STOCHASTICITY MODELING AND ANALYSIS OF ELASTIC MODULUS OF RANDOMLY ORIENTED STRANDS
Yi Wan, Xin Zhang, Yu Sato, Jun Takahashi
Abstract
Randomly oriented strands (ROS) exhibit multiscale inhomogeneous and significant stochastic variations even in elastic properties because of their complex internal geometries. To quantify the variations and expand the applicability of ROS in industrial fields, a laminate analogy analytical model was developed in combination with Monte Carlo simulation-based statistical analysis to compute the mean and variance of the elastic modulus of ROS. Two semi-empirical fitting equations were formulated to calculate the mean and variance of the elastic modulus, thus reducing the need for onerous experiments. The effects of strand morphology on the stochasticity of the elastic properties are discussed in detail for the first time. The results of the analytical model showed significant applicability and good agreement with the experimental data under various conditions, and the formulated equations were able to provide satisfactory values for the mean and variance of the elastic modulus with different strand properties and strand/specimen morphologies. The role of the strand morphology (thickness, width, and length) was analyzed from a statistical perspective and the thickness ratio of the strand to the specimen was found to be the most effective factor for reducing the stochasticity of ROS.