The structural effects on the impact response of ultra-high-molecular-weight polyethylene plain weaves

Abstract

This paper investigates the penetration and energy absorption mechanisms of ultra-high-molecular-weight polyethylene plain weaves with different fabric properties. Impact tests along with finite element (FE) analysis were used to study the impact response of the fabrics. In this research, the impacting projectile did not cause any fiber or yarn failure on the samples. It was found that structural parameters determine the yarn pull-out behavior and the softness of the resultant fabrics. Fabrics formed by loosely interlaced yarns tend to exhibit higher softness and less resistance against yarn pull-out. When the projectile velocity is not sufficient to initiate yarn pull-out, material softness determines the depth of the backface signature on the clay witness. This trend is more pronounced in a multi-ply fabric system than in a single-ply system; when yarn pull-out occurs, the projectile-slowing mechanism depends on the frictional force between the warp and weft yarns. Therefore, fabric softness becomes less important, and the yarn pull-out behavior of the fabric plays a predominant role in energy absorption. FE prediction showed that tightly woven fabrics exhibit a larger area of stress distribution and material deformation than those with severe yarn pull-out and, consequently, these tight fabrics tend to absorb more kinetic energy and sustain higher impact load from a projectile.

Keywords

woven fabrics yarn pull-out impact backface signature finite element simulation energy absorption

Modern soft body armor needs to be made more impact resistant with reduced weight. To fulfill the dual requirements of light weight and protection, the material used in anti-ballistic liners must be low in density and high in strength. One of the most widely used materials is unidirectional (UD) fabric made of ultra-high-molecular-weight polyethylene (UHMWPE) fibers.

UD fabric is formed by cross-laying UHMWPE filaments in a 0°/90° configuration. Due to the non-crimped fiber profile, the stress waves travel faster in this construction than conventional woven fabrics, enabling more material to participate in energy dissipation. Many researchers have studied the mechanisms of penetration of UD fabrics upon ballistic impact. It was reported that the cross-plied structure exhibits a two-stage response: progressive failure of the penetrated layers and transverse deflection of the un-penetrated layers.^1–4 For the former mechanism, fabric layers are damaged by a shear plug, the evidence of which has been provided not only for rigid carbon and glass fiber-reinforced composites,^5–7 but also for compliant composites made with UHMWPE fiber.^8,9 When the impact velocity is comparatively low (less than 300 m/s), the material failure mode turns into a tensile one.¹ For the latter mechanism, the transverse deflection of un-penetrated layers is effective in dissipating the residual kinetic energy of a projectile and in reducing the blunt trauma on the human torso caused by the impact force. It is believed that if the progressive failure process can be further impeded or delayed, subsequent un-penetrated layers would sustain more impulse load and provide better ballistic protection.

When a composite panel is impacted, the projectile tends to exhibit different failure modes in the through-the-thickness directions.^10–12 Similar phenomena were also observed on dry fabric panels by Guo et al.¹³ The fact that different layers of fabric exhibit various responses to impact suggests an advantage in combining more than one type of material in a panel. Mixing different materials in an appropriate sequence would make the best use of their corresponding properties and consequently enable the panel to be more energy absorbent. Inspired by this design guidance, engineering hybrid panels containing more than one type of material is a reasonable approach.^13,14 One of the most studied approaches is the incorporation of woven fabrics in UD fabric panels.^9,15 Although a woven fabric alone does not provide desirable anti-blunt-trauma properties, this type of structure is useful in impeding the progressive failure of fabric plies when placed near the impact face of a panel. There are many studies focused on the optimization of woven-fabric-based panels,^16,17 but the potential of this type of structure in ballistic applications has not been fully explored.

In most research studies, the fabric or sample panels are tightly clamped on the edges so that the target is well engaged with an impacting projectile.^18,19 When the fabric edges are not constrained in a ballistic liner, the penetration mechanism is essentially yarn pull-out and yarn “windowing.” “Windowing” refers to the mode of penetration where a projectile pushes the principal yarns aside.²⁰ The fact that fabric penetration is more often than not accompanied by yarn pull-out significantly hinders high-performance fibers from exploiting their mechanical properties, resulting in the lower impact resistance in woven constructions.²⁰ There are many publications that correlate the yarn pull-out behavior of fabrics with their corresponding ballistic performance. Dong and Sun²¹ found that the energy absorption capability of woven fabrics varies consistently with their yarn pull-out force. It was established that the application of shear thickening fluid (STF) on the woven fabrics could increase inter-yarn frictional force.^22–26 Other recent findings showed that the application of ZnO nanowires^27,28 and graphene oxide coating²⁹ on the fiber surface increases inter-yarn friction, and the resultant fabrics exhibit improved protection against impact. Nevertheless, there are some disadvantages related to the aforementioned chemical treatments. The resultant composites could become stiffer and heavier than the neat fabrics. For an untreated woven fabric, the yarn pull-out behavior depends largely on the structural parameters, for example, thread density and yarn count. In this research, we investigate the structural effect on the impact response of plain UHMWPE fabrics. The interplay between the yarn pull-out behavior and fabric softness during the penetration mechanisms of the plain weave was evaluated. Finite element (FE) analysis was used to identify the stress distribution, deformation, and energy absorption of the corresponding fabric structure.

