郑传宇, 黄江涛, 周铸, 刘刚, 高正红, 许勇. 飞翼翼型高维目标空间多学科综合优化设计[J]. 空气动力学学报, 2017, 35(4): 587-597. DOI: 10.7638/kqdlxxb-2017.0079
引用本文: 郑传宇, 黄江涛, 周铸, 刘刚, 高正红, 许勇. 飞翼翼型高维目标空间多学科综合优化设计[J]. 空气动力学学报, 2017, 35(4): 587-597. DOI: 10.7638/kqdlxxb-2017.0079
ZHENG Chuanyu, HUANG Jiangtao, ZHOU Zhu, LIU Gang, GAO Zhenghong, XU Yong. Multidisciplinary optimization design of high dimensional target space for flying wing airfoil[J]. ACTA AERODYNAMICA SINICA, 2017, 35(4): 587-597. DOI: 10.7638/kqdlxxb-2017.0079
Citation: ZHENG Chuanyu, HUANG Jiangtao, ZHOU Zhu, LIU Gang, GAO Zhenghong, XU Yong. Multidisciplinary optimization design of high dimensional target space for flying wing airfoil[J]. ACTA AERODYNAMICA SINICA, 2017, 35(4): 587-597. DOI: 10.7638/kqdlxxb-2017.0079

飞翼翼型高维目标空间多学科综合优化设计

Multidisciplinary optimization design of high dimensional target space for flying wing airfoil

  • 摘要: 传统的优化方法在处理高维多目标问题上面临多重困难,针对高维多目标优化问题,进行合理的目标降维是一种实用高效的方法。本文首先讨论、研究了PCA目标降维算法在高维多目标应用中的思路,通过典型高维度目标测试函数中的应用,验证了降维方法的有效性,进一步将该方法推广应用于翼型气动隐身多学科综合设计中,对综合设计高维度目标空间进行主成分分析。利用主成分分析提取主成分并辨识"冗余"或者不重要的目标,将冗余目标去除或者作为约束加入到重要目标的优化过程中。结果显示,目标空间降维以后的优化设计结果满足力矩、阻力发散、巡航升阻比、低速升力特性以及隐身特性等综合设计的要求。进一步探讨、展望了该方法在飞行器多目标、多学科优化设计中的应用前景。

     

    Abstract: There are many difficulties when traditional optimization methods are used to deal with high-dimensional multi-objective problems.For the high-dimensional multi-objective optimization problem, it is a practical and effective method to reduce dimensionality.In this paper, we first discuss the idea of principal component analysis (PCA) target dimension reduction method in high-dimensional multi-objective applications.The effectiveness of the dimension reduction method is verified by the application of the test function of typical high dimensional objects.After that, the method is applied to the aerodynamics and stealth multidisciplinary integrated design of airfoils, and principal component analysis is carried out on the integrated design of the high dimensional target space.The PCA is used to extract principal components and identify redundant or unimportant targets.Redundant targets are removed or added as constraints to the optimization process of important targets.The results show that the optimal design results of the target space after dimension reduction satisfy the requirements of torque, drag divergence, cruise lift-drag ratio, and low-speed lift characteristics as well as stealth characteristics.The application prospect of this method in the multi-objective and multi-disciplinary optimization design of aircrafts is further discussed.

     

/

返回文章
返回