Optimization of soft armor: the response of single-ply para-aramid and ultra-high molecular weight polyethylene fabrics under ballistic impact

Abstract

Typical soft armor systems are constructed of multiple layers of a single fabric type. This empirical research sought to begin optimization of these systems through hybridization, sequencing dissimilar armor fabrics to maximize their ballistic protective performance, by first investigating single plies with a spectrum of properties to determine their behavior and response to impact. Eight individual plain weave fabrics with varying yarns and thread counts were manufactured from para-aramid and ultra-high molecular weight polyethylene (UHMWPE) yarns and physical and ballistic characterizations were conducted. The ballistic impact tests established the specific energy absorption (SEA) of each fabric across a range of impact velocities (340–620 m·s^–1) and the transverse displacement wave velocity across the rear of the fabric was found using digital image correlation. Low cover factor (C_fab) fabrics (0.74–0.84) consistently showed faster transverse wave speed than the high C_fab fabrics (0.84–0.96) for any given yarn type. The relative SEA of the fabrics varied dependent on both the impact velocity and number of plies impacted. It was found that lower C_fab fabrics had the highest SEA, critical velocity and transverse wave velocity. UHMWPE fabrics were not considered suitable for a woven hybrid system as they had a significantly lower SEA compared to all the para-aramid fabrics. Results indicate that a hybrid system, when considered as a theoretical spaced system, would benefit from higher C_fab fabrics as rearward layers. However, transverse wave results suggest the lower response of these fabrics may inhibit lower C_fab fabrics at the front of a combined hybridized system.

Keywords

soft armor ballistic impact energy absorption para-aramid ultra-high molecular weight polyethylene

Body armor is a broad term used to describe a range of protective equipment from sportswear to personal protection for police forces or armed forces. Military personal protection typically consists of hard and soft armor elements where the hard armor is present in the form of ballistic plate inserts that cover the critical bodily organs and provide protection against high-velocity rifle rounds. However, the soft armor is also a critical component. Soft armor is used in military personal protection to protect against fragments from explosive devices, such as hand grenades or mortars, and generally sits behind the hard armor and extends around the body, providing greater anatomical coverage. The necessity of such armor systems can be seen from casualty analysis data from historic and recent conflicts. Surveys of the Korean war identified that soft armor vests defeated the majority of low-velocity fragments with a 70% reduction in torso wounds when the armor was worn.¹ The recent conflict in Afghanistan highlighted the efficiency of soft armor where data from the UK/US Joint Theatre Registry 2011 Report shows that, of injuries sustained as the result of an explosive device, seven out of 10 were to the relatively unprotected extremities and only one in 10 were to the protected torso.²

In 1915, documented research by the British Bureau of Munitions identified silk as the best performing fiber of its time.³ However, the ban on Japanese silk exports to America during WWII instigated the development of alternative materials into soft armor. By the end of the war, the M1951, a service issue nylon system, was introduced marking the beginning of synthetic fiber body armors.¹ Nylon remained the prominent constituent fiber in soft armor fabric until the development of Kevlar® 29, a para-aramid yarn commercialized by Dupont in 1972. High modulus and high tenacity being advantageous properties of armor fibers, Kevlar® represented a step change in performance, showing a strength 2.5 times that of nylon and an elastic modulus an order of magnitude greater.^4–6 In 1990, ultra-high molecular weight polyethylene (UHMWPE) was subsequently introduced to market by DSM under the brand name Dyneema®, meaning strong fiber in Greek. Dyneema® has certain strength and modulus advantages over Kevlar®, but the unique properties of each mean that they both feature in today’s soft armor market. Other companies that have since introduced similar yarns include Teijin, who produce a para-aramid branded as Twaron®, and Honeywell, who produce an UHMWPE under the name Spectra®.

The hierarchy of military survivability places ‘not being hit’ above ‘being adequately protected’ and, as such, the balance between protection and agility has long been an issue in the development of personal armor systems. Its cumbersome and weighty design led to the demise of the body armor in the middle ages and its reluctant use in both world wars is testimony that the design needs to meet the needs of the user as well as the prospective threat.^7,8 It is, however, the threat that defines the type of materials that are incorporated into soft armor systems and, therefore, their net weight. The current UK military service issue armor system includes soft armor protection for the torso, neck, shoulder and groin to combat the spectrum of threats in operational theatres.⁹ Consequently, modern personal armor systems can weigh up to 16 kg, including hard armor, and may lack flexibility, inhibiting the wearer from performing their function.¹⁰ In addition to this weight, the user can be subjected to a large thermal burden if they have to wear the armor for long periods or in hot climates. Human factor studies have shown that this burden can affect a soldier’s physical and cognitive function, therefore increasing their vulnerability.^11,12 Both the UK Ministry of Defence (MOD) and US Department of Defense (DoD) have active burden reduction programs that encompass armor materials, creating pressure on industry to improve existing body armor systems.^13,14 In terms of personal protection, burden reduction is commonly associated with a reduced mass system, although increased ventilation, flexibility and fit have also been considered.⁹ A reduced mass system can either be gained from a reduction in threat or from improvements in material performance. To this end, research has not only focused on developing better, higher performing fibers for use in body armor, but has also looked at how these fibers can be formed into fabrics to improve the qualities and performance of the end product.

