基于双层代理模型的无人机气动隐身综合设计

摘要

基于双层代理模型方法展开飞翼布局无人机气动隐身综合设计方法研究.DSM模型集成了插值型代理模型和回归型代理模型的优点,在同样样本的条件下,具有更高的预测精度.文中分别以Navier-Stokes方程数值求解方法和稀疏矩阵方法分析无人机气动、隐身特性,并构造了两种分析方法的DSM模型,验证了DSM模型特性.文章最后基于自适应代理模型优化设计策略,结合多目标遗传算法将DSM模型用于无人机气动隐身综合设计,使其在满足一定约束的前提下最小化全机阻力和头向雷达反射截面,取得了令人满意的设计结果,验证了文中方法的实用性.%In this paper,an aerodynamic and stealth synthesis design optimization method is presented based on double-stage metamodel(DSM).DSM integrates advantages of both interpolation and regression metamodel and expresses higher fidelity than traditional metamodels.Navier-Stokes equations numerical solution method and sparse matrix method are applied to analysis aerodynamic and stealth performance parameters.And then DSM models of both simulation models are constructed and the characteristics of DSM are validated.In the end,based on adaptive metamodel-based design optimization strategy,the fly-wing UAV aerodynamic and stealth synthesis optimal design using DSM and multi-objective genetic algorithm is presented.The design target is to minimize the drag coefficient and heading radar cross section while meeting performance and geometric constraints.Satisfying results are obtained and practicality of the design method is verified.