著者
金崎 雅博 奈良 拓矢 四谷 智義
出版者
日本計算工学会
雑誌
計算工学講演会論文集 (ISSN:1342145X)
巻号頁・発行日
no.15, 2010-05

In this paper, airfoil geometry representation is proposed for the efficient design optimization and the knowledge discovery. Here, PARSEC airfoil representation is improved with several modifications. To investigate the possibility to solve the unknown real-world design problem, two type airfoil geometries is considered using genetic algorithms. One is a conventional transonic airfoil in the Earth atmosphere, the other is a airfoil which can be used in Martian atmosphere To evaluate the aerodynamic performance, the structured Navier-Stokes solver with Baldwin-Lomax turbulent model is used. According to this result, the proposed airfoil representation method can generate the airfoil geometry which achieves high aerodynamic performance as same as NURBS representation. These results suggest that the proposed method can useful for the design exploration methods which include data mining, because the airfoil geometry can be defined with the design variables which show aerodynamic performance directly.