279 mm by 279 mm. All specimens were aligned with the warp direction horizontal. For each single-ply specimen, 15 shots were taken. This pattern was designed such that none of the principal yarns for each of the 15 shots were shared, and the spacing between each impact, and between each impact and the inner edge of the target frame, was at least 50.8 mm. A pair of light chronographs measured the impact velocity and a light bar chronograph measured residual velocity, which helped to identify penetrating or perforations versus non-penetrating impacts (“stops”) in cases when it was otherwise unclear solely from the damage to the fabric. Perforation regions were examined using scanning electron microscopy (SEM) after testing. Failures occurred via fiber breakage as well as pull-out, as pictured in Fig. 7.6c. V50, the velocity at which the probability of penetration is estimated to be 50 %, was estimated using velocities from 10 shots. For each calculation, half of the values were the highest non-penetrating shot velocities and half of the values were the lowest penetrating values. In each case, the correlating V50 span was calculated as the difference between the highest non-penetration and the lowest penetration. The V50 for treated (1.6 wt%) Kevlar was nearly double (214 m/s) that of the untreated Kevlar (110 m/s), suggesting that yarn pull-out resistance may be a critical factor under the ballistic conditions evaluated. This example illustrates the utility of the yarn pull-out tests for down-selecting to the best armor material synthesis conditions for full scale testing. 7.3 Modeling and Simulation Models are being used to identify candidate mechanisms responsible for observed changes in material performance and to understand how desirable macroscopic properties can be realized by modifying the underlying molecular and fibril structures. These models are being employed from the molecular and fibril level (molecular dynamics simulations) to yarn and fabric level (FEM models) to elucidate how the hierarchical structure and properties at different length scales contribute to the system performance. Molecular Dynamics (MD) simulations help to show the relationship between the addition of carbon to the Kevlar fibrils and the mechanical properties of the fibers themselves [16]. Simulated and experimental tensile data can be compared, and by understanding how the material properties and structure interact at different length scales, we can begin to understand the mechanisms responsible for improved ballistic performance at the macroscale. 7.3.1 Pull-Out Model Since yarn pull-out is an important mechanism that is active during ballistic impact, it must be accurately represented in the FEM ballistic model. A FEM pull-out model uses experimentally derived parameters such as strength and stiffness from tensile tests, and friction coefficients from the static friction test and yarn pull-out test. Simulations of the pull-out experiment, pictured in Fig. 7.7a, can be validated against experimental data (Fig. 7.7b), and enable verification that friction and yarn interaction are accurately represented in the ballistic model. A preliminary effort towards validation shows promising results out to 4 mm of displacement, and expansion of these simulations to greater length scales for improved comparison with experimental data is underway. Following completion and validation of these models, parametric studies can elucidate the relationship between friction and ballistic performance. 7.3.2 Single-ply Model Analogous to the experimental case, a single-ply sheet of fabric modeled in quarter or half symmetry using LS-DYNA can enable simpler testing of material properties and performance. The single-ply system, impacted at a lower-range of velocities, allows for comparison of either the residual velocity of projectiles that penetrate the fabric, or the achievable ballistic limits of the fabric as the material properties are changed. An orthotropic elastic constitutive model was utilized, allowing for the fabric to be characterized as strong along the length of the yarn and comparatively weak in all other directions. Input material properties and failure characteristics for the orthotropic constitutive model were extracted from tensile tests, static friction and pull-out tests, pull-out models, and molecular models. The model was partitioned into three regions (Fig. 7.8a) as described in detail by Thomas et al. [16]: a high-resolution impact region, a medium-resolution shell region, and a low-resolution global membrane region. Figure 7.8b shows a simulation image of a single ply of Kevlar 129 7 Multi-scale Testing Techniques for Carbon Nanotube Augmented Kevlar 65
RkJQdWJsaXNoZXIy MTMzNzEzMQ==