Soft armor is produced as a multiple layered system of approximately 20–40 ballistic protective plies.¹⁵ These plies may take the form of uni-directional (UD) laminates or woven fabrics and, in a few cases, felts. However, this research only focuses on woven fabrics. A soft armor system works by dissipating the strain induced by an impacting projectile through the thickness of the system and across each layer and the failure mode of each ply changes from front to rear.¹⁶ In the front layers of the system the strain in the fabric is often localized and yarns are sheared by the impacting projectile, creating a material plugging failure. Rearward layers tend to undergo a tensile failure, having been cushioned from the initial projectile impact. A tensile failure is marked by the propagation of strain along the impacted yarns and a transverse fabric deflection.¹⁷ Recently, hybrid systems have been appearing on the market where the composition may contain two or more different ply types.^18,19 Hybridizing the system is thought to take advantage of the changing failure mode through the thickness of the system and the layer-to-layer interactions, although relatively few studies have been conducted in this field.

This research sought to investigate the potential of hybridization through manipulation of the weave parameters and yarn type of each constituent ply. It is concerned with improving the ballistic protective performance of a multiple ply soft armor system by selectively stacking dissimilar layers within the system. The literature indicates that the holistic ballistic performance of armor systems can be influenced by the constituent fabric properties, but few studies capture trends in the fundamental impact response with these various properties,^18,20,21 with most recent studies focusing on specific fabrics or the modeling of them.^22–24

Materials and methods

The two types of yarn used during this study were Teijin Twaron® para-aramid and Dyneema® UHMWPE. A summary of the yarns used with their test references can be found in Table 1. The test reference is an abbreviation of the yarn type, yarn linear density and thread count. All fabrics from the detailed yarns were manufactured as single layer plain weaves using a Bonas Jacquard loom providing fabrics of 35 cm width. For SK/1760 yarns, 8 × 8 was the maximum achievable thread count due to difficulties and a quality issue evident above this value. Similarly, for CT/840 yarns the maximum thread count was 10 × 10.

Table 1.

Specifications of yarns and woven fabric samples

Yarn type	Reference	Yarn	Thread count warp × weft/yarns·cm^–1
Dyneema® UHMWPE	SK/1760/6	SK76 1760 dTex	6 × 6
	SK/1760/7	SK76 1760 dTex	7 × 7
	SK/1760/8	SK76 1760 dTex	8 × 8
Twaron® para-aramid	CT/840/8	CT2000 840 dTex	8 × 8
	CT/840/10	CT2000 840 dTex	10 × 10
	CT/550/10	CT2040 550 dTex	10 × 10
	CT/550/12	CT2040 550 dTex	12 × 12
	CT/550/14	CT2040 550 dTex	14 × 14

UHMWPE: ultra-high molecular weight polyethylene.

The cover factors (C_fab) for the warp ( _wp ) and weft ( _wf ) yarns were initially calculated from the ratio between the physical yarn diameter d and the pitch $p_{w, f}$ , Equation (1). The net or fabric cover factor was then calculated by combining the individual warp and weft cover factor values, removing the area where the yarns overlap, Equation (2). The cover factor was calculated using the data provided in the material specification and by direct measurement of the woven fabric

C_{wp, wf} = d / p_{wp, wf}

(1)

C_{fab} = C_{wp} + C_{wf} - C_{wp} C_{wf}

(2)

Two types of ballistic test were used to investigate the impact response of the fabric. A residual velocity test was used to capture the energy absorption capability of each fabric sample. The second ballistic test captured the back face displacement of the fabric samples under impact with two offset high-speed cameras (Photron FASTCAM SA5, 50 mm Nikon lens and Bowen’s flash lamp). Ballistic testing was conducted on a helium gas gun with a pressure of up to 300 bar. A schematic of the test set-up used can been seen in Figure 1. A stainless steel 0.7 g sphere projectile was used in place of a 1.1 g chisel-nosed Fragment Simulating Projectile (FSP). The lower mass of the sphere enabled greater differentiation between pre-impact and post impact velocities. Unlike a FSP, the sphere projectile has a consistent strike face regardless of impact angle, thus allowing consistent and accurate comparisons between test fabrics. The projectile was fired from a 7.62 mm caliber 1 in 12 twist rifled barrel using a bespoke nylon sabot used to help achieve the maximum velocity possible with the given firing system. Projectile impact properties ranged from 320 to 640 m s^–1 and were bracketed into eight 40 m s^–1 intervals with five tests performed at each interval for all test fabrics. Test fabrics were cut to 150 mm × 150 mm then mounted and clamped in a bespoke aluminum frame. The frame was designed to clamp the four corners, leaving the edges free.

