航空学报 > 2011, Vol. 32 Issue (4): 636-648   doi: CNKI:11-1929/V.20101130.1743.000

多铺层碳纤维蜂窝板模型修正

秦玉灵1, 孔宪仁1, 罗文波2   

  1. 1. 哈尔滨工业大学 航天学院, 哈尔滨 黑龙江 150001;
    2. 中国空间技术研究院, 北京 100086
  • 收稿日期:2010-07-01 修回日期:2010-09-30 出版日期:2011-04-25 发布日期:2011-04-25
  • 通讯作者: Tel.: 0451-86402357 E-mail: kongxr@hit.edu.cn E-mail:kongxr@hit.edu.cn
  • 作者简介:秦玉灵(1982- ) 女,博士研究生。主要研究方向:卫星结构动力学及模型修正。 Tel: 0451-86402357 E-mail: erica2004ren@163.com孔宪仁(1961- ) 男,博士,教授,博士生导师。主要研究方向:卫星结构动力学及热力学控制。 Tel: 0451-86402357 E-mail: kongxr@hit.edu.cn罗文波(1968- ) 男,博士,教授。主要研究方向:卫星结构动力学及环境试验。 E-mail: luowb999@sohu.com
  • 基金资助:

    "微小型航天器系统技术"长江学者创新团队发展计划(IRT0520)

Model Updating for Multi-layered Carbon Fiber Honeycomb Sandwich Panel

QIN Yuling1, KONG Xianren1, LUO Wenbo2   

  1. 1. School of Astronautics, Harbin Institute of Technology, Harbin 150001, China;
    2. China Academy of Space Technology, Beijing 100086, China
  • Received:2010-07-01 Revised:2010-09-30 Online:2011-04-25 Published:2011-04-25

摘要: 蜂窝板是现代飞行器的主要承力结构,通过分析各形式响应面适用范围,提出Linear-and-Gaussian组合核支持向量机(SVM)响应面和基于分组控制策略的改进粒子群优化(IPSO)算法。用ANSYS的SHELL91单元建立多铺层碳纤维蜂窝板的有限元模型(FEM),并通过正交试验设计和F值检验确定待修正结构参数,构造Linear-and-Gaussian响应面以拟合待修正结构参数与蜂窝板模态频率的关系并检验响应面模型有效性。最后,用基于分组控制策略的IPSO算法对响应面模型中的结构参数进行修正,修正后参数代入原有限元模型得到修正模型。通过对修正前后模型模态频率与基准模型模态频率在测试频段内外的对比,证实了修正后模型具有良好的复现能力和预测能力。

关键词: 响应面, 多铺层碳纤维蜂窝板, 组合核支持向量机, 分组控制策略, 粒子群优化算法

Abstract: The honeycomb sandwich panel is the main load-carrying structure of modern aircraft. Linear-and-Gaussian combined kernel function support vector machine (SVM) response surface and the group-control-based improved particle swarm optimization (IPSO) algorithm are proposed in this paper through an analysis of various response surfaces, and a finite element model (FEM) of the multi-layered carbon fiber honeycomb sandwich panel is constituted in ANSYS using SHELL91. The non-updated parameters are chosen by orthogonal design and F-test, which are then employed to constitute the Linear-and-Gaussian response surface to simulate the relationship between the structure parameters and the responses, and then verify the validity of the response surface model. The group-control-based IPSO algorithm is employed to update the non-updated parameters and the updated parameters are substituted into the non-updated FEM. A comparison of the modal frequencies of the non-updated and updated FEM and the benchmark FEM proves the reappearance and prediction ability of the updated FEM.

Key words: response surface, multi-layered carbon fiber honeycomb sandwich panel, combined kernel function support vector machine, group-control strategy, particle swarm optimization algorithm

中图分类号: