摘要
采用数字图像相关技术(digital image correlation,DIC)结合有限元模型修正技术(FEMU),通过2维编织高铝纤维增强多孔气凝胶基复合材料正轴及偏轴拉伸实验,同时识别获得了材料面内多个工程弹性常数.通过力学实验结合有限元数值虚拟实验,研究了识别过程中目标函数构成、优化方法、参数初值、位移场随机误差水平对识别结果和效率的影响.研究表明针对2维编织多孔气凝胶基复合材料,选取DIC实测位移和有限元数值计算位移的方差作为目标函数,通过L-M非线性最小二乘优化方法,计算参数敏感度矩阵,可同时识别获得多个加载面内材料力学性能参数,识别效率高,识别过程对参数初值、位移场随机误差水平不敏感,鲁棒性好.
Using digital image correlation technology(DIC)and finite element model correction technique(FEMU),multiple mechanical engineering constants in the two-dimensional woven high aluminum fiber reinforced porous aerogel matrix composites were identified through on-axis and off-axis tensile tests.The effects of objective function construction,optimization method,initial value of parameter and random error in the measured displacement field on the identifying results and efficiency were investigated by mechanical experiments and finite-element-based numerical virtual experiments.The results show that,for two-dimensional woven porous aerogel matrix composites,when the variance of the DIC measured displacement and finite element calculated displacement is chosen as the objective function and the L-M nonlinear least-square optimization method is used to calculate the sensitivity matrix,the multiple mechanical properties in the loading surface can be identified simultaneously with high efficiency.The proposed method is insensitive to the differences in initial values of material constitutive parameters and random error in the DIC displacement measurement and has strong robustness.
作者
刘刘
姬晓慧
郝自清
曲宏亮
贺体人
LIU Liu;JI Xiao-hui;HAO Zi-qing;QU Hong-liang;HE Ti-ren(School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, China;Beijing Key Laboratory of Commercial Aircraft Structures and Composite Materials, BASTRI, COMAC, Beijing 102211, China)
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2020年第10期1033-1042,共10页
Transactions of Beijing Institute of Technology
基金
国家自然科学基金面上项目(11472043)
国家重大专项资助课题。
关键词
多孔气凝胶基复合材料
数字图像相关技术
有限元模型修正
参数识别
L-M非线性最小二乘优化
aerogel-based matrix composites
digital image correlation technology
finite element model updating
parameter identification
L-M nonlinear least-square optimization
