A New Design Method for Market-Oriented Aircraft

针对国产或亚洲类似空中客车联合体研制的飞机，进行了市场研究、计算流体动力学（CFD）研究和设计优化研究。市场研究的重点是超音速区域运输的可行性，该运输在一天的往返时间内连接亚洲城市。提出了一种超声速区域喷气机的概念设计。复杂进化多目标优化（EMO）已被发展并应用于超音速机翼设计。该研究是最现实的EMO应用之一，已得到国际上的认可。EMO同时搜索多个目标函数之间的折衷解（帕累托解）。解决方案的结果集揭示了相互冲突的设计目标之间的设计权衡。将该方法应用于具有四个设计目标的超声速机翼设计，并对各种设计权衡进行了检验。两个目标之间的权衡研究是微不足道的：然而，随着设计目标数量的增加，它将变得更加困难。因此，针对EMO的实际应用，将Pareto解的可视化作为后处理，与EMO本身的改进和应用一起进行研究。自组织映射（SOM）是一种用于高维空间的可视化技术。利用目标函数空间中Pareto解的SOM对多个设计目标之间的设计权衡进行数据挖掘。SOM还被证实是一种有用的数据挖掘工具，它可以映射设计变量空间中的解决方案，因为得到的SOM通过设计变量簇代表了机翼设计中的气动知识。除此之外，非结构化伴随方法作为一种可选的优化方法，以及将CFD代码扩展到流固耦合以实现多学科设计优化的研究也在进行。这些活动已在国际会议上得到报告。少

Toward aircraft built domestically or developed by Asian Airbus-like consortium, researches have bee conducted for market, Computational Fluid Dynamics (CFD), and design optimization. Market study focused on feasibility of supersonic regional transport, which connects Asian cities within one-day round trip. A conceptual design of supersonic regional jet has been presented. Sophisticated Evolutionary Multiobjective Optimization (EMO) has been developed and applied to supersonic wing design. This research is one of the most realistic EMO applications and has been recognized internationally.EMO searches many compromised solutions (Pareto solutions) between multiple objective functions simultaneously. The resulting set of solutions reveals design tradeoffs between conflicting design objectives. This method has been applied to supersonic wing design with four design objective and various design tradeoffs have been examined. The tradeoff study is trivial between two objectives : however, it … More will become more difficult as the number of design objectives increases. Therefore, aiming for real-world applications of EMO, visualization of Pareto solutions are studied as post-processing together with the improvement and application of EMO itself. Self-Organizing Map (SOM) has been utilized as a visualization technique of high dimensional space. Datamining of design tradeoffs between multiple design objectives are performed by using SOM of Pareto solutions in the objective function space. SOM is also confirmed to be useful as the datamining tool by mapping the solutions in design variable space because the resulting SOM represents the aerodynamic knowledge in wing design througjh the clusters of design variables.In addition to these works, researches have been performed for unstructured adjoint approach as an alternative optimization method and for extension of CFD code to fluid-structure interaction toward multidisciplinary design optimization. These activities have been reported at international conferences. Less