Fabric specifications

Four types of UHMWPE plain weave made of Tekmilon® I were purchased from Mitsui & Co., Ltd, and were used in this study. The fabric is made of Tekmilon® I multi-filament yarns, having a breaking strength of 291 cN/tex, a modulus of 996 cN/tex, and a breaking elongation of 3%. Other specifications are given in Table 1. Fabric tightness was calculated based on yarn counts and thread densities, and the diametral quotient was assumed to be 0.037. Figure 1 shows close-up images of each fabric sample. The fabric samples were cut into 25 cm × 25 cm for ballistic testing.

Figure 1.

Optical images for ultra-high-molecular-weight polyethylene fabrics.

Table 1.

Specifications of ultra-high-molecular-weight polyethylene fabrics

Sample no.	1	2	3	4
Yarn count (denier)	200	400	400	1300
Areal density(g/m²)	90	130	180	240
Warp density (threads/cm)	20	14	20	9
Weft density (threads/cm)	20	12	17	8
Crimp of the warp yarn (%)	1.2	4.6	3.7	2.4
Crimp of the weft yarn (%)	1.2	8.2	12.8	3
Fabric tightness (%)	57.5	53.8	70.5	61.3

Testing of fabric properties

Three types of tests were performed, the fabric softness test, yarn pull-out test, and impact test. The results obtained were analyzed and combined to ascertain the penetration mechanisms of flexible materials during an impact event.

Fabric softness tests were performed according to the European standard EN ISO 17235:2015 for determination of leather softness. Figure 2 shows a schematic diagram of this testing method. Fabric samples were fixed by a circular aperture and were transversely dented by a cylindrical loading pin at a predetermined loading force. The circular aperture has a diameter of 25 mm. The distension of the sample was given in millimeters by the vertical displacement of the loading pin. The reading of the pin displacement was used as an indicator of fabric softness.

Figure 2.

A schematic diagram of the method for softness testing.

Yarn pull-out tests were performed to characterize the frictional force between the warp and weft yarns. The method is described in detail in our previous paper.³⁰ The set-up is schematically illustrated in Figure 3. Samples were cut into 6 mm × 12 mm. The length of the fabric zone is 6 cm. Yarn tails were left on the top (5 cm) and bottom (1 cm) sides. A slot was maintained for the very yarn to be pulled out on the bottom while the rest of the tails were clamped by the bottom jaw. The upper jaw grips the selected weft yarn and moves at a constant rate of 250 mm/min. Both the single-yarn and multiple-yarn pull-out tests were performed using this method.

Figure 3.

A schematic diagram of the method for yarn pull-out testing.

The ballistic tests were performed using the set-up shown in Figure 4. In this set-up, the spherical projectile used was 2 g in weight and 8 mm in diameter. The projectile was propelled by compressed gas, and the impact velocity varied in the range of 0–170 m/s. The projectile can be captured by a set of infra-red chronographs, and the impact velocity was obtained from the velocity reader. Non-penetration tests were performed both on single-layer and multilayer fabrics. For the former case, one layer of fabric was backed by a clay witness without any clamping device; for the latter case, targets were made from multi-layer fabrics forming a panel. Preliminary impact tests showed that a panel with an areal density of around 720 g/m² was sufficient to stop a projectile with an impact velocity of 150 m/s. The panel specifications are revealed in Table 2. As it is impossible to separate the fabric in half along the through-the-thickness direction for Sample 2, six layers were laid up to form a panel with an areal density of 780 g/m². The panels were cut into 25 cm × 25 cm for ballistic testing, and were fastened against the witness clay using elastic bondage. All samples were impacted once in the center, and the backface signature (BFS) depth was measured to characterize the impact performance. The softness of the clay witness met the requirements of the HOSDB Body Armor Standards for UK Police (2007). This method provides a first-order estimate of the degree of yarn pull-out at various velocities. Note that the fabric selvedges in all samples were removed so yarns were not constrained, that is, the resistance against yarn pull-out came solely from the frictional force at the crossovers.

Figure 4.

Schematic diagram of the testing range for the non-penetration test.

Table 2.

