基于遗传算法的带筋壁板条带激光喷丸成形工艺优化研究

Optimization of Strip Laser Peen Forming by Genetic Algorithm for Stiffener Integral Panels

为实现带筋壁板条带激光喷丸成形目标曲面,对基于遗传算法的激光喷丸条带分布工艺优化方法进行了研究。建立了固有应变的带筋整体壁板有限元模型,将条带的优化转换为仿真所获曲面与理想目标曲面偏差最小问题,用遗传算法对喷丸条带宽度与间隔进行优化,给出了优化流程,以及遗传算法操作参数的设置。以单筋壁板成形圆柱面为实例,实现了其条带分布优化,获得成形曲面曲率半径与目标圆柱半径偏差为1%,对应节点最大距离0.003 3mm。优化结果表明:遗传算法可较好地适应本优化问题中目标函数难求导的状况,同时执行多方向搜索提高了优化效率。

To obtain the specified surface, the layout optimization of the peening strip in strip laser peen forming based on genetic algorithm was discussed. The finite element model of stiffener integral panel with inherent stain was established. The layout optimization was transferred to the problem of minimum deviation between the curved surface obtained by simulation and the ideal curved surface. The width and distance of peening strip were optimized by genetic algorithm. The optimization flowchart was given. And the parameters were set for genetic algorithm operation. A unidirectional stiffener integral panel forming cylindrical shape was taken as an example and its optimal layout was realized by this approach. It was 1 % deviation of curvature radius between simulated curve and target cylinder and the maximum distance between the corresponding nodes was 0. 003 3 mm. It demonstrated that genetic algorithm was effective to solve the difficulty with the derivation problem of objective function, and improved the optimizing efficiency by searching candidate solutions multi-orientatedly.