Figure 1.

Schematic of the test set-up used for ballistic testing of fabrics for energy absorption and transverse and strain wave propagation.

Energy absorbed by the fabric was calculated using Equation (3), where m is the projectile mass and V_i and V_r are the impact and residual velocities, respectively. The projectile remains intact and non-deformed during impact so it can be assumed that there is no mass lost and all energy lost is transferred or absorbed by the target

E_{abs} = \frac{1}{2} m (V_{i}^{2} - V_{r}^{2})

(3)

The two high-speed cameras were set-up to capture the in-plane and out-of-plane response of the target fabrics. Digital image correlation (DIC) was then used to analyze the relative velocities of the transverse waves induced by impact. Target fabrics were cut to 300 mm × 300 mm and stenciled with a unique speckle pattern using a foam roller and Marsh stencil ink. Fabrics were then clamped on all four edges causing the strike area to be reduced to 250 mm × 250 mm.

Single yarn ballistic impact testing was performed using a similar method to that for fabrics except with a different compressed air gas gun, allowing the specimen to be positioned vertically and closer to the barrel, thus improving the accuracy and frequency of impact. Target yarn of ∼1.8 m was wrapped around a dowel at one end and tied off, ensuring not to twist the yarn. Ink marks were made at 10 mm intervals along the first 1.2 m to enable tracking of the strain wave in video analysis. The dowel was then mounted to the roof of the firing chamber with the bottom end being wrapped around a second dowel so the final gauge length was 1.64 m. Ten twists per meter were added to ensure the yarn was an aligned and compact target. A mass of 150 g was then suspended from the bottom dowel to minimize movement caused by gases escaping from the barrel prior to impact. The projectile was fired without a sabot from a smooth bore 5.56 mm barrel for the two lower impact velocities (220 and 290 m s^–1), and a further impact velocity of 430 m s^–1 was achieved using a 7.62 mm sabot and rifled barrel using an air cylinder. The preferential impact velocity for these tests would be 340 m s^–1, corresponding with the fabric tests. However, this velocity was not within the limits of either system and as such the wave velocities of 340 m s^–1 were interpolated from the results.

Crimp was measured following the procedure outlined in BS ISO 7211-3. A section of fabric is measured and the yarns are harvested and clamped at either end. Force is applied to one end of the specimen yarn and the extension is measured. Crimp is the ratio of the extension to the original length and is usually expressed as a percentage.

Results and discussion

Single-ply fabric energy absorption

The mean specific energy absorption (SEA) for each fabric type across all velocity brackets is shown in Figure 2.

Figure 2.

Mean specific energy absorption for all fabrics across all velocity brackets. Means displayed with standard error. Letters indicate statistically different results (Tukey’s honest significant difference p < 0.05). UHMWPE: ultra-high molecular weight polyethylene.

The one-factor analysis of variance (ANOVA) conducted identifies statistically significant differences in the SEA across all fabrics (F7,56 = 20.55, p = 2.6 × 10^–5). This was followed by Tukey’s honest significant difference (HSD) test to make specific pairwise comparisons (Tukey’s HSD α < 0.05) with statistically significant different results being indicated by letters according to Tukey (Figure 2). On average, the para-aramid fabrics had an 81% greater SEA than the UHMWPE fabrics in this study. The para-aramid with the highest SEA (CT/550/12) had over twice the energy absorption capacity of the highest performing UHMWPE fabric (SK/1760/8). Within each fabric set, the individual fabrics generally appeared to show a reduced SEA with increasing C_fab, although the only statistically significant difference was noted between the highest C_fab fabric and two lower C_fab fabrics in the CT/550/XX fabric set.

The SEA profiles at each velocity for para-aramid and UHMWPE fabrics are shown in Figures 3 and 4, respectively. Both para-aramid and UHMWPE fabrics indicate a reduction in energy absorption with increasing impact velocity. The data suggests that the reduction in energy absorbed is disjointed and not a gradual decline across the velocity range tested.

Figure 3.

Specific energy absorption profiles for single-ply para-aramid fabrics individually labeled with fabric type at the top.

Figure 4.

Specific energy absorption profiles for single-ply ultra-high molecular weight polyethylene fabrics individually labeled with fabric type at the top.

