首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于Kriging模型的结构耐撞性优化
引用本文:何欢,朱广荣,何成,等.基于Kriging模型的结构耐撞性优化[J].南京航空航天大学学报,2014,46(2):297-303.
作者姓名:何欢  朱广荣  何成  
作者单位:南京航空航天大学振动工程研究所
摘    要:提出了基于Kriging模型的耐撞性优化方法。首先就Kriging模型的构造方法及其精度评估问题进行了讨论;然后,以薄壁管为研究对象,采用瞬态非线性有限元分析程序作为计算核心,以薄壁圆管的直径和壁厚为优化变量,以最大撞击载荷为目标函数,薄壁管的最大压缩量等作为约束函数,构造了基于Kriging模型的全局近似函数来逼近真实的优化目标函数与约束函数;随后,提出了提高全局近似函数精度的Kriging模型更新方法,改进了优化设计分析流程;最后,在所构造的全局近似函数的基础上,采用遗传算法进行优化分析。算例分析结果表明,该方法构造的最大撞击载荷与最大压缩量的全局近似函数在最优解处与真实解非常吻合,说明了Kriging模型的有效性。

关 键 词:Kriging模型  全局近似函数  耐撞性  优化

Crashworthiness Optimization Based on Kriging Metamodeling
Abstract:A crashworthiness optimization method based on Kriging metamodeling is proposed. Construction of the Kriging metamodeling method and its evaluation of the accuracy criteria are discussed firstly. Transient nonlinear finite element method is employed to model the crush process of thin walled tube. Secondly, radius and thickness of the tube are defined as independent design variables and the global approximation model is built based on Kriging metamodling with the maximum impact load and the maximum crush length defined as optimal object and constraints function respectively. Then, an updating strategy of Kriging metamodeling is introduced to improve accuracy of the global approximation and the optimization process is introduced too. Finally, both object and constraint functions obtained through global approximation model are coupled with genetic algorithms to perform the objective optimization.
Keywords:Kriging metamodeling  global approximation function  crashworthiness  optimization
本文献已被 CNKI 等数据库收录!
点击此处可从《南京航空航天大学学报》浏览原始摘要信息
点击此处可从《南京航空航天大学学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号