High‐fidelity random fiber distribution algorithm based on fiber spreading process

Abstract

AbstractThe performance of fiber‐reinforced composites is not only determined by their constituents, but also by the spatial distribution of inclusions. The traditional random distribution algorithm only considers the randomness of its relative position, and every newly generated fiber has to be checked with all the existing fibers, which requires a large amount of calculation. In this paper, inspired by the spreading process that the fibers are not arranged randomly, but the forces acting on the upper and lower rows of fibers. A random fiber distribution algorithm based on partition randomness is proposed, and a microscopic model of in‐plane fibers with random distribution is established. The maximum volume fraction of fibers generated by the algorithm is 75%, and the generation time is less than 20s. Statistical characterization and finite element analysis of the structure generated by the algorithm are compared with the experimental data to verify the randomness and effectiveness of fibers. The algorithm overcomes the jamming limit of the traditional hard‐core method, and can efficiently generate random microstructure models with different volume fractions, which can be a promising alternative for modeling and analysis of unidirectional fiber‐reinforced composites.