论文部分内容阅读
提出了基于代理模型的两步优化方法,用于翼型在黏性流场中气动外形的优化设计。第一步优化使用基于代理模型的遗传算法(GA)获得全局最优解的大致范围,以本征正交分解(POD)方法作为第一步优化中气动力计算的代理模型方法,降低遗传算法的计算量,并对其采样解的生成方法进行改进,提高了计算精度;第二步优化使用基于Navier-Stokes方程的最速下降法(SDA),既改善了第一步优化结果,又修正了代理模型所引入的误差。针对传统型函数方法在翼型后缘表达不足的缺陷进行改进,提高了型函数对翼型的表达精度。不同外形翼型的反设计结果以及不同来流状态下的优化设计结果表明,本文提出的两步优化方法是一种高效且具有实用价值的优化方法。
A two-step optimization method based on agent model was proposed to optimize the aerodynamic shape of the airfoil in the viscous flow field. The first step optimizes the approximate range of the global optimal solution using genetic algorithm (GA) based on the agent model (GA), and uses the method of intrinsic orthogonal decomposition (POD) as the proxy model method for the first step optimization of aerodynamic forces to reduce the genetic algorithm The second step is to optimize the steepest descent method (SDA) based on the Navier-Stokes equation, which not only improves the first step optimization result but also modifies The error introduced by the proxy model. Aiming at the defect that the traditional function method is not expressed enough in the trailing edge of the airfoil, the expression accuracy of the type function on the airfoil is improved. The anti-design results of different profile airfoils and the optimization design results under different flow regimes show that the proposed two-step optimization method is an efficient and practical optimization method.