Multi-disciplinary analysis of aerodynamics stealth of airfoil based on data mining
-
摘要:
在翼型气动隐身多学科设计中，涉及目标的多样性以及变量之间的相互耦合关系，增大了其优化设计的计算成本和研发周期。针对翼型升力系数、阻力系数、俯仰力矩系数、升阻比、垂直极化雷达散射面积、水平极化雷达散射面积这6个目标，开展了基于随机森林、自适应增强集成学习、自组织映射、等度量映射这4种算法的数据挖掘。在目标与设计变量的分析中，翼型的气动隐身性能受设计变量前缘和后缘弯度影响较大，而受弦长段的影响次之。较大的前缘弯度可以减小阻力，改善隐身性能但增大俯仰力矩系数；较小的后缘弯度可以改善升力系数、升阻比和隐身性能，同时减小俯仰力矩系数。通过数据挖掘，给出了设计变量的具体参考范围以得到气动隐身性能较优的翼型。
Abstract:In the multidisciplinary design of aerodynamic stealth for airfoil profiles, the diversity and coupling relationships among objectives and variables increased the computational cost and development cycle of the optimization design. Focusing on data mining using four types of algorithms: random forest, adaptive boosting algorithm, self-organizing maps and isometric mapping, the data mining considered six objectives: aerodynamic lift coefficient, drag coefficient, pitching moment coefficient, and lift-to-drag ratio, as well as vertical polarized radar cross-section and horizontal polarized radar cross-section. Result showed that, in the analysis of objectives and design variables, the aerodynamic and stealth performance of the airfoil profiles was greatly influenced by the curvature of the leading and trailing edges, followed by the chord length. Larger curvature of the leading edge reduced the drag and improved the stealth performance but increased the pitching moment coefficient. Smaller curvature of the trailing edge improved the lift coefficient, lift-to-drag ratio, and stealth performance while reducing the pitching moment coefficient. Through data mining, specific reference ranges for design variables could be provided to obtain airfoil profiles with superior aerodynamic stealth performance.
-
图 8 随机森林得到的设计变量重要程度排序
Figure 8. Order of importance of design variables obtained by random forest
图 9 自组织映射网络目标函数染色结果
Figure 9. Objective function dyeing results of self-organizing map
图 11 自组织映射网络的设计变量染色结果
Figure 11. Design variable dyeing results for self-organizing map
表 1 设计变量与扰动范围
Table 1. Design variables and disturbance range
变量 初始值 扰动值范围 A0 0.3424 [−0.1, 0.1] A1 0.4398 [−0.2, 0.2] A2 0.2418 [−0.2, 0.2] A3 0.4284 [−0.1, 0.1] A4 0.3162 [−0.2, 0.2] A5 − 0.3424 [−0.1, 0.1] A6 − 0.1657 [−0.2, 0.2] A7 − 0.3835 [−0.2, 0.2] A8 − 0.0838 [−0.1, 0.1] A9 − 0.2851 [−0.2, 0.2] 
下载: 导出CSV
表 3 升力系数CL的设计准则
Table 3. Design rules of CL
序号 规则 数量 1 A7（（−0.293, *）） and A5（（−0.309, *）） and A9（（−0.213, *）） and A6（（−0.277, *）） and A8（（−0.149, *））→high 46 2 A7（（−0.293, *）） and A5（（−0.309, *）） and A9（（−0.213, *）） and A6（（*, −0.277]） and A1（（0.497, *））→high 4 3 A7（（*, −0.293]） and A9（（*, −0.118]） and A2（（*, −0.043]）→high 3
下载: 导出CSV
表 4 阻力系数CD的设计准则
Table 4. Design rules of CD
序号 规则 数量 1 A1（（*, 0.339]） and A6（（−0.296, *）） and A0（（*, 0.341]） and A4（（*, 0.397]） and A2（（*, 0.269]） and A3（（0.441, *）） and
A5（（−0.37, *））→high42 2 A1（（*, 0.339]） and A6（（−0.296, *）） and A0（（*, 0.341]） and A4（（*, 0.32]） and A2（（*, 0.155]） and A3（（*, 0.441]）→high 21
下载: 导出CSV
表 5 俯仰力矩系数CM的设计准则
Table 5. Design rules of CM
