TY - JOUR
T1 - The niching-based adaptive space reconstruction method for airfoil aerodynamic/stealth design
AU - Zhang, Wei
AU - Gao, Zhenghong
AU - Wang, Chao
AU - Xia, Lu
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.
PY - 2023/7
Y1 - 2023/7
N2 - The airfoils for flying wing aircraft need good aerodynamic, trim and stealth performance, and these unconventional requirements would lead to multimodality in airfoil design optimizations. It would be time-consuming for a single-point sequence sampling method, like the Efficient Global Optimization method (EGO), to obtain the global optima over the entire design space, and the traditional multi-start sampling approach would tend to find the local minima which were well-explored by the surrogate model. Thus, the Niching-based Adaptive Space Reconstruction method (NASR) was introduced to enhance the performance of the EGO for aerodynamic/stealth optimizations in this paper. The proposed method adopted a Niching-based multiple points sampling method to simultaneously search out the locations of multiple local minima of the EI function, which did not rely on the starting points and avoided converging to the same local minima. Moreover, it would check the neighborhood of the local minima and focus on the local minima which were not well-explored to decrease computational cost. To decrease the risk of missing promising regions, different expansion strategies were applied to adjust the size of design intervals in different iterations. Several numerical and structural optimization problems were studied to test and verify the efficiency and robustness of NASR. Then, airfoil optimizations considering aerodynamic, stealth and aero/stealth performance by NASR were conducted. It was shown that the aerodynamic performance of the aerodynamic optimized result was comparable with the aero/stealth optimization, and the stealth performance of the stealth optimized result was better than the aero/stealth optimization. The conflictions between Radar Cross Sections (RCS) reduction and drag reduction were not very prominent. Moreover, although the pitch positive moment constraint would damage the lift-drag characteristics, it could be beneficial to RCS reduction based on the current study.
AB - The airfoils for flying wing aircraft need good aerodynamic, trim and stealth performance, and these unconventional requirements would lead to multimodality in airfoil design optimizations. It would be time-consuming for a single-point sequence sampling method, like the Efficient Global Optimization method (EGO), to obtain the global optima over the entire design space, and the traditional multi-start sampling approach would tend to find the local minima which were well-explored by the surrogate model. Thus, the Niching-based Adaptive Space Reconstruction method (NASR) was introduced to enhance the performance of the EGO for aerodynamic/stealth optimizations in this paper. The proposed method adopted a Niching-based multiple points sampling method to simultaneously search out the locations of multiple local minima of the EI function, which did not rely on the starting points and avoided converging to the same local minima. Moreover, it would check the neighborhood of the local minima and focus on the local minima which were not well-explored to decrease computational cost. To decrease the risk of missing promising regions, different expansion strategies were applied to adjust the size of design intervals in different iterations. Several numerical and structural optimization problems were studied to test and verify the efficiency and robustness of NASR. Then, airfoil optimizations considering aerodynamic, stealth and aero/stealth performance by NASR were conducted. It was shown that the aerodynamic performance of the aerodynamic optimized result was comparable with the aero/stealth optimization, and the stealth performance of the stealth optimized result was better than the aero/stealth optimization. The conflictions between Radar Cross Sections (RCS) reduction and drag reduction were not very prominent. Moreover, although the pitch positive moment constraint would damage the lift-drag characteristics, it could be beneficial to RCS reduction based on the current study.
KW - Aerodynamic/stealth design
KW - EGO
KW - Global optimization
KW - Kriging model
KW - Niching-based multiple points sampling
UR - http://www.scopus.com/inward/record.url?scp=85163764344&partnerID=8YFLogxK
U2 - 10.1007/s00158-023-03608-7
DO - 10.1007/s00158-023-03608-7
M3 - 文章
AN - SCOPUS:85163764344
SN - 1615-147X
VL - 66
JO - Structural and Multidisciplinary Optimization
JF - Structural and Multidisciplinary Optimization
IS - 7
M1 - 159
ER -