Identification of fiber misalignment in continuous fiber composites

Misaligned fibers are invariably present in nominally unidirectional high-performance composites. Such misaligned fibers are known to affect key mechanical properties of the composite, such as the longitudinal compressive strength, longitudinal tensile modulus, fatigue endurance, shear strength, and delamination resistance (1). In this paper we present a method for the automated detection of large angle fiber misalignment (theta > 40degrees) in continuous fiber-reinforced composite materials. The method relies on the application of a series of geometrical criteria based upon measurements routinely obtained during optical scanning of polished sample cross-sections. As such, the technique is ideal for the automated identification of highly misaligned fibers in large-area (similar tocm(2)) specimens that may contain several millions of individual fiber images. The criteria applied take into account the fact that prepared cross-sections of such materials contain many damaged fibers as a result of attrition during polishing. Data obtained from three pultruded unidirectional rods reinforced with continuous carbon filaments are used to illustrate the effectiveness of this method in identifying regions where large angle misalignment occurs.