Critical velocity, failure mode and cover factor

The high-speed video (HSV) of each impact was analyzed to determine whether the reduction in SEA was related to the critical velocity of the single-ply fabrics. The defeat of the fabric was observed to establish how the fabrics failed, noting fractured yarns, yarns pulled from the weave and the initiation of a transverse wave. The HSV revealed a distinct variation in impact response with velocity, illustrated in Figure 5. Above the critical velocity, the yarns failed instantaneously, typical of a shear failure, and below this velocity, the fabric deflected and the yarns failed under a tensile failure or were pulled from the weave. These failure modes were in keeping with the failure modes reported by Susich et al.²⁵ and Lee et al.²⁶ It was inferred from the association between experimental observation and published literature that a high-velocity impact is equivalent to the high strain rate response experienced at the front of an armor system. It was similarly inferred that a low-velocity impact is equivalent to the low strain rate response indicative of the rear of an armor system. Given these mixed failure modes, it is difficult to define a precise critical velocity for each fabric.

Figure 5.

Typical impact damage: (a) low-velocity regime; (b) high-velocity regime.

Therefore, Table 2 summarizes a critical velocity range, bounded by the lowest velocity fracture failure and the highest velocity slip failure observed. For any given C_fab, the UHMWPE SK/1760/XX fabrics had higher critical velocity than both para-aramid fabric sets, which can be linked to the strain wave velocity of single yarns (Figure 6), which exceeded that of para-aramids by 16–24%. For all fabric sets the critical velocity reduces with increasing C_fab, suggesting that the higher C_fab fabrics are more susceptible to early failure from the higher strain rate loading associated with higher impact velocities. While limited statistical significance could be drawn between the differences in energy absorbed and the projectile impact velocity, the footage supports the connection between SEA and the critical velocity. The reduction in critical velocity with C_fab suggests that the strain wave velocity also reduces with C_fab, contrary to Stempien’s findings,²⁷ and is corroborated by the transverse wave of each fabric (Figure 6). No values for percentage crimp were given in Stempien’s paper and, as such, this disparity could be the result of increased tension in the warp yarns as the weft density increases. In this study, the thread count of fabrics was increased only in the weft direction, which would limit the crimp in the warp yarns, effectively straightening them.²⁸ A straight yarn under tension would then dissipate the strain more rapidly, as it will not need to uncrimp prior to propagating the strain. If this assumption is correct it could be inferred that critical velocity decreases with higher C_fab in this study due to their higher percentage crimp (Table 2) and, hence, hindering the dissipation of strain energy from the point of impact, causing the strain to accumulate more rapidly and the principal yarns to fail earlier.

Figure 6.

(a) Relationship between strain wave velocity along yarn and projectile impact velocity. (b) Relationship between impact velocity and transverse wave velocity of all single yarn types.

Table 2.

Critical velocity range and crimp for all fabrics

Fabric	Thread count warp x weft/yarns·cm^–1	C_fab	Crimp (%)	Critical velocity range (HSV) m.s^–1
CT/840/10	10 × 10	0.89	4.97	389–389
CT/840/8	8 × 8	0.78	2.14	453–457
CT/550/14	14 × 14	0.93	5.93	316–390
CT/550/12	12 × 12	0.86	3.78	382–404
CT/550/10	10 × 10	0.76	2.25	427–455
SK/1760/8	8 × 8	0.97	16.18	328–520
SK/1760/7	7 × 7	0.92	14.32	429–464
SK/1760/6	6 × 6	0.84	9.48	496–503

HSV: high-speed video.

A comparison of the SEA of the fabrics above and below the critical velocities is shown in Figure 7. The fabric with the greatest SEA both above and below its critical velocity was CT/550/10, closely followed by CT/840/8. These two fabrics have the lowest fabric C_fab of all the fabrics tested (Table 1). In addition, their C_fab values of 0.76 and 0.78 are both close to the optimum C_fab of 0.75 proposed by Figucia.²⁹ The trend between C_fab and the SEA of the fabric must be explained within the confines of the energy absorption mechanisms available. These have been summarized to include energy dissipation through the following: yarn fracture; strain propagation along the impacted yarns; the relative movement of yarns and fibers; and the kinetic response of the fabric.³⁰

Figure 7.

Relationships between specific energy absorption and C_fab for all yarn types below (a) and above (b) the critical velocity.

To understand these trends, the mean actual energy absorbed (E_abs) for both failure modes was calculated for each fabric based on the failure mode identified through HSV, shown in Figure 8. The E_abs rather than SEA allows comparison of the relative energy absorbed through yarn fracture and the frictional effects of the weave between fabrics of a given yarn type. These features are not clear when considering energy absorption performance only in terms of mass efficiency. As expected, given the variation in SEA either side of the critical velocity, there is a significant difference in E_abs under the two failure modes for each fabric type.

