基于应变场数据的C/C材料力学参数辨识
Identification of mechanical parameters of C/C composite basedon strain field data
Received:August 20, 2021  Revised:January 24, 2022
DOI:10.7520/1001-4888-21-190
中文关键词:  C/C复合材料; 参数辨识  奇异值分解; 数字图像相关
英文关键词:C/C composite material  parameter identification  singular value decomposition  digital image correlation
基金项目:一院高校联合创新基金(CALT基金);国家自然科学基金青年科学基金(11902101)资助
Author NameAffiliation
GAO Bo National Key Laboratory of Science and Technology for National Defense on Advanced Composites in Special Environments, Harbin Institute of Technology, Harbin 150001, Heilongjiang, China 
GONG Wenran Beijing Institute of Structure and Environment Engineering, Beijing 100076, China 
WANG Long Beijing Institute of Structure and Environment Engineering, Beijing 100076, China 
SU Yunquan Beijing Institute of Structure and Environment Engineering, Beijing 100076, China 
YANG Qiang* National Key Laboratory of Science and Technology for National Defense on Advanced Composites in Special Environments, Harbin Institute of Technology, Harbin 150001, Heilongjiang, China 
Hits: 589
Download times: 86
中文摘要:
      提出了一种基于复杂构型试样DIC(Digital Image Correlation,数字图像相关法)全场应变测试结果,实现单次试验辨识C/C复合材料面内多个力学参数的方法。在该参数辨识方法中,采用复杂构型试样以使得全场应变数据对不同的材料参数敏感;建立与试验条件同等的有限元模型,将待辨识的材料力学参数作为设计变量,应变场模拟结果与试验结果的偏差视为目标函数,通过多岛遗传优化方法构建了参数辨识模型。采用奇异值分解对试验获取的应变场进行重构,实现DIC应变场测试结果的降噪。结果表明,尽管试验获取的应变场由于试验噪声、纤维变形等原因与有限元模拟的分布特征有所差异,但在加载方向的应变辨识精度偏差依然保持在4%以内。采用奇异值分解可提升与加载方向相垂直方向的弹性模量辨识精度,与标准测试结果相比,辨识误差从33.12%降低到4.48%。上述研究可为建立复合材料高通量试验方法提供参考。
英文摘要:
      In the paper, a method to identify multiple mechanical parameters of C/C composite by a single test, based on the results of full-field strain measurement of the sepcimen with complex configuration by DIC (Digital image correlation), is proposed. The complex configuration specimen is used to make the full-field strain sensitive to different material parameters; The finite element model is built under the same test conditions, the material mechanical parameters to be identified are taken as the design variables, and the deviation between the simulated strain field results and the test results is regarded as the objective function. The parameter identification model is constructed by multi-island genetic algorithm, and the singular value decomposition is used to reconstruct the strain field obtained from the test to reduce the noise. The results show that although the strain field obtained from the test is different from the distribution characteristics of the results of finite element simulation, the error of strain identification accuracy in the loading direction remians within 4%. Singular value decomposition can improve the identification accuracy of elastic modulus perpendicular to the loading direction. Compared with the standard test results, the identification error is reduced from 33.12% to 4.48%. The above research provides a reference for the establishment of high-throughput test methods for composites.
View Full Text  Download reader
Close