Modeling Random Fiber Networks in Hyperelastic Composites

Resumo

The mechanical behavior of elastomers reinforced with random fiber networks (RFNs) depends strongly on the microstructural organization. This work presents a computational approach for generating three-dimensional RFNs using a persistent random walk with orientation via the von Mises-Fisher distribution. The algorithm incorporates curvature control, fiber smoothing, and a self-avoidance criterion with collision detection to avoid fiber overlap. Fiber geometries are generated according to a target volume fraction through an iterative adjustment of the number of fibers and are exported for finite element analysis. The geometries of polyester fiber modeled as linear-elastic REINF264 elements in ANSYS are embedded into a hyperelastic RTV silicone rubber matrix modeled by a two-term Mooney-Rivlin formulation. Fibers were modeled as linear elastic reinforcement elements using REINF264 elements. Seven RVE sizes from 1 mm to 4 mm were tested under uniaxial deformation, and homogenized Cauchy stresses were computed to obtain the second Piola-Kirchhoff stress tensor and the effective material elasticity component C_1111. Results indicate convergence of the elastic response for L=3.5mm, suggesting this as the minimum RVE size for representing the given microstructure.