基于遗传算法的复合材料圆柱壳屈曲多目标优化设计

Buckling multi-objective optimum design based on genetic algorithm for composite cylinder shell

针对复合材料圆柱壳屈曲载荷最大化和缺陷敏感度最小化问题,提出以铺层纤维方向角为设计变量的基于遗传算法的多目标优化方法,采用层合壳屈曲理论计算临界载荷和统计方法计算缺陷敏感度,对由多个个体组成的群体施加选择、交叉和变异等演化操作,通过多代进化得到问题的pareto解。算例优化结果与采用离散方法得到的目标空间解边界比较吻合,验证了方法的准确性和有效性。

In this paper, a multi-objective design method based on genetic algorithm for maximizing buckling load and minimizing imperfection sensitivity of CFRP cylinder shell under axial compression is proposed. Design variables are fiber orientations of composite laminate. The population consists of individuals which contain critical load and imperfection sensitivity computed by cylinder shell buckling theory and statistics method respectively. The operator such as selection crossover and mutation is imposed on this population. The results of two examples show that this method has high accuracy compared with function space obtained by discrete method and is effective in searching pareto solutions of problem.