材料工程2026，Vol.54Issue(1)：62-73,12.DOI:10.11868/j.issn.1001-4381.2025.000049
基于同步辐射CT与深度学习的C/SiC复合材料带孔试件损伤演化分析
Damage evolution analysis of C/SiC composite specimens with holes based on synchrotron radiation CT and deep learning
摘要
Abstract
The damage evolution behavior of center-notched C/SiC composite components is investigated under uniaxial tensile loading by conducting in situ mechanical experiments based on synchrotron radiation X-ray CT observations with deep learning algorithms,3D digital image-based finite element method(3D IB-FEM),and digital volume correlation(DVC)method.The intelligent identification of different types of internal damage,including fiber bundle cracking,matrix cracking,and delamination,in center-notched C/SiC composite specimens under uniaxial tensile loading is achieved.Furthermore,the relationship between damage evolution and strain concentration is established.Deep learning-based damage identification and quantitative analysis indicate that both the center notch and initial pores influence the locations of damage initiation.The results of the three-dimensional digital image-based finite element analysis reveal the effect of initial pore geometry on crack initiation,while DVC results demonstrate the correlation between widespread strain concentration in fiber bundle regions and delamination damage,as well as final fracture.The three damage forms—fiber bundle cracking,matrix cracking,and delamination—are interrelated to some extent.As the load increases,adjacent delamination areas connect to form matrix cracks,which may further propagate into fiber bundle cracks.Under tensile loading,the primary failure modes of fiber bundles include fiber bundle fracture,fiber bundle splitting,and fiber bundle sliding.The presence of a center notch does not change the failure modes of the fiber bundles.关键词
C/SiC复合材料/X射线CT原位拉伸实验/损伤/深度学习/数字体积相关/图像有限元Key words
C/SiC composite material/X-ray CT in-situ tensile testing/damage/deep learning/digital volume correlation/image-based finite element分类
通用工业技术引用本文复制引用
刘乔雨,王龙,侯传涛,李志强,刘武刚..基于同步辐射CT与深度学习的C/SiC复合材料带孔试件损伤演化分析[J].材料工程,2026,54(1):62-73,12.基金项目
国家自然科学基金(52202082) （52202082）