Panel specifications

Panel no.	Sample no.	No. of layers	Areal density (g/m²)
1	Sample 1	8	720
2	Sample 2	6	780
3	Sample 3	4	720
4	Sample 4	3	720

Experimental results

Fabric softness test

For the fabric softness test, it is not appropriate to compare single-layer fabrics or to normalize the softness of fabric by areal density; therefore, fabrics were layered up, and the results are shown in Figure 5. Some trends are immediately apparent. The value of distension increases as the number of fabric layers decreases, as expected. Sample 1 exhibits the greatest value of distension for a single-layer fabric. As the layer number increases, Sample 2 becomes the softest material; Samples 1 and 4 give similar values, and Sample 3 seems to be the stiffest one. No linear relationships could be established between fabric softness and their corresponding areal densities. Softness is closely associated with the bending rigidity of a material. Other things being equal, the bending rigidity of a woven fabric is associated with the length of the yarn float³¹ and the frictional force between the yarn-forming fibers.³² The frictional force is produced by the inter-yarn pressure, which is related to fabric tightness. According to the value of fabric tightness listed in Table 1, tighter structures lead to stiffer fabrics, and loosely interlaced structures with a large gap between the neighboring yarns exhibit softer properties.

Figure 5.

Structural effects on the softness of different woven fabrics.

Yarn pull-out tests

Yarn pull-out tests were performed in both the warp and weft directions of the fabrics due to different crimp ratios. Figures 6 and 7 display the junction rupture forces (JRFs) or peak load forces^33–35 of the load–displacement curves obtained from yarn pull-out tests. The JRF or peak load force is an indicator of the resistance against yarn pull-out. Both single and multiple yarns were withdrawn to determine the effect of structural parameters on the yarn pull-out behavior. As can be clearly seen from Figures 6 and 7, greater forces are required to withdraw the warp yarns than the weft yarns from the woven fabrics tested. The behavior can be explained by the following: higher tension is required on the warp yarns to avoid filament entanglement during the weaving process and the warp yarns are twisted and sized to protect against fiber damage due to repeated bending, abrasion, and flexing. These two factors could make the warp yarns stiffer and more difficult to bend when interlaced with the weft yarns. Therefore, despite their lower crimp ratio, the warp yarns need to overcome greater resistance forces due to crimp exchange during the yarn pull-out process, resulting in higher JRFs in the warp direction.

Figure 6.

Comparison of peak load force for multiple-yarn pull-out tests in the weft direction.

Figure 7.

Comparison of peak load force for multiple-yarn pull-out tests in the warp direction.

In the single-yarn pull-out tests in both the weft and warp directions, the JRFs of Sample 1 are consistently lower than other samples, while Sample 2 ranks the second lowest. Samples 3 and 4 exhibit similar values when a yarn is pulled from the weft direction. However, the JRF of Sample 3 is almost double that of Sample 4 when a yarn is pulled from the warp direction. In multiple-yarn pull-out tests, the JRFs are not proportional to the number of yarns withdrawn. For Sample 3, the JRF for double, triple, and quadruple yarn pull-outs are around 2, 8, and 16 times higher than that for single-yarn pull-out in the warp direction. In addition, it seems that the influence of the number of yarns being withdrawn is less significant on fabrics with lower areal densities. It is difficult to quantify the yarn pull-out behavior of fabrics with different physical characteristics, such as linear density, thread density, and fabric tightness. Nevertheless, it can still be concluded from Figures 6 and 7 that tightly woven fabrics with higher crimp ratios exhibit higher resistance against yarn pull-out. For instance, the JRF of Sample 3 is almost five times than that of Sample 1 when pulling single warp yarns. The difference is more pronounced for multiple-yarn pull-out tests. When pulling four yarns together in the warp direction, the JRF of Sample 3 is higher than that of Samples 2 and 4 by factors of ∼13 and ∼9, respectively.

Impact tests

The BFS depths for various types of single-layer fabrics are plotted as a function of impact velocity in Figure 8. The impact results are subjected to analysis of variance (ANOVA), which is shown in appendix Tables A1–A3. It has been found that the values of F for all the velocities are much greater than the critical value of F_(3,16) corresponding to α = 0.05, which is 5.29 (the F value is the ratio of the mean squares; it is an estimate of population variance that accounts for the degree of freedom used to calculate that estimate). We are 99% confident (1–α) that the depths of the BFS are not statistically equal for Samples 1–4. Sample 3 outperformed all other samples tested, and exhibited no yarn pull-out nor material damage at any of the impact velocities tested (Figure 7). The superior performance of Sample 3 could be explained by its higher stiffness and higher resistance against yarn pull-out.

Figure 8.

Backface signature (BFS) depth as a function of projectile impact velocity for different samples.

At an impact velocity of 40 m/s, the BFS depth varies inversely with the areal density. Yarn displacement was not found on all the samples (Figure 9(a)). At an impact velocity of 60 m/s, yarn pull-out and “windowing” are evident in Samples 2 and 4. Although yarn pull-out and windowing are not evident in Sample 1, it performs similarly to Samples 2 and 4. This is because fewer yarns are involved in stopping the projectile. At a velocity of 80 m/s, all fabrics exhibited yarn pull-out and windowing, except Sample 3. The number of warp and weft yarns pulled out at this velocity is displayed in Table 3. It appears that fabric samples with no withdrawn yarns exhibit low and less random BFS depths than those with severe yarn pull-out. For Samples 1, 2, and 4, the number of principal yarns pulled out varies, influencing the BFS depth on the clay witness. When the projectile engaged more principal yarns, more kinetic energy was dissipated as the yarns pulled out, and therefore there was a less defined BFS; when the projectile spanned fewer principal yarns, less kinetic energy was absorbed during the yarn pull-out process and hence there was a more defined BFS.

