Date of Award

Spring 2020

Document Type

Open Access Thesis


Mechanical Engineering

First Advisor

Subramani Sockalingam


High performance fibers such as ultrahigh molecular weight polyethylene (UHMWPE) are often used for ballistic impact applications in the form of textile fabrics and composite laminates. In order to predict the ballistic performance of such materials, single-fiber experiments are performed to quantify the material behavior at smaller length scales, which can be applied to larger length scales as a result. Failure of UHMWPE is well understood as a function of simple tension at low and high strain rates, as well as under various multiaxial loading states. However, experimental characterization of single UHMWPE fibers under transverse loading at high strain rates (4000-7000 s-1) has not yet been performed due to the lack of available methodology.

In this work, a single fiber transverse impact experimental technique is developed at the Army Research Laboratory (ARL) labs. A small-diameter Hopkinson bar is modified to launch custom-designed loading geometries on individual fibers mounted transversely to the path of motion. Load cells at the grips record forces experienced by the fiber, and a high framerate camera captures the test progression and deformation behavior. Loading geometries are all circular with varying radius including a razor (~2 μm), a sharp indenter (20 μm), and a blunt indenter (200 μm), and two impact velocities are chosen, 10 m/s and 20 m/s, which correlate to strain rates of approximately 4320 and 6846 s-1.

This novel apparatus and experimental design is used to study the transverse impact behavior of UHMWPE Dyneema® SK76 single fibers with average diameters of 17 um. Failure strain for all groups is significantly reduced relative to existing tensile and quasi-static (QS) transverse loading data. For all the geometries, failure strains are reduced by 46-51%, compared to QS tensile and 12-19% compared to QS transverse, as strain rates increased from 4320-6846 s-1. Compared to high strain rate (1156 s-1) tensile failure strain, significant reduction in failure strains are measured due to transverse impact loading. Failure strains (i) reduced by 28-34% for blunt impact at strain rates 4369-6952 s-1; (ii) reduced by 32-39% for sharp impact at strain rates 4285-6797 s-1 and (iii) reduced by 58-61% for razor impact at strain rates 4307-6789 s-1. For all the geometries, change in strength ranges from +6% to -2%, compared to QS tensile, as strain rates increased from 4320-6846 s-1. Compared to high strain rate tensile strength, changes in strength can range from a slight increase to a significant reduction due to interactions between the rate-dependent increases in stiffness and strength, and strength degradation due to transverse loading. Strength measurements (i) range from +6% to -2% for blunt impact at strain rates 4369-6952 s-1; (ii) range from +4% to -8% for sharp impact at strain rates 4285-6797 s-1 and (iii) range from -28% to -42% for razor impact at strain rates 4307-6789 s-1. The reduction in tensile properties are attributed to the failure mechanism induced by different geometries. While all geometries induce axial compression due to the impact, the loading radius affects the degree of applied transverse shear, where little to no transverse shear is observed in the blunt indenter, an intermediate amount of shear is applied in the sharp indenter, and a high degree of shear is applied by the razor indenter. This conclusion is supported by failure surface images, where blunt impact results in fibrillation characteristic of tensile failure, razor impact results in fiber shearing characteristic of the cutting action of the razor, and the sharp impact demonstrates a mixed amount of both failure modes.

The experiments are modeled in LS-DYNA using a custom user material model (UMAT) to incorporate nonlinear inelastic transverse compressive behavior. Model predictions correlate well to the experimental observations in terms of load and strain values as well as in qualitative characterization of the material response to impact loading. A previously-developed strain-based single fiber multiaxial failure criterion is discussed and applied to the model output, but more development is necessary for this criterion to have predictive capabilities for high strain rate impact of UHMWPE.