Fibre architectural statistics-based modelling of unidirectional composites facilitated by X-ray computed tomography

In this study, we developed a statistics-based modelling approach that considers the actual fibre architecture of unidirectional composite materials and validated its accuracy. Initially, fibre centre coordinates and misalignment were extracted from a three-dimensional micro-computed tomography image. Subsequently, we confirmed that the fibre arrangement aligns with a random distribution, and the fibre misalignment follows a normal distribution, providing statistical data for setting up the numerical model. The fibres were generated in MATLAB based on the statistics via reverse modelling. Finally, the comparison between the finite element model (FEM) with fibre misalignment and the FEM with fully straight fibres verified the feasibility of the proposed approach.