Figure 9.

Post-impact close-ups at different impact velocities. (a) Sample 1 (b) Sample 2 (c) Sample 3 (d) Sample 4.

Table 3.

Number of yarns pulled out and backface signature (BFS) depth on single-layer fabrics at 80 m/s impact velocity

Sample no.	BFS depth (mm)	Mean value (mm)	Standard deviation	No. of yarns pulled out (threads)
Sample no.	BFS depth (mm)	Mean value (mm)	Standard deviation	Warp	Weft
1	29.5			3	3
	24.3			4	4
	27.2	26.6	1.94	3	4
	26.3			3	4
	25.6			4	4
2	33.4			3	3
	38.1			2	3
	35.6	36.7	2.32	2	3
	39.7			2	2
	36.5			2	3
3	15.2			-	-
	15.4			-	-
	16.3	15.0	0.98	-	-
	14.1			-	-
	13.9			-	-
4	33.5			1	1
	26			2	2
	24.7	27.3	3.57	2	2
	26.8			1	2
	25.3			2	2

Fabrics were layered differently in each panel to normalize the areal densities as much as possible in testing. Results of the ANOVA are listed in appendix Table A4. Also, the values of F for all velocities are much greater than the critical value of F_(3,16) corresponding to α = 0.05, indicating that the performance of different samples are not statistically similar. It can be seen in Figure 10 that Panels 2 and 4 show a similar performance, which is outperformed by that of Panels 1 and 3. Post-impact close-ups of each panel are shown in Figure 11. Here, “Layer 1” refers to the panel at the impact face of a panel; “Layer 2,” “Layer 3,” and so on refer to the subsequent layers away from the impact face. On Panel 2, slight yarn displacement is noticeable for layers close to the clay witness. The extent of yarn pull-out is more noticeable on Panel 4 than on the other panels. The actual number of yarns pulled from each layer is indicated in Table 4. The results suggest that the projectile pulled more yarns on layers near the impact surface of a panel and less on layers away from the point of impact.

Figure 10.

Depth of the backface signature (BFS) for different panels at an impact velocity of 150 m/s.

Figure 11.

Post-impact close-ups of different panels at 150 m/s impact velocity. (a) Panel 1 (b) Panel 2 (c) Panel 3 (d) Panel 4.

Table 4.

Number of yarns pulled out from Panel 4 at 150 m/s impact velocity

Shoot no.	BFS depth (mm)	Layer 1		Layer 2		Layer 3
Shoot no.	BFS depth (mm)	Warp	Weft	Warp	Weft	Warp	Weft
1	20.1	3	3	2	1	1	0
2	20.6	2	3	2	2	1	0
3	21.6	2	2	1	2	0	1
4	22.03	2	2	2	1	0	1
5	21.35	2	2	2	1	1	0

BFS: backface signature.

Woven fabrics are formed by yarn interlacement. The interlaced structure has comparatively lower dimensional stability than composite materials such as UD fabrics, and yarns are more prone to be pulled out and/or pushed aside by an impacting projectile. If the kinetic energy of a projectile is not sufficient to cause fiber damage, the two major factors that influence the impact performance seem to be the fabric softness and yarn pull-out behavior. The two factors are closely associated with fabric tightness: firstly, tighter structures lead to stiffer fabrics; secondly, a tighter structure enables the fabric to have greater resistance against yarn pull-out.

Fabric softness plays a more important role in determining the BFS property of fabrics when the impact velocity is not high enough to initiate yarn pull-out. For example, when testing a single fabric, Sample 1 is the softest material and exhibits the highest value of distention, followed by Samples 2, 4, and 3. This trend is identical to the results at an impact velocity of 40 m/s when yarn pull-out is not evident in all single-layer samples (Figure 7). On multilayer fabrics, Sample 2 becomes the softest material, followed by Samples 1, 4, and 3. When the corresponding panels are subjected to a non-penetration test at an impact velocity of 150 m/s, the differences in impact performance become more apparent, with the softer panel showing a deeper BFS, with the exception of Panel 4. In Panel 4 yarns were pulled out, and a deep cavity was created (Figure 9(d)).

When the impact velocity is sufficient to initiate yarn pull-out, fabrics are essentially perforated, and a deep hole is created in the clay witness; the projectile-slowing mechanism depends largely on the yarn pull-out behavior of fabric samples and the witness material. Consequently, inter-yarn friction in the interlaced structure and the deformation of the clay witness dissipate the projectile kinetic energy. From the results obtained on single-layer fabrics, Samples 1, 2, and 4 exhibit a considerable extent of yarn pull-out at an impact velocity of 80 m/s (Figure 7). The number of yarns pulled out on Sample 1 is greater than those on Samples 2 and 4. Sample 1 provides similar resistance against yarn pull-out as Samples 2 and 4, as indicated by Table 3 and Figures 8 and 9. Table 3 also shows that a shallower cavity is created in the clay witness if more yarns are involved in the pull-out event. A high-velocity impact test differs from a low-velocity drop-weight impact test in that the penetration process for the former case occurs almost instantly, resulting in a significant out-of-plane shear force on the material. For rigid fiber-reinforced composites, shear failure occurs;^5–7 for flexible fiber assemblies, especially woven constructions, yarn pull-out is involved in the penetration mechanisms.

Impact simulation

Finite element models for plain weaves

Commercial FE software ABAQUS® was used to simulate the impact and further study the fabric deformation, stress distribution, and energy absorption processes during an impact event. In this research, FE simulation is limited to the collision between a projectile and a flexible single-layer fabric supported by a clay witness. The projectile model is spherical with an 8 mm diameter and 2 g mass, identical to that used for the practical tests. Fabrics were simulated at the yarn level to capture yarn pull-out during an impact event. The yarn was modeled as a three-dimensional (3D) solid geometry with undulated crimps and elliptical cross-sections according to their corresponding fabric parameter (Figure 12(a)). The application of lenticular cross-sections avoids large element deformation between a pulled yarn and its orthogonal yarns. Yarns were assembled in the warp and weft directions to construct a plain weave of 100 mm × 100 mm. Since the sample targets were backed by a clay witness in the impact test, free boundary conditions were assigned for the outer edges of the fabric models. Symmetric boundary conditions about the X- and Z-axes were applied to the other two edges to save computational resources. The clay witness was also modeled as a 3D solid continuum block. The FE models are shown in Figure 12(b).

Figure 12.

Meshing of finite element models for Sample 4. (a) Yarn model (b) A projectile colliding clay-witness-backed plain weave.

Yarns were modeled as homogenous continua to represent the filament-level architecture. The material was assumed to be isotropic whereby woven fabrics could be recreated with acceptable fidelity at a comparatively low computational cost. The Young’s modulus for the yarn model was 997 cN/dtex according to data listed in Table 1. The tensile modulus was converted to an engineering modulus by multiplying material density, which was calculated to be 96.7 GPa. The mechanical properties of the clay witness were taken from a study with a similar impact simulation, including the density, Poisson’s ratio,³⁶ Young’s modulus, and yield stress.³⁶ However, it was found that the application of the Young’s modulus (6.58 × 10^–3GPa) and yield stress (6 × 10^–5GPa) leads to model instability and large deformation of the elements, and therefore the two values were increased by three orders of magnitude. The material was modeled as perfectly plastic with no hardening. A yield stress of 2.82 GPa was entered with a corresponding initial plastic strain of 0.0. Ductile damages of both yarns and the clay witness were assumed and applied in the model. Damage is initiated when the following condition is satisfied³⁷

\int \frac{d {\bar{ɛ}}^{pl}}{{\bar{ɛ}}_{D}^{pl} (η, {\overset{\cdot}{\bar{ɛ}}}^{pl})} = 1

(1)

where η is the stress triaxiality,

{\overset{\cdot}{\bar{ɛ}}}^{pl}

is the equivalent plastic strain rate,

{\bar{ɛ}}^{pl}

is the equivalent plastic strain, and

{\bar{ɛ}}_{D}^{pl}

is the equivalent plastic strain rate at the onset of damage. Damage evolution energy is assigned to be 500 J and material softening is defined as exponential. The projectile is modeled as a rigid body and was subjected to no deformation during the impact process. The material properties are listed in Table 5. A general contact interaction was used to define the contact between the projectile and the sample target, and between the interlaced warp and weft yarns. The coefficient of friction was determined to be 0.14.⁹

Table 5.

Material properties

Material properties	Projectile	Yarn	Clay witness
Yield stress (GPa)	/	2.82	6 × 10^–2
Fracture strain (%)	/	3	/
Young’s modulus (GPa)	/	96.7	6
Poisson’s ratio	/	0.3	0.35
Mass density (kg/m³)	7800	970	1539

Initial model validation

The clay witness model was subjected to a validation process to check its accuracy. The calibration of the clay witness was performed according to HOSDB Body Armor Standards for UK Police (2007) 8 Part 2: Ballistic Resistance. The test consists of a free fall sphere onto the backing material at a height of 1.5 m. The sphere has a mass of 1.043 kg and a diameter of 63.5 mm. The depth and width of the BFS were recorded and compared with the FE models. The difference between numerical predictions and experimental results was found to be within 1 mm (Figure 13).