Figure 8.

E_abs under the two failure modes: yarn pull and slip failure and yarn fracture.

Above the critical velocity, the energy absorption mechanisms are limited to the accumulation of strain in the impacted yarns leading to yarn fracture. Previous studies by Rodriguez et al.³¹ and Tan et al.³² have considered the strain rate dependency of the failure stress and strain for both UHMWPE and para-aramid polymers.

Both Rodriguez et al.³¹ and Tan et al.³² concluded that the polymer’s modulus, failure strain and failure stress increase with strain rate. While this data is not available for the yarns tested, it can be concluded that the energy absorbed above the critical velocity depends upon the linear density, dynamic tenacity and strain of each yarn. The volume of the yarn strained will depend upon the strain wave velocity within the yarn and the number of yarns impacted for each fabric, where the number of impacted yarns is dictated by the C_fab and the yarn diameter with the more densely woven structures having effectively a greater volume of yarns in the projectile’s path. From Figure 8 under the yarn fracture failure mode, the E_abs for these fabrics is about equal at approximately 4 J for all CT fabrics, contrary to the expected higher net energy absorption for high C_fab fabrics, given the additional yarns. This incongruity is explained by an increased strain wave velocity in the lower C_fab fabrics, as suggested by their higher critical velocities. In this regard, the displacement of the strain wave front along the impacted yarns will travel further prior to the yarn fracturing, each individual yarn absorbing greater energy and compensating for the fewer yarns fractured.

Below the critical velocity, secondary energy absorption mechanisms are engaged. Based on the failure modes identified from the HSV, all the fabrics showed an increased energy absorption capacity under a slip failure rather than yarn fracture of between 1 and 3 J (Figure 8). HSV revealed the early failure of the UHMWPE SK/1760/XX fabrics in the low-velocity regime. Whilst the para-aramid fabrics tended to engage with the projectile casting a net across its path, the projectile was able to roll off the larger diameter SK/1760 yarns. This is likely to be augmented by the secondary yarns migrating down the principal yarns away from the point of impact, allowing an easy opening through which the projectile may pass and could explain the lower SEA across all velocity brackets. For the lower C_fab para-aramid fabrics, CT/550/10 and CT/840/8, the benefit of transitioning between failure modes was less pronounced than for the higher C_fab fabrics, the slip failure showing only a 21% and 33% increase compared to 55% and 61% for the higher C_fab fabrics. This indicates that the secondary energy absorption mechanisms dominated in the higher C_fab fabrics.

Ballistic testing has shown that the SEA below the critical velocity also decreased with increasing C_fab and that the SK/1760 fabrics had a significantly lower SEA than the para-aramids (Figure 7). If the E_abs below the critical velocity was governed by strain development and fiber fracture, it would follow a similar pattern to the SEA above the critical velocity and potentially decrease more rapidly with C_fab as the strain wave velocity in the fabrics reduces. Instead, this trend shows a more gradual decline than above the critical velocity (Figure 7). To illustrate this, the data presented in Figure 7 is shown as E_abs rather than SEA in Figure 9.

Figure 9.

Relationship between net energy absorbed below the critical velocity and C_fab for the three fabric sets.

It can clearly be seen that E_abs actually increases with C_fab for all fabric sets, a feature that is hidden when mass efficiency is considered for SEA. Although partitioning of energy absorption between the various mechanisms means that the exact contribution of each is unknown, the secondary mechanisms below the critical velocity must be responsible for the additional energy absorbed. Previous numerical studies have all shown that these secondary energy absorption mechanisms are highly sensitive to the inter-yarn friction of the weave. These mechanisms are summarized as yarn pull out, yarn migration from the point of impact, shearing at yarn crossovers and the global deflection of the fabric.^33–36 It has further been shown that for plain woven fabrics of any given yarn type that the energy absorption would increase with thread count as the friction within the weave increases.³³ Within the impact velocities tested, the increase in energy absorbed through these secondary energy absorption mechanisms was insufficient to compensate for the additional mass held by higher C_fab fabrics.

Layered system

Determination of the critical velocity and SEA data above and below the critical velocity for the single-ply fabrics has resulted in a generalization that the lower C_fab fabrics are best front, middle and rear of a layered soft armor system. Whilst the lightest CT/550/10 fabric showed a superior energy absorption capacity as a mean value above and below its critical velocity, this may vary across the eight impact velocity brackets tested, which the previous generalization fails to account for. To address this, fabrics are considered on a layer-to-layer basis by a theoretical spaced system, that is, a system consisting of layers of fabric separated so that no single layer interacts with a successive layer within the system. This idealized system undergoes an impact from a projectile of known size, shape and impact velocity. From the energy absorption data presented, the first layer of the system would be a fabric with maximum energy absorption capacity for that impact velocity. Each successive layer would be selected dependent on the residual velocity of the projectile after it has emerged from the previous layer, thus creating a hybridized system. Assuming the projectile to be the 5.56 mm steel sphere impacting the first layer at 620 m s^–1, Figure 10 presents theoretical spaced systems for each fabric type and for an optimized hybrid system detailing the number of layers and corresponding mass per unit area for each system.

