基于工程环境的气动多目标优化设计平台研究

工程环境中,飞机气动力设计面临在多个目标和多种约束条件下寻找最优值,需在较短时限内完成设计优化,并保证最终方案可靠。基于高性能计算环境,采用现代计算流体力学(CFD)数值模拟技术和优化技术等构建了面向实际工程的飞行器气动多目标优化设计平台:采用基于非均匀有理B样条(NURBS)方法的自由曲面变形技术,实现对工程复杂气动外形的参数化表达;采用网格变形技术,实现优化过程中计算网格的自动更新;采用基于有限体积方法和多块结构网格的雷诺平均Navier-Stokes(RANS)方程并行解算器进行气动力求解;采用基于精英保留策略的非支配排序的多目标遗传算法(NSGA-II)进行多目标全局优化求解;采用非线性单纯形算法进行局部优化求解,优化过程中,通过人工调整优化种群,引入人工经验,构建"人在回路"的设计流程。以某翼型/机翼气动力优化设计为例对该平台技术进行验证:多目标优化设计可得到清晰的Pareto前沿解分布;优化后的翼型/机翼在满足各项约束的前提下,具备更高的综合气动性能。结果表明:所发展的气动多目标优化设计平台具有很好的工程适用性。

Engineering environment-based multi-objective optimization platform for aerodynamic design

In the engineering environment,the aim of aerodynamic design of an aircraft is to find the optimal value under the multi-objectives and multi constraints.The design optimization should be completed in a short time and the final scheme must be reliable.Based on high-performance computing environment,an engineering-practical multi-objective optimization platform for aerodynamic design is constructed by adopting modern computational fluid dynamics(CFD) numerical simulation and optimization techniques.The free form deformation method based on non-uniform rational b-splines(NURBS) is used to obtain the parametrized representation of aerodynamic shapes;the mesh deformation method is taken to realize the automatic deformation of computational mash in optimization;a Reynolds averaged Navier-stokes solver based on finite volume method and multi-block structured mesh is used to find out aerodynamic forces;the non-dominated sorting-based multi-objective genetic algorithm(NSGA-II) is used to search for global optimum,while the nonlinear simplex algorithm method is used to search for local optimum.During the optimization process,a "human-in-the-loop" design process is configurated through artificially changing optimum population and introducing artificial experience.For the validation of this platform,optimization design of an airfoil/wing's aerodynamic forces is taken as an example.A clear Pareto-optimal front can be obtained by multi-objective optimization design;on the premise of meeting all constraints,the comprehensive aerodynamic performance of optimized airfoil/wing is significantly improved.The result shows that the multi-objective optimization platform for aerodynamic design developed in this paper can be well applied to engineering practice.