Multiple Linear Regression for Separating Static and Dynamic Derivatives in Free Flight
摘要
Free-flight experiments were conducted for the HB-2 damping model to acquire both static and dynamic derivatives. The Hough algorithm was utilized to detect the model’s straight line and endpoints. This facilitated the tracking of angle of attack and position information from images, which contributes to the determination of aerodynamic coefficients. The experimentally acquired aerodynamic coefficients combined static and dynamic derivatives. Therefore, for a detailed analysis, these coefficients need to be separated into static and dynamic derivatives. Multiple linear regression was applied to separate the normal force and pitching moment coefficient, assuming linearity in the static derivatives and uniformity in the dynamic derivatives at low angles of attack. The aerodynamic coefficients obtained from the free-flight experiments lie on a single linear plane. As a result, the mean squared error of the separation was assessed to be approximately 0.01%, indicating that multiple linear regression can successfully separate static and dynamic derivatives. The separated static derivatives were validated against slope values from computational fluid dynamics, Missile DATCOM, and traditional balance methods conducted by Vojnotehnički Institute and Arnold Engineering Development Complex. The proposed methodology was efficient, as it acquired both static and dynamic derivatives for multiple angles of attack in a single experiment, unlike conventional methods that necessitate additional experiments to obtain dynamic derivatives.