考虑飞行条件变化和制造误差的气动优化方法

提出一种渐近全局代理模型方法以提高稳健优化中的代理模型的精度.基本思路是连续成批地在样本空间的全局和局部均加入新样本点,不断提高代理模型的全局拟合精度.将基于渐近全局代理模型稳健优化方法应用于高亚声速翼型设计,结果表明不仅目标值阻力系数具有稳健性,对飞行条件的小幅度变化和制造误差不敏感,而且力矩系数的约束也具有稳健性.

Aerodynamic optimization under operation variations and manufacture errors

To improve the accuracy of the surrogate model, a sequential global surrogate model is proposed in this paper. The global accuracy of the surrogate models is continually improved by a strategy of adding global and local sampling points in the design space sequentially. The robust optimization based on the sequential global surrogate model has been applied to transonic airfoil design under the uncertainties, including operation variations and manufacture errors. The results show that the objective drag coefficient is insensitive to the uncertainties, and the constraint of pitch moment coefficient is still satisfied under the uncertainties.