ls-svm modeling method for nonlinear unsteady aerodynamic characteristics of missiles

Abstract

The invention relates to a LS-SVM modeling method for nonlinear unsteady aerodynamic characteristics of a missile. The method decomposes the missile aerodynamic force into the sum of static aerodynamic force and aerodynamic force increments generated by dynamic motion, and adopts the LS-SVM modeling method to establish dynamic aerodynamic force. Dynamic incremental model, in which the motion history is approximately described with finite sampling points to reflect the influence of the motion history on aerodynamic characteristics; the linear equations of LS-SVM are solved by singular value decomposition method, and the model is trained; the training-verification method is used to determine Penalty factor and kernel width to improve the generalization performance of the LS‑SVM model. The purpose is to provide an aerodynamic modeling method for missile maneuvering at high angle of attack based on machine learning.

Description

technical field [0001] The invention belongs to the technical field of aerodynamics, and in particular relates to a LS-SVM modeling method for missile nonlinear unsteady aerodynamic characteristics. Background technique [0002] High angle of attack maneuverability is an important design indicator of modern tactical missiles. Especially for air-launched missiles, the use of large angle of attack maneuvers to achieve over-the-shoulder launch can greatly shorten the combat response time, improve the survival probability of the carrier aircraft and the missile combat effectiveness. Regardless of whether it is a traditional over-the-shoulder launch or a new type of self-flipping over-shoulder launch, the missile will experience a large angle of attack and a stall state, and the aerodynamic force exhibits highly unsteady and strong nonlinear characteristics. At this time, the aerodynamic derivative model based on the quasi-steady assumption is no longer applicable, and a nonline...

AI Technical Summary

Problems solved by technology

However, in this literature, the Cholesky decomposition method is used to solve the linear equations in LS-SVM. When the sample size is large, there is also the problem that the coefficient matrix A has a singular value and the training algorithm cannot be performed.

Images