序号 规则 数量 1 A7（（*, −0.442]） and A6（（−0.193, *）） and A1（（0.608, *）） and A2（（0.148, *）） and A9（（*, −0.327]）→high 12 2 A7（（*, −0.442]） and A6（（−0.193, *）） and A1（（*, 0.608]）） and A5（（*, −0.442]）→high 2
下载: 导出CSV
表 6 垂直极化Ste的设计准则
Table 6. Design rules of Ste
序号 规则 数量 1 A5（（−0.294, *）） and A9（（−0.14, *）） and A4（（0.344, *）） and A0（（*, 0.301]）→high 7 2 A5（（*, −0.294]） and A0（（0.252, *）） and A4（（*, 0.476]） and A1（（*, 0.241]）→high 2
下载: 导出CSV
表 7 水平极化Stm的设计准则
Table 7. Design rules of Stm
序号 规则 数量 1 A0（（0.301, *）） and A3（（*, 0.35]） and A8（（0.006, *）） and A4（（*, 0.427]）→high 6 2 A0（（0.301, *）） and A3（（*, 0.347]） and A8（（*, 0.006]） and A5（（−0.259, *））→high 4
下载: 导出CSV
表 2 升阻比K的设计准则
Table 2. Design rules of K
序号 规则 数量 1 A6（（−0.113, *）） and A7（（−0.229, *）） and A9（（−0.222, *）） and A0（（0.428, *）） and A2（（*, 0.397]）→high 41 2 A6（（*, −0.113]） and A2（（0.362, *）） and A8（（−0.018, *）） and A4（（0.474, *））→high 4 3 A6（（−0.113, *）） and A7（（−0.229, *）） and A9（（*, −0.222]） and A3（（*, 0.337]） and A1（（0.397, *））→high 1
下载: 导出CSV
-
[1] JEONG S, SHIMOYAMA K. Review of data mining for multi-disciplinary design optimization[J]. Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering, 2011, 225(5): 469-479. doi: 10.1177/09544100JAERO906 [2] SIMPSON T, TOROPOV V, BALABANOV V, et al. Design and analysis of computer experiments in multidisciplinary design optimization: a review of how far we have come-or not[R]. AIAA 2008-5802, 2008. [3] NUNEZ M, GUENOV M D. Design-exploration framework for handling changes affecting conceptual design[J]. Journal of Aircraft, 2013, 50(1): 114-129. doi: 10.2514/1.C031825 [4] CHIBA K, JEONG S, OBAYASHI S, et al. Knowledge discovery in aerodynamic design space for flyback-booster wing using data mining[R]. AIAA 2006-7992, 2006. [5] CHIBA K, OYAMA A, OBAYASHI S, et al. Multidisciplinary design optimization and data mining for transonic regional-jet wing[J]. Journal of Aircraft, 2007, 44(4): 1100-1112. doi: 10.2514/1.17549 [6] OYAMA A, NONOMURA T, FUJII K. Data mining of Pareto-optimal transonic airfoil shapes using proper orthogonal decomposition[J]. Journal of Aircraft, 2010, 47(5): 1756-1762. doi: 10.2514/1.C000264 [7] BATTISTON A, PONZA R, BENINI E, et al. Multi-criteria multi-constrained aerodynamic optimization of civil tiltrotor empennage surfaces[R]. AIAA 2021-3029, 2021. [8] JEONG S, CHIBA K, OBAYASHI S. Data mining for aerodynamic design space[J]. Journal of Aerospace Computing, Information, and Communication, 2005, 2(11): 452-469. doi: 10.2514/1.17308 [9] OBAYASHI S, SASAKI D. Self-organizing map of Pareto solutions obtained from multiobjective supersonic wing design[R]. AIAA 2002-991, 2002. [10] KUMANO T, JEONG S, OBAYASHI S, et al. Multidisciplinary design optimization of wing shape with nacelle and pylon[R]. Egmond aan Zee, the Netherlands: Proceedings of the European Conference on Computational Fluid Dynamics, 2006. [11] KANAZAKI M, JEONG S. Data mining based multipoint design of next generation transonic wing with small sweep back[R]. Nice, France:27th Congress of the International Council of the Aeronautical Sciences, 2010. [12] 钱炜祺, 赵暾, 黄勇, 等. 工程翼型气动特性数据挖掘与建模[J]. 