaces deflecting, and the modeling results are compared with the results of another modeling method of orthogonal least squares regression(OLS). The results show that PLS is a feasible and effective method for aerodynamic modeling and the modeling results are generally better than OLS, and the method is of bright prospect in engineering practices.

Key words: Partial Least Squares regression, Aerodynamic modeling, Axi symmetric flight vehicle with control surfaces