含初始缺陷复合材料格栅加筋圆柱壳的鲁棒优化设计

ROBUST OPTIMIZATION OF ADVANCED GRID COMPOSITE CYLINDRICAL SHELL CONSIDING INITIAL IMPERFECT EFFECT

针对复合材料格栅加筋圆柱壳具有多约束、连续变量和非连续变量混合的特性,基于复合材料中厚圆柱壳体的非线性稳定性理论,考虑优化目标和约束鲁棒性,采用混合遗传算法,提出了一种非完善复合材料格栅加筋圆柱壳的鲁棒优化设计方法.应用正交设计理论确定主要的鲁棒优化设计变量,简化了运算过程,从而提高了优化分析效率;同时又采用五点差分格式代替梯度运算,解决了在优化计算过程中不可求导的问题.通过典型算例表明,鲁棒优化设计所得到的优化分析结果与传统确定性优化方法相比差别较大,此外壳体的初始缺陷亦将对优化结果产生很大的影响.

Based on the characters of optimal design for the composite grid-stiffened structures （AGS）, such as multi-loading cases, muhi-constraints, and mixed discrete continuous design variables, etc, a robust optimization design method using a hybrid genetic algorithm provided by the authors has been proposed for the advanced grid composite cylindrical shell with initial-imperfect. A nonlinear stability shell theory, based on Reissner-Mindlin deformation assumption, was applied in the re-analysis process of the shell stability. Five point difference scheme was adopted to solve the non-differentiable problem and the orthogonal design method was used to select the main design variables for reducing the computational cost. From discussion of a typical example, it can be seen that （ 1 ） the robuse optimization results are quite different from that by general deterministic optimization method; （2） the initial imperfection of the shells has an evident influence on the optimal results.