Generation of random fiber distribution using the RSE-Monte Carlo algorithm to investigate the mechanical properties of UD composite

This work presents a new algorithm designed to generate random spatial fiber distributions in the transverse cross-section of unidirectional composite materials. The approach establishes a three-dimensional representative volume element (RVE) within a numerical framework to estimate strain, stress fields, and stiffness tensors via numerical homogenization. The process is automated using Python scripting, from geometry creation to predicting elastic properties. To manage fiber arrangement effectively, the algorithm employs a user-defined probability function to control inter-fiber distances and includes a novel method named Random Sequence Expansion Monte Carlo. This method addresses common issues such as fiber clustering and unrealistic matrix regions. The algorithm’s performance was evaluated through statistical analysis of various geometric descriptors, showing good agreement with experimental data from the literature. Finite element analysis was used to predict the elastic properties of the generated microstructure, with results closely matching experimental data and exhibiting a coefficient of variation of less than 5 %. The study confirmed the transverse isotropy of unidirectional composites, showing that the random fiber distribution in the transverse plane of fiber direction was accurately represented. Overall, the proposed numerical framework demonstrates a strong correlation with experimental outcomes, providing a useful tool for material design and optimization. Future work will focus on validation of the framework for prediction of damage onset and strength of unidirectional composite materials.