点击次数：

论文类型：期刊论文

第一作者：Hao Peng

通讯作者：Hao, P (reprint author), Dalian Univ Technol, Dept Engn Mech, State Key Lab Struct Anal Ind Equipment, Dalian 116023, Liaoning, Peoples R China.

合写作者：Wang Yutian,Liu Chen,Wang Bo,Tian Kuo,Li Gang,Wang Qun,Jiang Liangliang

发表时间：2018-10-01

发表刊物：AIAA JOURNAL

收录刊物：SCIE

卷号：56

期号：10

页面范围：4180-4194

ISSN号：0001-1452

摘要：The concept of a curvilinearly stiffened panel is promising for aerospace and aircraft structures with cutouts, since the stiffness distribution and loading path can be flexibly tailored in terms of cutouts. However, uncertainties from the manufacturing process can significantly reduce the practical load-carrying capacity of a curvilinearly stiffened panel. Traditional nondeterministic design optimization methods would suffer from a high computational burden because numerous discrete-continuous variables, in which thousands of finite element analyses are usually required, are involved. Therefore, a hierarchical nondeterministic optimization framework of a curvilinearly stiffened panel is established, containing two steps (global search and local search) and four main parts (layout design, nondeterministic design optimization, reshape layout design, and stress validation). In the first-step deterministic optimization, all variables are considered, and a rough result set will be obtained by a global search. In the second-step nondeterministic optimization, all discrete-continuous layout parameters are fixed in the initial stage, and the efficient adaptive-loop method is applied to significantly increase the computational efficiency, in which the variations of material properties and geometric dimensions are considered as uncertainties. The layout also will be reshaped to explore the potential of the load-carrying capacity. Finally, the stress constraint and crippling constraint are verified for the optimum design. An illustrative example indicates the advantage of the proposed framework not merely provides a competitive efficiency and robustness over other nondeterministic approaches but also improves the rationality of deterministic design optimization.