Figure 10.

Mass per unit area of the monolithic and hybrid spaced armor system required to defeat a 5.56 mm steel sphere projectile traveling at 620 m^.s^–1. The numbers above the columns indicate the total number of layers for each system.

When considered in this way, it is apparent that the CT/550/10 is the best fabric choice for a monolithic system. It has the lowest mass per unit area, theoretically defeating the projectile with 32 layers of fabrics totaling 3.7 kg·m^–2. The nearest competitor to the CT/550/10 system is the other lower C_fab para-aramid CT/840/8, weighing an additional 540 g m^–2, or 15%. This would agree with the findings of both Cunniff²⁰ and Figucia,²⁹ which independently concluded that it is better to have many lighter layers rather than fewer heavier layers within an armor system.

The SEA capacity of the para-aramid fabrics was significantly higher than that of the UHMWPE fabrics tested. This was a trend that carried across all impact velocities and is reflected in the relative mass of the UHMWPE and para-aramid systems. The UHMWPE fabrics were made from higher linear density yarn and produced fabrics with a higher mass per unit area. When considering SEA, the UHMWPE fabrics would have had to absorb two to three times the actual energy per layer from the projectile to match the lighter para-aramid fabrics. Above the critical velocity, any layer-to layer-interaction should be localized and limited to the loading of the fabric layers at the point of impact. In this respect, any system effects noted must be linked to the compression of the fabric layers directly in front of the projectile, which may have an effect on their critical breaking strength or the relative strain through successive layers of the system.

Transverse wave interference

An example of transverse wave images taken from the vertical high-speed camera for tested fabrics at both low and high velocity can be seen in Figure 11. These images highlight a clear difference in transverse wave formation of the same fabric under different velocity regimes and it is important to analyze to understand how subsequent layers may interact. Representative images of transverse wave and strain wave measurement, including subsequent DIC, can be seen in Figure 12.

Figure 11.

Single-ply CT/840/10 fabric one and three frames post-impact under low- and high-velocity regimes. Projectile impacting at 331 m s^–2 (a) and 621 m s^–2 (b).

Figure 12.

Transverse fabric deflection wave overlaid with normal distribution (a) and digital image correlation output from high-speed images.

Below the critical velocity, there is potential for further interference between layers as the transverse waves propagate. The transverse wave velocities for both single yarns and fabrics were measured and the single yarn data was interpolated to give the wave velocity at 340 m s^–1, which was compared to DIC analysis for the fabrics. Wave velocities for single yarns and fabrics are shown in Figures 6 and Figure 13, respectively.

Figure 13.

Relationship between transverse wave velocity and C_fab for each fabric.

Results show that UHMWPE SK 1760 yarn had both a higher strain wave velocity and transverse wave velocity than both the para-aramid yarns. It is evident that fabric transverse wave velocities were distinctly lower than the transverse wave velocities of their constituent yarns, with UHMWPE fabrics having a lower transverse wave velocity than para-aramid fabrics at all C_fab. Furthermore, the transverse wave velocity of a fabric increased with decreasing C_fab for all fabric sets.

Theory suggests that the transverse wave velocity for both single yarn and fabric are dependent upon the velocity of the preceding strain wave along the yarns.³⁷ There was direct evidence of this for the single yarns, where both the transverse wave and strain wave velocity increased with impact velocity. While the strain propagation could not be directly measured for the fabrics, there was indirect evidence of this dependency. The increase in critical velocity with decreasing C_fab indicated that the strain wave velocity was greater in the low C_fab fabrics, which coincided with the decrease in transverse wave velocity with C_fab seen through the DIC analysis.

The reduction in transverse wave velocity from yarn to fabric and with increasing C_fab confirms that the fabric parameters inhibit the propagation of the strain and transverse waves. Furthermore, there is evidence to suggest that the weave architecture has more influence on wave propagation than the yarn elastic modulus amongst the fabrics tested. Figure 14 shows the transverse wave velocity in relation to the percentage crimp of each fabric. Based on the yarn elastic modulus alone, the UHMWPE fabrics should exhibit a higher transverse wave velocity, as they did in the single yarn tests (Figure 6).

Figure 14.

Relationship between transverse wave velocity and percentage crimp for all fabrics.

