基于自洽聚类分析的2D C/SiC压缩性能快速预报  

Fast prediction of 2D C/SiC compression performance based on self-consistent clustering analysis

作　　者：戴新航 许承海[1] 王琨杰 高博[1] DAI Xinhang;XU Chenghai;WANG Kunjie;GAO Bo(National Key Laboratory of Science and Technology on Advanced Composites in Special Environments,Harbin Institute of Technology,Harbin 150001,China)

机构地区：[1]哈尔滨工业大学特种环境复合材料技术国家级重点实验室,哈尔滨150001

出　　处：《复合材料学报》2024年第8期4386-4397,共12页Acta Materiae Compositae Sinica

摘　　要：本文利用自洽聚类分析(Self-consistent clustering analysis,SCA)方法研究了2D C/SiC在单轴压缩载荷下的渐进损伤行为,SCA方法通过应变集中张量对网格单元进行聚类,在不显著降低计算精度的前提下,大幅度降低了模型的自由度,使模型的计算效率得以提高。整个方法由离线和在线两个阶段组成:离线阶段,利用k-means算法对高保真度的复合材料单胞进行分解、聚类并计算不同聚类间的相互作用张量,最终生成降阶模型;在线阶段,基于降阶模型求解离散的Lippmann-Schwinger方程组获取力学响应。将SCA方法应用于2D C/SiC压缩强度的预报,当聚类总数量为64时,与试验相比,压缩强度求解的计算精度与传统有限元相比降低了1%,但整体计算效率提升了34倍。当不考虑离线阶段花费的聚类时间,即事先已知材料的细观构型对其力学行为进行求解时,其一次在线计算的时间仅为6 s,计算效率比传统有限元提升了104倍,在结构性能快速设计、结构状态快速预报等领域,有着广阔的应用前景。In this paper,the self-consistent clustering analysis(SCA)method was used to investigate the progressive damage behavior of 2D C/SiC under uniaxial compression load.The SCA method clusters the grid elements by strain concentration tensor,which greatly reduces the degree of freedom of the model and improves the computational efficiency of the model without significantly reducing the computational accuracy.The whole method consists of two stages:Offline and online.In the offline stage,the k-means algorithm was used to decompose and cluster the high-fidelity composite unit cells and calculate the interaction tensor between different clusters,and finally the reduced-order model was generated.At the online stage,the mechanical response was obtained by solving the discrete Lippmann-Schwinger equations based on the reduced-order model.The SCA method was applied to predict the compressive strength of 2D C/SiC.When the total number of clusters is 64,compared with the experiment,the calculation accuracy of the compressive strength solution is reduced by 1%compared with the traditional finite element method,but the overall calculation efficiency is improved by 34 times.When the clustering time spent in the offline stage is not considered,that is,the meso-structure of the material is known in advance to solve its mechanical behavior,the time of one-time online calculation is only 6 s,and the calculation efficiency is 104 times higher than that of the traditional finite element method.It has broad application prospects in the fields of rapid design of structural performance and rapid prediction of structural state.

关 键 词：2D C/SiC 自洽聚类分析 渐进损伤 压缩强度 细观模型 

分 类 号：TB332[一般工业技术—材料科学与工程]