Multiscale modeling of randomly interwoven fibers for prediction of KM2 Kevlar yarn strength and damage

Modeling Kevlar yarn response as a function of twist requires creating a model at the filament level that incorporates capturing the mechanical interaction of numerous fibers. The inherent complexity of building a multiscale interwoven fibrous structure manually is prohibitive for such an endeavor; therefore, computer-aided simulations are preferred. In this study, a random walk methodology was employed to generate a fibrous structure along the axis of a yarn. Since the directionality of the fibers is randomly oriented along the axis, the fibers can wind around each other and tangle or terminate on demand. The resultant geometry can represent the tortuous path that yarn filaments experience. Yarn twist can be introduced through imposing a rotation matrix to the geometry or conducting an initial analysis that applies the preload. The analysis method employed in this paper captures the correct prestress of the twisted yarns at zero, three, and ten twists per inch. The analysis then loaded the yarn until filament fracture occurred. The predicted ultimate load was within 5% for all three twists per inch analyzed. The zero twists per inch linear response matched test results to within 5%.