Figure 13.

Validation of the clay witness model. (a) Clay witness (b) Numerical results.

Validation of the fabric model was performed using experimental data taken from Samples 1–4. A comparison of experimental results and FE predictions is shown in Figure 14. It was found that the numerical BFS depth values were greater than those of the experimental results. The underestimation of fabric performance by the simulation can be attributed to the reduced size of the model, and probably less material was involved in energy dissipation. The difference becomes more pronounced as the impact velocity increases. At an impact velocity of 80 m/s, the primary projectile-slowing mechanisms of Samples 1, 2, and 4 were found to be yarn pull-out and clay deformation. Due to the limited sample size, the energy dissipated by yarn pull-out was underestimated by the model and hence the greater depth of the BFS.

Figure 14.

Comparison of experimental results and finite element (FE) predictions. BFS: backface signature.

Results and discussion

In this section, fabric deformation, stress distribution, and energy absorption during the impact event are studied in detail to further explore the penetration mechanism of plain fabrics. Figure 15 reveals the counter plots of stress distribution for different fabric models taken 30 μs after impact. It can be seen that fabrics become more deformed and stress is distributed to a larger area as the impact velocity increases. Also, yarn pull-out is evident in Samples 1, 2, and 4. Sample 3 exhibits a wider transverse deflection than other samples due to its tight structure. It is also observed that fabric away from the impact point was stretched in Samples 3 and 1, indicating that more yarns were involved in energy dissipation. Apart from Sample 3, yarns were more or less unraveled on the boundaries. By comparing the FE results in Figure 15 with the post-impacted fabrics in Figure 9, it is found that yarn pull-out is more likely to occur on Samples 2 and 4 than on Sample 3, and is more likely to occur at high velocities than at low velocities.

Figure 15.

Contour plots of single-layer plain weaves subjected to impact from a spherical projectile at velocities of (a) 40 m/s, (b) 60 m/s, and (c) 80 m/s, 30 μs after impact.

Figure 16 shows von Mises stress distribution along a pre-selected primary yarn on different plain weave models 30 μs after impact. It can be seen that the magnitude of stress increases as the impact velocity increases on all the plain fabrics. The magnitude of stress on Sample 3 was the highest among all samples. The difference between Sample 3 and the other samples widened as the impact velocity increased. This is because yarn pull-out releases the primary yarns from the structure. Without positive constraints from the fabric structure, the primary yarns could not sustain the impact load and the forces will likely be transmitted to the witness clay. In addition, it was also found that the stress values presented in Figure 16 are far from reaching that of plastic yield stress, and therefore material failure will not occur.

Figure 16.

Stress distribution in a primary yarn of single-layer plain weave at the impact velocities of (a) 40 m/s, (b) 60 m/s, and (c) 80 m/s, 30 μs after impact.

When a clay-witness-backed sample is impacted, a portion of the projectile kinetic energy is absorbed by the fabric, and the rest is transmitted to the backing material, causing the BFS. Some researchers measure the BFS volume to characterize the energy absorption capability of sample targets.^38,39 This approach provides a first-order estimate of the amount of energy transfer to the backing material. Two issues, however, need to be considered. Firstly, the clay witness might respond differently to a drop-sphere impact and projectile impact, indicating that the energy per unit volume obtained from the former case cannot be used for the latter case. Secondly, shallow and wider BFSs can exhibit a similar clay crater volume as deeper and narrow BFSs. The former case indicates better energy absorption capacity than the latter. In this research, the energy absorption of different sample targets was quantified by the sum of the fabric strain energy, fabric kinetic energy, and energy dissipated due to friction. For mass-normalized metrics, total energy was also divided by fabric areal density to obtain the normalized energy absorption, as shown in Figure 17.

Figure 17.

Energy absorption for single-layer plain weaves at the impact velocities of (a) 40 m/s, (b) 60 m/s, and (c) 80 m/s, 30 μs after impact.

It can be seen in Figure 17 that the FE method predicts an increase in the total energy absorption of plain weaves when the impact velocity increases. Sample 3 was found to be more sensitive to impact velocity than the other samples, indicating that Sample 3 is more energy absorbing at higher impact velocities. Energy absorbed by the fabric is broadly equal to the sum of fabric strain energy, kinetic energy, and frictional energy. In contrast to earlier reports,^40,41 however, this research finds that frictional energy accounts for a major proportion of the total energy. This is probably because earlier research did not consider the interaction between the yarns pulled out and inner peripheral surface of the hole created inside the clay, which also slows down an impacting projectile and enables a large amount of energy to be dissipated by friction. When yarn pull-out is not initiated at an impact velocity of 40 m/s, the total energy absorption increases with areal fabric density. When the total energy absorption is normalized by areal density, energy absorption exhibits a decreasing trend with an increase in areal density (Figure 17(a)). The results indicate that Sample 1 yields the best energy absorption capability despite the greatest BFS depth. When yarn pull-out is initiated at an impact velocity of 80 m/s, the normalized energy absorption for Sample 1 ranks second to that of Sample 3 (Figure 17(c)). It can be seen in Figure 9 that yarn pull-out is not noticeable on Sample 3, which shows the largest fabric stress area compared with the other samples. The yarn pull-out phenomena weakens the fabric’s capability to absorb and dissipate the kinetic energy of the projectile.

