CHARACTERIZATION OF ORIENTATION CLUSTERING IN SHORT-FIBER COMPOSITES

During flow of concentrated fiber suspensions, the proximity of the fibers induces interactions in their orientation space. These fiber‐fiber interactions may lead to clustering in orientation of fibers and a nonhomogeneous fiber orientation field. An orientation cluster is a group of fibers in a specific neighborhood sharing the same orientation state. It is important to characterize clustering to understand and quantify fiber‐fiber interactions. Such characterization will also be useful in the prediction of the mechanical and physical properties of the resulting composite material. An image of the fiber distribution at a location in a composite is divided into a number of partitions, and the variance of the fiber orientation states within the partitions, between the partitions and of the entire sample are compared to establish the existence and the extent of clustering in the orientation space. Statistical methods such as analysis of variances and techniques analogous to the scale of segregation used in mixing theory are modified and adopted to identify and quantify orientation clusters. Simulated examples that describe clustering quantitatively in magnitude and direction of planar orientations are presented. This method is then extended to identify clustering in model injection molding experiments performed in a center‐gated square mold with suspensions of 1% and 10% volume fractions. Copyright © 1990 John Wiley & Sons, Inc.