刘书田

个人信息Personal Information

教授

博士生导师

硕士生导师

性别:男

毕业院校:大连理工大学

学位:博士

所在单位:力学与航空航天学院

学科:工程力学. 计算力学. 航空航天力学与工程

办公地点:综合实验I号楼 512

扫描关注

论文成果

当前位置: 刘书田 >> 科学研究 >> 论文成果

Mapping method for sensitivity analysis of composite material property

点击次数:

论文类型:期刊论文

发表时间:2002-09-01

发表刊物:STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION

收录刊物:Scopus、SCIE、EI

卷号:24

期号:3

页面范围:212-217

ISSN号:1615-147X

关键字:mapping method; sensitivity analysis; homogenization; composite materials

摘要:Composite properties are dependent on the microstructure of materials, which is depicted with a base cell. The parameters for representing the microstructure should include the shape parameters of the base cell and those used to describe the distribution of materials in the base cell. The goal of material design optimization is to find appropriate values of these parameters to make the materials have specific properties. Design optimization needs the sensitivity information of the material properties with respect to the shape parameter of the base cell and the material distribution parameters. Moreover, sensitivity calculation is often expensive. Thus, it is very important to develop an efficient sensitivity analysis method. In this paper, a mapping method is proposed for predicting the material properties and computing their sensitivities with respect to the shape parameters of the base cell. Through mapping transformation, solutions to the micro-scale homogenization problem defined on the domain of a base cell can be obtained by solving a homogenization problem defined on an initial given domain. The composite properties and their sensitivities with respect to the shape parameters of the base cell are explicitly expressed in terms of the properties and their sensitivities of a virtual material with respect to the distribution parameters. This virtual material has an initially given base cell domain. Thus re-meshing for discretizing the problem is avoided and computing cost savings are realized. Numerical examples show that the proposed method is accurate and efficient in both the prediction of material properties and sensitivity calculation.