空气动力学学报, 2021, 39(6): 175-183. QIAN Weiqi, ZHAO Tun, HUANG Yong, et al. Modeling and data mining of engineering airfoil aerodynamic characteristics[J]. Acta Aerodynamica Sinica, 2021, 39(6): 175-183. (in Chinese doi: 10.7638/kqdlxxb-2021.0219QIAN Weiqi, ZHAO Tun, HUANG Yong, et al. Modeling and data mining of engineering airfoil aerodynamic characteristics[J]. Acta Aerodynamica Sinica, 2021, 39(6): 175-183. (in Chinese) doi: 10.7638/kqdlxxb-2021.0219 [13] 段焰辉, 吴文华, 范召林, 等. 基于本征正交分解的气动优化设计外形数据挖掘[J]. 物理学报, 2017, 66(22): 138-147. DUAN Yanhui, WU Wenhua, FAN Zhaolin, et al. Proper orthogonal decomposition-based data mining of aerodynamic shape for design optimization[J]. Acta Physica Sinica, 2017, 66(22): 138-147. (in ChineseDUAN Yanhui, WU Wenhua, FAN Zhaolin, et al. Proper orthogonal decomposition-based data mining of aerodynamic shape for design optimization[J]. Acta Physica Sinica, 2017, 66(22): 138-147. (in Chinese) [14] 刘深深, 陈江涛, 桂业伟, 等. 基于数据挖掘的飞行器气动布局设计知识提取[J]. 航空学报, 2021, 42(4): 524708. LIU Shenshen, CHEN Jiangtao, GUI Yewei, et al. Knowledge discovery for vehicle aerodynamic configuration design using data mining[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524708. (in ChineseLIU Shenshen, CHEN Jiangtao, GUI Yewei, et al. Knowledge discovery for vehicle aerodynamic configuration design using data mining[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 524708. (in Chinese) [15] 李海泉, 陈小前, 左林玄, 等. 基于随机森林的飞行载荷代理模型分析方法[J]. 航空学报, 2022, 43(3): 225640. LI Haiquan, CHEN Xiaoqian, ZUO Linxuan, et al. Surrogate model for flight load analysis based on random forest[J]. Acta Aeronautica et Astronautica Sinica, 2022, 43(3): 225640. (in ChineseLI Haiquan, CHEN Xiaoqian, ZUO Linxuan, et al. Surrogate model for flight load analysis based on random forest[J]. Acta Aeronautica et Astronautica Sinica, 2022, 43(3): 225640. (in Chinese) [16] 李润泽, 张宇飞, 陈海昕. 针对超临界翼型气动修型策略的强化学习[J]. 航空学报, 2021, 42(4): 523810. LI Runze, ZHANG Yufei, CHEN Haixin. Reinforcement learning method for supercritical airfoil aerodynamic design[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 523810. (in ChineseLI Runze, ZHANG Yufei, CHEN Haixin. Reinforcement learning method for supercritical airfoil aerodynamic design[J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(4): 523810. (in Chinese) [17] 何磊, 钱炜祺, 刘滔, 等. 基于深度学习的翼型反设计方法[J]. 航空动力学报, 2020, 35(9): 1909-1917. HE Lei, QIAN Weiqi, LIU Tao, et al. Inverse design method of airfoil based on deep learning[J]. Journal of Aerospace Power, 2020, 35(9): 1909-1917. (in ChineseHE Lei, QIAN Weiqi, LIU Tao, et al. Inverse design method of airfoil based on deep learning[J]. Journal of Aerospace Power, 2020, 35(9): 1909-1917. (in Chinese) [18] 廖鹏, 姚磊江, 白国栋, 等. 基于深度学习的混合翼型前缘压力分布预测[J]. 