Mesoscale ballistic damage mechanisms of a single-layer woven glass/epoxy composite

Highlights

• Single-layer composite isolates perforationphase and facilitates optical and X-ray CT mesoscale damage characterization.</span>

• Experimentsshowed more energy required to perforate tow crossovers than interstitialregions of woven composite.</span>

• Transversedeformation wave velocity, time-resolved back-face deflection, and mesoscaledamage mechanisms observed.

• The data in this paper are useful for validating mesoscale finite element models of woven fabric composites.

Abstract

Understanding damage mechanics across multiple length scales is necessary for design of enhanced protection by composite armor systems. In this work, we investigated the perforation mechanics of a single layer, woven composite target transversely impacted below and above the ballistic limit by a rigid projectile sized on the order of a tow width. To visualize mesoscale damage mechanisms of woven composites, a thin translucent composite target was used providing access to both impact and back-face surfaces. High-speed video, high-speed digital image correlation, high-resolution photography, and X-ray computed tomography were used to gather data. Impact and residual velocity data, ballistic limit velocity, and projectile impact location relative to the weaving architecture were quantified. It was found that impact on a tow-tow crossover requires more energy to perforate than impact on a matrix-rich interstitial site or on adjacent, parallel tows. Dynamic deflection and the transverse deformation cone wave-front velocity of the thin composite were measured. Meso-mechanical damage in thin, woven composites was characterized for impact velocities below and above the ballistic limit. Four mesoscale damage modes were identified: transverse tow cracks, tow-tow delamination, 45° matrix cracks, and punch-shear. These damage modes were observed both on the surface and inside the composites. The data presented here are useful for mesoscale modeling of woven composites, and this work provides new insight into meso‑mechanical modes of damage in thin composite structures under dynamic impact.

Introduction

Woven thermoset composite materials typically possess excellent mass efficiency and are therefore widely used in ballistic protection applications [1], [2], [3]. Understanding the ballistic impact damage modes and mechanisms of composite armor is important for the design of improved protection by armor systems. Experimental studies have investigated the ballistic response of woven thermoset composites of various thicknesses to normal projectile impact, and experimental results are often used in subsequent continuum simulations. Jordan et al. [4] conducted quasi-static punch-shear testing, low-velocity impact, and ballistic impact of 7.52 mm diameter steel right circular cylinders (RCC) into plain weave (PW) E-glass/phenolic laminates as thick as 50 mm (100 layers) to as thin as 4 mm (8 layers). Jordan et al. [5] then used these data to parameterize the MAT162 composite model in LS-DYNA, and conduct continuum simulations of these impacts. Gama and Gillespie [6] impacted 12.7 mm (22 layers) and 20.6 mm (33 layers) PW S-2 Glass/SC-15 composite plates with 12.7 mm diameter steel RCCs at velocities below and above the ballistic limit velocity, and used these data and MAT162 and LS-DYNA to numerically investigate damage evolution at different phases of ballistic penetration. Using steel 7.62 mm diameter conical-nosed and 9 mm diameter hemispherical-nosed projectiles, Gower et al. [7] impacted Kevlar® panels as thick as 18.5 mm (37 layers) and as thin as 9.5 mm (19 layers), and they used backplane displacement-time data to evaluate their LS-DYNA numerical results. Zhu et al. [8] conducted quasi-static and dynamic experiments of 6.35 mm, 9.525 mm, and 12.7 mm diameter cylindrical steel projectiles with conical and blunt noses impacting Kevlar® panels as thick as 15 mm (24 layers) and as thin as 3.1 mm (5 layers). The thinnest Kevlar® targets, 3.1 mm (5 layers), were impacted by 60° conical nose 12.7 mm diameter steel projectiles with velocities below and above the ballistic limit, and Zhu et al. [8] showed that blunt-nosed projectiles have greater ballistic limit velocity than conical-nosed projectiles. Silva et al. [9] impacted Kevlar® panels, 2.4 mm (7 layers) thick, with blunt-nosed, cylindrical fragment simulating projectiles (diameters not specified) at velocities below and above the ballistic limit, and found that AUTODYN numerical simulations with a calibrated material model accurately estimate the ballistic limit velocity, residual velocity, and global damage. Using 6.36 mm diameter steel RCC projectiles, Pandya et al. [10] impacted hybrid composites of PW E-glass and satin weave T300 carbon fabric and epoxy matrix that were as thin as 2.5 mm, and determined ballistic limit velocity and global damage patterns for various hybrid configurations. Gellert et al. [11] studied woven E-Glass/vinylester composites as thick as 20 mm (44 layers) impacted by steel 6.35 mm diameter blunt-nosed cylinders, 4.76 mm diameter blunt-nosed cylinders, and 4.76 mm diameter 90° and 45° conical-nosed cylinders. Gellert et al. [11] define a thick target as one that exhibits both indentation and “dishing”, which is a cone of deforming material opening outward in the direction of projectile travel. Gellert et al. [11] define a thin composite target as one in which the only perforation mechanism is dishing, but the thinnest composite target they perforated was 4.5 mm (11 layers).