However, the UHMWPE fabrics were manufactured from a high linear density yarn and were shown to have a much higher percentage crimp than the finer para-aramid fabrics (Table 2). Consequently, both strain and transverse wave must have been retarded more severely in the UHMWPE fabrics as a result of the fabric architecture overriding the advantage of their higher elastic modulus.

The dependency on the fabric crimp shows that the strain and transverse wave velocities can be controlled by fabric geometry, which could be used to manipulate a layered system response.

With respect to designs for a hybridized system, the layers could be forced to interact with each other by placing fabrics of a given yarn type in order of increasing or decreasing C_fab. If a fabric of higher C_fab were placed on the strike face, the transverse wave would pass more slowly across the surface of this fabric than the successive layers, and all layers would propagate the transverse wave freely, as illustrated in Figure 15(a). Alternatively, the lowest C_fab fabric could be placed at the strike face and its faster transverse wave would drive the successive higher C_fab layers in the system (Figure 15(b)). Zhou et al.¹⁹ showed that hybridization of layers between aramid and UHMWPE can result in increased energy absorption. Further development of this hybridization using varying cover factors as well as material could see further improvements in hybridized systems.

Figure 15.

Theoretical response of multiple ply plain weave systems with (a) increasing C_fab and (b) decreasing C_fab.

The transverse wave velocities could also be manipulated using fabrics with different constituent yarn types, as was intended by Cunniff²⁰ when testing the two-ply UHMWPE/para-aramid system. However, these results have shown that it would be incorrect to use the elastic modulus alone to set the relative transverse wave velocities of each layer without considering the weave architecture; for this study, the transverse wave in the UHMWPE fabrics traveled more slowly than in the para-aramids despite their higher modulus.

Conclusion

It was identified from single-ply impact tests that the lower C_fab fabrics (0.76–0.84) within each fabric set had the following:

the highest SEA across the impact velocities tested;

the highest critical velocities; and

the greatest transverse wave velocities.

Of the low C_fab fabrics, the CT/550/10 fabric had the highest mean SEA of all fabrics across the range of impact velocities tested. It was determined that this fabric would provide the lowest mass solution for a homogenous spaced system in which the layers act independently of one another. Based on the single-ply SEA data, only a small mass efficiency could be gained by hybridizing the system to include the CT/840/8 fabric, which proved superior in the highest velocity bracket; the hybrid system weighing only 3% less. The UHMWPE fabrics were not considered suitable for a woven hybrid system as they had a significantly lower SEA compared to all the para-aramid fabrics.

It was speculated, but not verified, that at velocities below 320 m s^–1 the balance of energy absorption would change in favor of the higher C_fab/higher friction fabrics, as the peak strain in the principal yarns reduces and frictional effects increase. If correct, a hybrid system would benefit from higher C_fab para-aramid fabrics (>0.85) as rearward layers. However, as a stacked system with layers combined, the higher C_fab fabrics have slower transverse wave propagation that, if placed at the rear of a system, may inhibit the motion of the lower C_fab fabrics in the preceding layers and limit the system E_abs. It is suggested that these conflicting theories should be tested and discussed further.

Footnotes

Acknowledgements

The authors would like to thank Axis Composites for weaving the test fabrics. Further thanks to Ulster University and Northern Ireland Advanced Composites and Engineering (NIACE) Centre for support and provision of testing equipment.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a Department for Employment and Learning (DEL) CAST award and supported by DSTL.

ORCID iD

Calvin Ralph

References

Laible

Barron

History of armour. In: Laible

(eds). Methods and phenomena 5: ballistic materials and penetration phenomena, 1st ed. Amsterdam: Elsevier Scientific Publishing Company, 1980, pp. 9–40.

Explosive Injury. Parliamentary office of science and technology, Explosive Injury, POSTnote 395; 2011.

Bashford

. Soft armor, its beginning, development and possible value. Helmets and body armor in modern warfare, 1st ed. London: Yale University Press, 1920.

Laible

Fibrous armour. In: Laible

(eds). Methods and phenomena 5: ballistic materials and penetration phenomena, 1st ed. Amsterdam: Elsevier Scientific Publishing Company, 1980, pp. 73–115.

Yang

. Kevlar aramid fiber, Chichester: Wiley, 1993.

Karahan

Kuş

Eren

. An investigation into ballistic performance and energy absorption capabilities of woven aramid fabrics. Int J Impact Eng 2008; 35: 499–510.

McDonald J. Medieval Warfare armour and shields, http://www.medievalwarfare.info/armour.htm (2015, accessed 17 June 2018).

Riley-Smith B. Soldiers complain new body armour makes it difficult to get up from the ground. Telegraph, 19th May, 2016.

UK MOD. Combat body armour, http://www.army.mod.uk/equipment/23216.aspx (2016, accessed 28 July 2017).

10.