航空动力学报, 2019, 34(8): 1751-1758. LIAO Peng, YAO Leijiang, BAI Guodong, et al. Prediction of hybrid airfoil leading edge pressure distribution based on deep learning[J]. Journal of Aerospace Power, 2019, 34(8): 1751-1758. (in ChineseLIAO Peng, YAO Leijiang, BAI Guodong, et al. Prediction of hybrid airfoil leading edge pressure distribution based on deep learning[J]. Journal of Aerospace Power, 2019, 34(8): 1751-1758. (in Chinese) [19] BREIMAN L. Prediction games and arcing algorithms[J]. Neural Computation, 1999, 11(7): 1493-1517. doi: 10.1162/089976699300016106 [20] BREIMAN L. Random forests[J]. Machine Learning, 2001, 45: 5-32. doi: 10.1023/A:1010933404324 [21] OSHIRO T M, PEREZ P S, BARANAUSKAS J A. How many trees in a random forest? [C]// Machine Learning and Data Mining in Pattern Recognition. Berlin: Springer, 2012: 154-168. [22] CUTLER A, CUTLER D R, STEVENS J R. Random forests[M]//Ensemble Machine Learning. New York: Springer, 2012: 157-175. [23] FREUND Y, SCHAPIRE R E. Experiments with a new boosting algorithm[J]. Machine Learning, 1996, 96: 148-156. [24] RICHARDSON T, NEKOLNY B, HOLUB J, et al. Visualizing design spaces using two-dimensional contextual self-organizing maps[J]. AIAA Journal, 2014, 52(4): 725-738. doi: 10.2514/1.J052181 [25] CHEN Shusheng, LI Jinping, LI Zheng, et al. Anti-dissipation pressure correction under low Mach numbers for Godunov-type schemes[J]. Journal of Computational Physics, 2022, 456: 111027. doi: 10.1016/j.jcp.2022.111027 [26] 陈树生, 张兆康, 李金平, 等. 一种宽速域乘波三角翼气动布局设计[J]. 航空学报, 2023, 44(24): 128441. CHEN Shusheng, ZHANG Zhaokang, LI Jinping, et al. A wide-speed aerodynamic layout adopting waverider-delta wing[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(24): 128441. (in ChineseCHEN Shusheng, ZHANG Zhaokang, LI Jinping, et al. A wide-speed aerodynamic layout adopting waverider-delta wing[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(24): 128441. (in Chinese) [27] 刘衍旭, 陈树生, 冯聪, 等. 高超声速滑翔飞行器锐边化气动隐身一体化设计[J]. 航空学报, 2023, 44(16): 128093. LIU Yanxu, CHEN Shusheng, FENG Cong, et al. Integrated sharp edge aerodynamic-stealth design for hypersonic glide vehicle[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(16): 128093. (in ChineseLIU Yanxu, CHEN Shusheng, FENG Cong, et al. Integrated sharp edge aerodynamic-stealth design for hypersonic glide vehicle[J]. Acta Aeronautica et Astronautica Sinica, 2023, 44(16): 128093. (in Chinese) [28] KNOTT E F. A progression of high-frequency RCS prediction techniques[J]. Proceedings of the IEEE, 1985, 73(2): 252-264. doi: 10.1109/PROC.1985.13137 [29] SCORNET E. On the asymptotics of random forests[J]. Journal of Multivariate Analysis, 2016, 146: 72-83. doi: 10.1016/j.jmva.2015.06.009 [30] MILJKOVIĆ D. Brief review of self-organizing maps[C]//2017 40th International Convention on Information and Communication Technology, Electronics and Microelectronics. Piscataway, US: IEEE, 2017: 1061-1066. [31] KOHONEN T. Essentials of the self-organizing map[J]. Neural Networks, 2013, 37: 52-65. doi: 10.1016/j.neunet.2012.09.018 [32] NAGAR D, RAMU P, DEB K. Visualization and analysis of Pareto-optimal fronts using interpretable self-organizing map (iSOM)[J]. Swarm and Evolutionary Computation, 2023, 76: 101202. doi: 10.1016/j.swevo.2022.101202 [33] TENENBAUM J B, DE SILVA V, LANGFORD J C. A global geometric framework for nonlinear dimensionality reduction[J]. Science, 2000, 290(5500): 2319-2323. doi: 10.1126/science.290.5500.2319 -
点击查看大图
计量
- 文章访问数: 409
- HTML浏览量: 182
- PDF量: 42
- 被引次数: 0