In this section, FE simulation was used to further explore the structural effects on the deformation, stress distribution, and energy absorption of different fabrics. FE predictions suggest that when UHMWPE plain fabric was impacted by a sphere projectile at impact velocities under 80 m/s, yarn pull-out was the major factor that determined the amount of energy absorbed and the ability of the primary yarns to sustain higher impact loads. It follows that constraining yarn mobility is beneficial in improving the impact performance of plain weaves. This conclusion is only established under the condition that the impacting force of a projectile is not sufficient to cause yarn failure. When the impact energy is sufficient to damage the fabrics, it is likely the constraint imposed by yarn mobility leads to stress concentration at the impact site, and hence early failure of the material.

Conclusions

This research identifies the influencing factors that determine the impact performance of flexible materials. A gas gun was used to propel spherical projectiles to target samples at impact velocities below 170 m/s. The two major factors that determine impact performance are yarn pull-out behavior and material softness if the kinetic energy of a projectile is not sufficient to cause fiber damage. Fabric softness plays a more important role in determining the BFS property of fabrics when the impact velocity is not high enough to initiate yarn pull-out. When yarn pull-out occurs, the projectile-slowing mechanism depends largely on the frictional force between the warp and weft yarns. Results show that woven fabrics with coarser yarns and looser weaves exhibit inferior performance to those with finer yarns and tighter weaves. Fabrics with finer yarns and tighter weave structures enable the projectile to span and engage more yarns during an impact event. Consequently, the yarn pull-out resistance increases exponentially, and therefore yarns are less likely to be pulled out. FE software (ABAQUS) was used to further study the penetration mechanisms of woven fabrics. Numerical simulations show that fabric samples with less or no yarn pull-out exhibit larger areas of stress distribution and deformation than those with severe yarn pull-out. Consequently, tightly woven fabrics tend to absorb more kinetic energy and sustain more impacting load from a projectile.

Footnotes

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 the National Natural Science Foundation of China (51708553) and a Hubei Province Technical Innovation Special Project (2019AAA005).

ORCID iD

Yi Zhou

Appendix

References

OMasta

Crayton

Deshpande

, et al. Mechanisms of penetration in polyethylene reinforced cross-ply laminates. Int J Impc Eng 2015; 86: 249–264.

Heisserer

Vander

Hendrix

. Ballistic depth of penetration studies in Dyneema® composites. Symp Proc 2013; 2: 1936–1943.

Karthikeyan

Russell

. Polyethylene ballistic laminates: failure mechanics and interface effect. Mater Des 2014; 63: 115–125.

Nguyen

Ryan

Cimpoeru

, et al. The effect of target thickness on the ballistic performance of ultra-high molecular weight polyethylene composite. Int J Impact Eng 2015; 75: 174–183.

Cantwell

Morton

. Impact perforation of carbon fibre reinforced plastic. Compos Sci Technol 1990; 38: 119–141.

Gama

Gillespie

. Punch shear based penetration model of ballistic impact of thick-section composites. Compos Struct 2008; 86: 356–369.

Gama

Gillespie

. Finite element modelling of impact, damage evolution and penetration of thick-section composites. Int J Impact Eng 2011; 38: 181–197.

Lee

Song

Ward

. Failure of spectra polyethylene fibre-reinforced composites under ballistic impact loading. J Compos Mater 1994; 28: 1202–1225.

Chen

Zhou

Wells

. Numerical and experimental investigations into ballistic performance of hybrid fabric panels. Compos Part B Eng 2014; 58: 35–42.

10.

Cheeseman

Boggeti

. Ballistic impact into fabric and compliant composite laminates. Compos Struct 2003; 61: 161–173.

11.

Yang

Chen

. Investigation of failure modes and influence on ballistic performance of ultra-high molecular weight polyethylene (UHMWPE) uni-directional laminate for hybrid design. Compos Struct 2017; 174: 233–243.

12.

Yang

Chen

. Investigation on energy absorption efficiency of each layer in ballistic armour panel for applications in hybrid design. Compos Struct 2017; 164: 1–9.

13.

Guo

Chen

Zheng

. Effect of replacement strike-face material on the ballistic performance of multi-ply soft armor targets. Text Res J 2019; 89: 711–725.

14.

Martinez

Ridruejo

Gonzalez

, et al. Ballistic performance of hybrid nonwoven/woven polyethylene fabric shields. Int J Impact Eng 2018; 111: 55–65.

