Statistical simulation of fiber fragmentation in a single-fiber composite

Abstract

A probabilistic model of fiber fragmentation in a single-fiber composite is presented. The model takes into account the random nature of both fiber and interfacial strengths. The Monte Carlo technique has been chosen for computational realization, and the interface is assumed to be elasto-plastic. In contrast with existing solutions, which use a random generation of flaw locations (or fragment lengths), the present approach is based on the generation of random fiber and interfacial strengths along the fiber and interface, respectively. Numerical calculations of the fragment length distributions are presented as functions of the statistical characteristics of the fiber and interfacial strengths. A significant influence of the fiber strength variability on the distribution and the average value of fragment lengths is shown. In general, the variability of the interfacial strength affects the fragment length distribution as well. A strategy for the measurement and calculation of the interfacial strength distribution from the single-fiber fragmentation test is presented. The strategy is illustrated by using a series of single-fiber glass glass/epoxy composites with five different fiber surface treatments. Statistical parameters for the interfacial strength are evaluated by using a set of dimensionless graphs and tables.