All of the above experimental studies of ballistic impact and perforation tested multi-layer composites. Lomov et al. [12] noted that single-layer composites are seldom tested (or used), but local variations due to lamina nesting, random tow placement, and interpenetration and overlapping of tows in multi-layer composites contributes to error in strain field characterization. The work by Lomov et al. [12] suggests that experiments on a single layer can provide insight into the material behavior by minimizing sources of error related to the interaction between multiple plies. The above experimental studies also concentrated on macroscale damage, and conducted numerical simulations in which individual woven composite layers were homogenized. While a homogenized continuum approach may be appropriate for modeling macroscale damage, it cannot capture mesoscale damage modes associated with the mesostructure. However, mesoscale modeling of composites is becoming more common, for example [13], [14], [15], [16], [17], [18], [19]. Cunniff [20] and Novotny et al. [21] studied ballistic impact of single-layer fabrics, but these were unreinforced fabrics and matrix reinforcement provides additional complexity to the perforation mechanics and damage mechanisms. The present work is concerned with impact induced, mesoscale damage mechanics in polymer reinforced woven fabric composites, so a single-layer, translucent composite target, which provides access to both impact and back-face surfaces, was used to visualize the mesoscale damage in an impacted woven composite.

Length scales associated with the weaving architecture of woven fabric composite materials play an important role in the energy dissipation mechanisms of these composites. Under transverse ballistic impact, compared to laminated, unidirectional composites, the weaving architecture of woven composites provides additional mesoscale energy dissipation mechanisms including primary tow matrix cracking, tension-shear and compression-shear tow failure, elastic deformation of secondary tows [22], additional delamination resistance due to the out-of-plane undulation [18], tow-tow and tow-matrix delamination [17], and coupled tension and bending due to local undulation of interwoven tows [12]. Mesoscale damage modes are comprised of microscale damage mechanisms such as fiber-matrix interfacial failure, fiber fracture, and any matrix damage smaller than a tow thickness.

Except for Gellert et al. [11], the above experimental studies of glass fiber reinforced polymer composite panels used projectiles that were larger in diameter than the width of a single tow. Impactor dimension relative to the composite architecture and impactor nose shape (e.g., conical, hemispherical, blunt) influence how kinetic energy is transmitted into the composite. Since mesoscale damage occurs at the length scale of a tow width, an impactor diameter on the order of a tow width was considered in the present work. Fig. 1 shows several conventional projectile calibers relative to a standard S-2 glass fabric [23] made with nominally 5 mm (0.2 in) wide tows. A projectile diameter less than a fabric tow width may impact on individual tows, tow-tow crossovers, and interstitial matrix pockets, however, larger diameter projectiles will impact on multiple tows. The present study considers a 5.6 mm (0.22 in) right circular cylindrical (RCC) blunt-nosed projectile. The diametric center of this projectile can impact the woven composite within either region A or B in Fig. 1. An impact centered in region A, a tow-tow crossover, is hereafter termed “center impact” and an impact centered in region B, an interstitial matrix pocket or an individual tow, is termed “off-center impact,” referring to impacts centered (or not) on primary tows (defined by Naik and Shrirao [20] as the tows directly beneath the impacting projectile).

In previous works, Haque et al. [5], [6], [24] described ballistic impact response in terms of three phenomenologically different phases: (1) penetration; (2) transition; and (3) perforation. Each of these phases is dominated by different deformation and damage mechanisms. Applying these phases to the present work, a thick composite is herein defined as one in which penetration, transition, and perforation occurs, and a thin composite is defined as one in which only perforation occurs.

Penetration and perforation damage mechanisms below and above the ballistic limit velocity are fundamentally different. Below the ballistic limit, the extent of damage increases with increasing impact velocity (kinetic energy), and the extent of damage is greatest at the ballistic limit. Above the ballistic limit, the extent of damage reduces with increasing impact velocity because the damage mechanisms occur under high energy density beyond the ballistic limit. Impact velocities greater than the ballistic limit exceed a target's capacity to dissipate energy without catastrophic failure in the neighborhood of the projectile, so the projectile perforates and exits the target with some residual velocity in the direction of impact. All transverse impacts with projectile velocities exceeding the ballistic limit include the perforation phase, regardless of target thickness. Therefore, improving protection by armor systems and understanding ballistic impact damage necessitates studying the perforation phase.

To isolate and examine the damage modes and mechanisms present in the perforation phase up to and exceeding the ballistic limit velocity, the present study uses a single layer woven composite plate. A single-layer composite ensures that neither the penetration nor transition phases will be present, but only the perforation phase will occur. To investigate the perforation dominated mesoscale damage modes and mechanisms associated with thin woven fabric composites, ballistic impact experiments are conducted using a blunt-nosed projectile scaled on the order of a tow width. To investigate kinetic energy transfer and mesoscale damage, impact velocities up to and exceeding the ballistic limit velocity are recorded, impact locations relative to tow-tow crossovers are determined, the transverse cone wave-front velocity is calculated from high-speed video, dynamic deflection of the target back face is measured, and mesoscale damage is characterized using high-resolution optical photography, X-ray computed tomography, and confocal microscopy.