摘要
碳纤维增强复合材料因其优越的力学性能,一直以来是材料科学的研究热点.常规的多尺度方法因其巨大的计算量很难对该类材料的研究提供实质性帮助.非均匀变换场分析是针对弹塑性多尺度分析十分有效的降阶均匀化方法.然而,由该方法推导出的降阶模型存在普适性不强,应用门槛较高的缺陷.为此,将非均匀变换场分析法与k-means聚类算法有机结合起来,提出了改进版降阶模型.用户可以根据自身需求将细观本构模型嵌入到本文降阶框架中,无需再次对降阶模型进行独立推导.基于聚类非均匀变换场分析,揭示材料细观塑性应变场演化对宏观响应的影响规律,对碳纤维增强复合材料力学性能进行预测.算例结果表明,新降阶模型能够准确地预测复合材料的宏观力学性能,其相较于传统有限元计算的加速率可达惊人的10^(3)-10^(4)量级.
Carbon-fiber-reinforced composites have attracted much attention in materials science due to their superior mechanical properties.On a structural level,such materials are often considered as homogeneous.However,the homogeneity is lost on a smaller scale such as microscale.This micro-heterogeneity significantly affects the macroscopic performance of the composites.The multiscale methods provide a powerful tool for an in-depth study of such materials.However,it is difficult for conventional multiscale methods to provide substantial assistance for the research of such materials due to their huge computational costs.The nonuniform transformation field analysis is an effective reduced order homogenization method for elastoplastic multiscale analysis.However,the reduced-order model in such a theory has to be separately derived for different kinds of microscale constitutive models.For higher universality,we propose an enhanced reduced order model by combining the nonuniform transformation field analysis with a k-means clustering algorithm in frame of the so-called CNTFA.Through clustering analysis,the microscopic representative volume element(RVE)is divided into a given number of subdomains to reduce the number of degrees of freedom,and thus improving the computational efficiency of the RVE.As a merit,one can easily embed the required microscopic constitutive model into the reduced order homogenization framework without the need to derive a new reduced order model.Based on the cluster-based nonuniform transformation field analysis,the influence of the microscopic plastic strain field evolution on the macroscopic material response is captured,while the mechanical responses of carbon-fiber-reinforced composites are predicted.The numerical results show that the proposed reduced order model can accurately predict the macroscopic mechanical responses of composite materials,with a striking acceleration rate of 103-104 compared with the conventional finite element(FE)computation.Finally,the convergence of related FE meshes and the influences of carbon-fiber content and spatial distribution on the macroscopic mechanical properties of the considered composites are numerically studied.
作者
江祎楠
陈娟娟
梁利华
许杨剑
鞠晓喆
Yinan Jiang;Juanjuan Chen;Lihua Liang;Yangjian Xu;Xiaozhe Ju(College of Mechanical Engineering,Zhejiang University of Technology,Hangzhou,310023;Zhejiang Huayun Information Technology Co.,Ltd.,Hangzhou,310000)
出处
《固体力学学报》
CAS
CSCD
北大核心
2023年第5期648-658,共11页
Chinese Journal of Solid Mechanics
基金
国家自然科学基金青年科学基金项目(12002309)
国家自然科学基金面上项目(51875523,52275164)
浙江省自然科学基金探索项目(LQ21A020002)资助
关键词
碳纤维增强复合材料
降阶模型
多尺度方法
聚类分析
非均匀变换场分析
carbon-fiber-reinforced composites
reduced order model
multiscale methods
clustering
nonuniform transformation field analysis
作者简介
通讯作者:鞠晓喆.E-mail:xju@zjut.edu.cn.