Erwin S. Army has few options to lessen weight of body armor. National Defence Magazine, 2009.

11.

Savage-Knepshield

Martin

Lockett

, et al. Designing soldier systems: current issues in human factors, 1st ed. London: CRC Press, 2012.

12.

Carter

Cheuvront

Williams

, et al. Epidemiology of hospitalisations and deaths from heat illness in soldiers. Med Sci Sport Exercise 2005; 37: 1338–1344.

13.

SEA. Reducing the burden on the dismounted soldier (RBDS), https://www.soldiermod.com/volume-8/des.html (2010, accessed 22 September 2017)..

14.

Walkinshaw L. Travelling light: reducing the burden on the close-combat soldier, http://www.army-technology.com/features/featuredssi-icg-des-dstl-soldier-as-asystem/ (2012, accessed 22 September 2017).

15.

Cavallaro P. Soft body armor: an overview of materials, manufacturing, testing, and ballistic impact dynamics. NUWC-NPT Technical Report 12,057, 2011.

16.

Prosser

. Penetration of nylon ballistic panels by fragment-simulating projectiles. Part 2: mechanisms of penetration. Text Res J 1988; 58: 161–165.

17.

ADM002075:2006. From filaments to fabric packs, simulating the performance of textile protection systems.

18.

Thomas JHL. Anonymous layered ballistic resistant material. Body Armour Patent, No. US 7,100,490 B2, USA, 2004..

19.

Zhou

Hou

Gong

, et al. Hybrid panels from woven Kevlar and Dyneema fabrics against ballistic impact with wearing flexibility. J Text Inst 2018; 109: 1027–1035.

20.

Cunniff

. An analysis of the systems effects in woven fabrics under ballistic impact. Text Res J 1992; 62: 495–509.

21.

Yang

Chen

. Influence of fabric architecture on energy absorption efficiency of soft armour panel under ballistic impact. Compos Struct 2019; 224: 111015–111015.

22.

Azrin Hani

Roslan

Mariatti

, et al. Body armour technology: a review of materials, construction techniques and enhancement of ballistic energy absorption. Adv Mater Res 2012; 488–489: 806–812.

23.

Nilakantan

Horner

Halls

, et al. Virtual ballistic impact testing of Kevlar soft armor: predictive and validated finite element modeling of the V0-V100 probabilistic penetration response. Defenc Technol 2018; 14: 213–225.

24.

Guo

Chen

Zheng

. A semi-empirical design parameter for determining the inelastic strike-face mass fraction of soft armor targets. Int J Impact Eng 2019; 125: 83–92.

25.

Susich

Dogliotti

Wrigley

. Microscopical study of multilayer nylon body armor panel after impact. Text Res J 1958; 28: 361–377.

26.

Lee

Walsh

Won

, et al. Penetration failure mechanisms of armor-grade fiber composites under impact. J Compos Mater 2001; 35: 1605–1633.

27.

Stempien

. Influence of a woven fabric structure on the propagation velocity of a tension wave. Fibre Text East Eur 2007; 15: 174–178.

28.

Sule

. Investigation of bending and drape properties of woven fabrics and the effects of fabric constructional parameters and warp tension on these properties. Text Res J 2012; 82: 810–819.

29.

Figucia F. Energy absorption of Kevlar® fabrics under ballistic impact. Report ADA090390, 1980, pp.29–41.

30.

Termonia

. Impact resistance of woven fabrics. Text Res J 2004; 74: 723–729.

31.

Rodriguez

Chocron

Martinez

, et al. High strain rate properties of aramid and polyethylene woven fabric composites. Compos B Eng 1996; 27: 147–154.

32.

Tan

Zeng

Shim

. Characterization and constitutive modeling of aramid fibers at high strain rates. Int J Impact Eng 2008; 11: 11–1303–1313.

33.

Sadegh A and Cavallaro P. Crimp imbalanced protective (CRIMP) fabrics: an analytical investigation into the relationship between crimp contents, energy absorption and fabric ballistic performance. Report ADA530736, 2010.

34.

Duan

Keefe

Bogetti

, et al. A numerical investigation of the influence of friction on energy absorption by a high-strength fabric subjected to ballistic impact. Int J Impact Eng 2006; 32: 1299–1312.

35.

Zeng

Shim

Tan

. Influence of boundary conditions on the ballistic performance of high-strength fabric targets. Int J Impact Eng 2006; 32: 631–642.

36.

Zhou

Chen

. Numerical investigations into the response of fabrics subjected to ballistic impact. J Ind Text 2016; 45: 1520–1547.

37.

Smith

McCrackin

Schiefer

. Stress-strain relationships in yarns subjected to rapid impact loading: Part V wave propagation in long textile yarns impacted transversely. Text Res J 1958; 28: 288–302.