摘要
从全局搜索性出发,分析模型修正中所使用的常用优化方法的不足,进而提出基于遗传算法的模型修正方法。以工程设计中常用的变截面梁模型为例,推导截面为环形的平面弯曲梁单元的刚度矩阵,建立综合考虑结构固有频率和模态振型的目标函数。以一个30个节点的变截面梁有限元模型为例,采用遗传算法对模型进行优化,计算结果表明,优化后的模型频率与振型都与试验结果符合较好,验证了基于遗传算法的模型修正方法的可实现性。
In this paper, the insufficiency of common optimization algorithms used in model updating is analyzed on the basis of global search, then, the method of model updating based on genetic algorithm is presented. By taking a non-uniform beam model commonly used in engineering for example, the stiffness matrices of the plane bending beam with annular section is deduced, the object function which considering structural natural frequencies and modal shapes is established. A non-uniform beam model with 30 nodes is updated by using genetic algorithm. The results show that the frequencies and vibration shapes of the optimized model are in good agreement with experimental data and the feasibility of the method is verified.
出处
《导弹与航天运载技术》
北大核心
2013年第2期60-63,共4页
Missiles and Space Vehicles
关键词
全局搜索
遗传算法
模型修正
Global search
Genetic algorithm
Model updating
作者简介
王乐（1984-），男，硕士研究生，主要从事载荷、力学环境和结构动力学研究
