使用类别形状函数的多目标气动外形优化设计
Multi-Objective Aerodynamic Shape Optimization Based on Class and Shape Transformation
-
摘要: 在飞行器的气动外形优化设计中, 参数化方法和优化算法具有十分重要的作用, 对优化的计算时间设计空间的数学特性有着深刻的影响.类别形状函数(class and shape transformation, CST)方法是一种简洁高效的参数化方法, 但对于复杂曲面很难使用统一的CST方法进行拟合.文章首先介绍了CST方法的三维实现, 分析了其数学性质, 提出了分块CST参数化方法, 保留CST方法的特性, 实现了分块曲面之间的光滑连接.针对气动外形优化设计的复杂情况, 需要根据具体的飞行任务提出设计目标, 并处理不同目标的矛盾问题.其次采用Pareto策略自动寻找最优方案集, 并基于分块CST参数化方法遗传算法和气动力快速计算方法, 对类乘波翼身组合飞行器进行了优化设计, 并改变原有问题的设定条件优化得到了全新外形.研究结果表明分块CST方法参数少, 精度高, Pareto策略处理多目标准确有效, 是气动外形优化设计中非常有用的工具.Abstract: A parameterization technique with fewer parameters and high fidelity and appropriate optimization algorithms was applied to improve the efficiency and accuracy of aerodynamic shape optimization. The class and shape transformation (CST) method is concise and efficient. Nevertheless, it is difficult to parameterize complex vehicle configurations using a unified CST method. This article presented a multi-block CST method after analyzing the mathematical description of the method, and this method joined adjacent surfaces smoothly and retained the good properties of the CST method. Pareto strategy was employed to deal with the multi-object problems in complicate flying task. A system of aerodynamic shape optimization was developed based on the multi-block CST method, genetic algorithms and hypersonic aerodynamic engineering analysis, and the aerodynamic characteristics of a quasi-waverider wing-body vehicle were improved significantly after optimization. A novel configuration was also created using the optimization system. The optimization results indicate that the multi-block CST method, which involves fewer design variables and yields higher fidelity, and Pareto genetic algorithms are powerful tools for aerodynamic shape optimization.