Variation of yarn fiber volume fraction, induced by the compression between adjacent yarns during the manufacturing process of textile composites, is difficult to be determined by using a single imaging method. A method combining scanning electron microscopy and micro-computed tomography is proposed to quantify the variation of yarn fiber volume fraction of textile composites, which is decomposed into systematic trend and stochastic deviation. The method takes the advantages of high resolution of scanning electron microscopy and wide 3D view of micro-computed tomography. Average fiber cross-sectional areas are acquired by analyzing hundreds of fiber cross-sectional areas in scanning electron microscopic images. Yarn cross-sectional area is determined by fitting ellipse to the labeled yarn cross-section in slices of micro-computed tomography images. The results of E-glass/epoxy and carbon/epoxy specimens show that their systematic trends of yarn fiber volume fraction combined with standard deviations of stochastic deviation, relative to the respective global means, fluctuate between [−11.4%, 15.3%] and [−12.9%, 10.7%], respectively. Yarn FVF varies in specimen obviously and needs to be considered in mechanical property prediction.

No abstract

Traditional equivalent inclusion method provides unreliable predictions of the stress concentrations of two spherical inhomogeneities with small separation distance. This paper determines the stress and strain fields of multiple ellipsoidal/elliptical inhomogeneities by equivalent inhomogeneous inclusion method. Equivalent inhomogeneous inclusion method is an inverse of equivalent inclusion method and substitutes the subdomains of matrix with known strains by equivalent inhomogeneous inclusions. The stress and strain fields of multiple inhomogeneities are decomposed into the superposition of matrix under applied load and each solitary inhomogeneous inclusion with polynomial eigenstrains by the iteration of equivalent inhomogeneous inclusion method. Multiple circular and spherical inhomogeneities are respectively used as examples and examined by the finite element method. The stress concentrations of multiple inhomogeneities with small separation distances are well predicted by equivalent inhomogeneous inclusion method and the accuracies improve with the increase of eigenstrain orders. Equivalent inhomogeneous inclusion method gives more accurate stress predictions than equivalent inclusion method in the problem of two spherical inhomogeneities.