Ce Xiao, Qiang Chen*, Haoming Zhang, Haoming Luo, Zun Yin, Jun Ren, Zhonghuai Su,
3D in-situ X-CT characterization of anisotropic damage in glass-fiber–reinforced polypropylene,
International Journal of Mechanical Sciences,
2026,
111368,
ISSN 0020-7403,
https://doi.org/10.1016/j.ijmecsci.2026.111368.
(https://www.sciencedirect.com/science/article/pii/S0020740326002249)
Abstract: Glass fiber-reinforced thermoplastics exhibit strong anisotropy due to fiber alignment induced by injection molding. This study develops an experiment-informed framework to interpret damage initiation and evolution by coupling quasi-static tensile testing inside an X-ray tomograph with digital volume correlation (DVC) and microstructure-informed finite element modeling. Longitudinal and transverse specimens were imaged at multiple strain levels. A U-Net model segmented fibers, voids, and matrix, enabling quantitative analysis of fiber orientation tensors and length distributions, which were then used to construct periodic representative volume elements (RVEs) with cohesive interfaces. DVC linked localized strain fields to damage behavior. Macroscopically, the longitudinal specimens exhibited higher tensile strength and elongation than the transverse ones, attributed to differences in fiber orientation. In-situ 3D observations revealed orientation-specific failure mechanisms: transverse specimens were dominated by interfacial debonding and planar cracking, while longitudinal specimens exhibited fiber pull-out and fracture. CT-informed statistical periodic RVEs reproduced the macroscopic stress–strain responses in both orientations and localized cohesive stresses at fiber ends, consistent with shear-lag theory and critical fiber length predictions. This work provides a mechanistic and orientation-resolved understanding of failure in PP-GF30, and demonstrates a closed-loop structure–simulation–performance workflow applicable to injection-molded short fiber composites.
Keywords: Fiber-reinforced composites; In-situ X-ray computed tomography; Damage mechanism; Digital volume correlation; Fiber orientation tensor; CT-informed RVE–cohesive modeling