基于特征正交分解的材料微结构参数化表征模型及等效性能优化设计

发布时间：2019-05-12 20:10

【摘要】：随着计算能力的不断发展,近年来基于材料微结构图像的材料等效性能数值模拟越来越受到学者们的重视.在此背景下,提出了一种针对材料微结构图像的高效参数化表征模型.通过特征正交分解(proper orthogonal decomposition,POD)对已有材料微结构图像数据进行特征分析,得到近似描述该类材料微结构的特征缩减基.应用移动最小二乘(moving least squares,MLS)法建立特征缩减基映射系数的响应面模型,拟合得到任意给定参量相应的缩减基映射系数.利用拟合缩减基系数可获得任意给定参量对应的微结构图像矩阵.该参数化表征模型被用于表征含椭球夹杂的两相材料(2-phase composite)的二维情形,并进一步应用于这类复合材料宏观等效力学性能的优化设计.
[Abstract]:With the development of computing power, the numerical simulation of material equivalent performance based on material microstructure image has been paid more and more attention by scholars in recent years. In this context, an efficient parametric representation model for material microstructure images is proposed. The image data of microstructure of existing materials are analyzed by feature orthogonal decomposition (proper orthogonal decomposition,POD), and the feature reduction base which approximately describes the microstructure of this kind of materials is obtained. The response surface model of feature reduced basis mapping coefficient is established by moving least square (moving least squares,MLS) method, and the corresponding reduced basis mapping coefficient of any given parameter is obtained by fitting. The microstructure image matrix corresponding to any given parameter can be obtained by using the fitting reduction base coefficient. The parametric characterization model is used to characterize the two-dimensional case of two-phase materials (2-phase composite) with ellipsoid inclusions and is further applied to the optimization design of macroscopic equivalent mechanical properties of this kind of composites.
【作者单位】：西北工业大学理学院数学与应用数学系;巴黎东部大学多尺度模拟与仿真实验室(CNRS
【基金】：国家自然科学基金青年科学基金(11302173) 陕西省自然科学基金(2017JQ1037)~~
【分类号】：TB33

【相似文献】