差分进化智能算法在高旋弹气动辨识中的应用

Application of Differential Evolution Intelligent Algorithm in High-vibration Aerodynamic Identification

针对传统方法对初值选取敏感问题,将差分进化算法(DE)应用到气动参数辨识中,取代传统极大似然方法中的牛顿迭代梯度优化方法,对高旋弹的气动参数进行全弹道辨识。文中首先对可辨识性问题进行分析。通过仿真计算,证明该方法与传统极大似然方法相比辨识结果更加稳定,并通过反算弹道的方法证明了该算法收敛速度快,全局寻优能力强,具有较高的精度,并且解决了传统方法对初值敏感问题。

Considering the sensitivity problem of traditional methods for initial value selection,a Differential Evolution Algorithm(DE)that is applied to the aerodynamic parameter identification and to replace Newton iterative gradient optimization in traditional maximum likelihood method,to identification aerodynamic parameters of high-speed spinning projectiles.And the canalizations of the aerodynamic identification parameters are given for high-speed spinning projectiles with sensitivity calculation.Through simulation calculation prove that the identification result of this method is more stable.And use the method of back calculating trajectory show the effective performance,fast convergence speed,strong global optimization ability,high accuracy and low-sensitive to initial values with the proposed method.