TY - JOUR
T1 - Adaptive aerodynamic optimization design method based on design variables space reconstruction concept
AU - Zhu, Jun
AU - Gao, Zhenghong
AU - Bai, Junqiang
AU - Zhan, Hao
PY - 2013/10
Y1 - 2013/10
N2 - Due to the strong empirical demand of setting the design space variation in optimization design process which strongly affect the design result and efficiency, the adaptive aerodynamic optimization design method based on design variables space reconstruction concept has been established by the spatial statistical analysis of the design variables spatial distribution in the optimization process, which resolves the issue of design variables spatial distribution selection and lead to a better flexibility and convergence capability of the optimization model. As the design variables distribution has been statistically analyzed in the optimization process with the concept of clustering level, on the one hand the design variables spatial variation is reconstructed, on the other hand the design variables of a part of samples is adjusted in the reconstructed design space, both of which result in a better population diversity and a faster convergence with reservation of good genetic information. The NACA 0012 airfoil and NLF(1) 0416 airfoil have been optimized by this method, the results of which have been compared and analyzed with that of optimization of fixed design space variation. The design method approached has been proved of good feasibility with the optimization results, which results in an optimum solution in a larger variables distribution scale and shows better optimization efficiency.
AB - Due to the strong empirical demand of setting the design space variation in optimization design process which strongly affect the design result and efficiency, the adaptive aerodynamic optimization design method based on design variables space reconstruction concept has been established by the spatial statistical analysis of the design variables spatial distribution in the optimization process, which resolves the issue of design variables spatial distribution selection and lead to a better flexibility and convergence capability of the optimization model. As the design variables distribution has been statistically analyzed in the optimization process with the concept of clustering level, on the one hand the design variables spatial variation is reconstructed, on the other hand the design variables of a part of samples is adjusted in the reconstructed design space, both of which result in a better population diversity and a faster convergence with reservation of good genetic information. The NACA 0012 airfoil and NLF(1) 0416 airfoil have been optimized by this method, the results of which have been compared and analyzed with that of optimization of fixed design space variation. The design method approached has been proved of good feasibility with the optimization results, which results in an optimum solution in a larger variables distribution scale and shows better optimization efficiency.
KW - Clustering
KW - Optimization design
KW - Optimization model
KW - Space reconstruction
KW - Statistical distribution
UR - http://www.scopus.com/inward/record.url?scp=84888992854&partnerID=8YFLogxK
U2 - 10.1177/0954410012445954
DO - 10.1177/0954410012445954
M3 - 文章
AN - SCOPUS:84888992854
SN - 0954-4100
VL - 227
SP - 1535
EP - 1544
JO - Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering
JF - Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering
IS - 10
ER -