15.

Chen

Zhu

Wells

. An analytical model for ballistic impact on textile based body armour. Compos Part B Eng 2012; 45: 1508–1514.

16.

Ralph C, Baker L, Archer E, et al. Optimization of soft armor: the response of single-ply para-aramid and ultra-high molecular weight polyethylene fabrics under ballistic impact. Text Res J 2020; 90: 1713--1729.

17.

Yang

Chen

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

18.

Tapie

Guo

Shim

VPW

. Effect of pre-tension on energy absorbed by fabric during projectile penetration. Int J Impact Eng 2018; 121: 134–150.

19.

Liu

Zhao

Chen

, et al. Influence of pre-tension on ballistic impact performance of multi-layer Kevlar 49 woven fabrics for gas turbine engine containment systems. Chin J Aeronaut 2018; 6: 1273–1286.

20.

Nilakantan

Merrill

Keefe

, et al. Experimental investigation of the role of frictional yarn pull-out and windowing on the probabilistic impact response of Kevlar fabrics. Compos Part B Eng 2015; 68: 215–229.

21.

Dong

Sun

. Testing and modeling of yarn pull-out in plain woven Kevlar fabrics. Compos Part A 2009; 40: 1863–1869.

22.

Neelanchali A, Hemant CH, Shishay AG, et al. Impact response of Shear Thickening Fluid (STF) treated ultra high molecular weight poly ethylene composites – study of the effect of STF treatment method. Thin Wall Struct 2017; 126: 16--25.

23.

Khodadadi

Liaghat

Vahid

, et al. Ballistic performance of Kevlar fabric impregnated with nanosilica/PEG shear thickening fluid. Compos Part B 2019; 1: 643–652.

24.

Sen

Jamal

MNB

Shaw

, et al. Numerical investigation of ballistic performance of shear thickening fluid (STF)-Kevlar composite. Int J Mech Sci 2019; 164: 105174–105174.

25.

Arora

Majumdar

Butola

. Soft armour design by angular stacking of shear thickening fluid impregnated high-performance fabrics for quasi-isotropic ballistic response. Compos Struct 2020; 233: 111720–111720.

26.

Yuan

Chen

, et al. Evaluation of ballistic performance of STF impregnated fabrics under high velocity impact. Compos Struct 2019; 227: 111208–111208.

27.

Hwang

Malakooti

Patterson

, et al. Increased interyarn friction through ZnO nanowire arrays grown on aramid fabric. Compos Sci Technol 2015; 107: 75–81.

28.

Malakooti

Hwang

Goulbourne

, et al. Role of ZnO nanowire arrays on the impact response of aramid fabrics. Compos Part B Eng 2017; 127: 222–231.

29.

Oliveira

Weber

Monteiro

, et al. Effect of graphene oxide coating on the ballistic performance of aramid fabric. J Mater Res Technol 2020; 9: 2267–2278.

30.

Zhou Y, Muhammad A, Gong XZ, et al. An experimental investigation of the yarn pull-out behaviour of plain weave with leno and knitted insertions. Text Res J 2019; 89: 4717--4731.

31.

Gong

Mukhopadhyay

. Fabric objective measurement: a comparative study of fabric characteristics. J Text Inst 1993; 84: 192–198.

32.

Grosberg

. The mechanical properties of woven fabrics. Part:: the bending of woven fabric. Text Res J 1966; 36: 205–205.

33.

Sebastian

Bailey

Briscoe

, et al. Extensions, displacements and forces associated with pulling a single yarn from a fabric. J Phys D Appl Phys 1987; 20: 130–130.

34.

Sebastian

Bailey

Briscoe

, et al. Effect of a softening agent on yarn pull-out force of a plain weave fabric. Text Res J 1986; 56: 604–611.

35.

Kirkwood

Lee

, et al. Yarn pull-out as a mechanism for dissipating ballistic impact energy in Kevlar® KM-2 fabric: Part I: quasi-static characterization of yarn pull-out. Text Res J 2004; 74: 920–928.

36.

Yang

Chen

. Study of energy absorption and failure modes of constituent layers in body armour panels. Compos Part B Eng 2016; 98: 250–259.

37.

AIA. Abaqus Analysis User's Manual. Available at: http://abaqus.software.polimi.it/v6.14/books/usb/default.htm.

38.

Mehmet

Abdil

Recep

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

39.

Kedzierski

Popławski

Gieleta

, et al. Experimental and numerical investigation of fabric impact behaviour. Compos Part B Eng 2015; 69: 452–459.

40.

Duan

Keefe

Bogettic

, et al. Modeling friction effects on the ballistic impact behavior of a single-ply high-strength fabric. Int J Impact Eng 2005; 31: 996–1012.

41.

Lim

Shim

VPW

. Finite-element modeling of the ballistic impact of fabric armour. Int J Impact Eng 2003; 28: 13–31.