Experimental investigation of the performances of thick CFRP, GFRP, and KFRP composite plates under ballistic impact

1 Introduction

Owing to the risks of high-velocity ballistic impact damage in various fields, such as military and aerospace industries, an adequate ballistic protection system is required. These concerns are vital because high-velocity impacts from different projectiles can endanger the prolonged functionality of impacted components, resulting in substantial damage and even cause fatalities in extreme cases.

With the advancement in military operations and weapons, the demand for ballistic body armour that can resist damage from ballistic threats while maintaining comfort, flexibility, and high mobility for the wearer has also increased. Furthermore, similar impact risks exist in other applications, such as interiors of turbines and aircrafts or spacecraft exteriors (e.g. runway debris or bird strikes). Consequently, ballistic-resistant structures are essential in these related fields [1].

Currently, fibre-reinforced polymer (FRP) composite laminates are commonly used in ballistic protection due to their high specific strength, weight-saving potential, and energy-absorbing characteristics. However, FRP plates are structurally more complicated than isotropic materials due to the macroscopic combination of fibres and matrices. During ballistic impact, they undergo a complex damage propagation process and failure mechanism. These damages include fibre breakage, matrix cracking, delamination, and interfacial debonding, which occur as the composite plates experience shear plugging, cone deformation, or perforation due to impact [2].

Understanding the response behaviour of these materials to relevant ballistic threats is indispensable for creating the most effective and optimized designs. The influence of changes in the impact parameters, targets, and projectiles on the ballistic performance of the FRPs has been previously analysed. Increasing the aspect ratio [3], thickness [4], bluntness of the projectile nose shape [5], and impact obliquity angles [6,7] was found to improve the ballistic performance significantly, as observed from the smaller residual velocity and greater energy absorption.

With respect to damage characterization analysis, studies on the damage characteristics of high-velocity impact damage in carbon-based FRP (CFRP) laminates and identification of crater formation, delamination, and f breakage have been conducted [8]. A numerical comparative study of glass-based FRP (GFRP) and CFRP composites with varying stacking sequences against a blunt-nose projectile revealed the superior ballistic performance of CFRPs owing to their higher strength and stiffness compared to GFRP [9], whereas an analysis of the ballistic resistance of polyurea-coated CFRP laminates under projectile impact loading revealed the improved performance of the coated plates [10]. Fibre breakage, matrix failure, debonding between the fibres and matrix, fibre pull-out, delamination, conically shaped deformations, and shear failure were identified.

Ballistic impact-response analysis or optimization is typically performed experimentally to ensure the reliability of the results. However, ballistic experimentation investigation of the ballistic performance of FRP laminates can be quite expensive, especially when CFRP and/or Kevlar^®/aramid-based fibre FRP (KFRP/AFRP) materials are involved. This limitation can be overcome by using finite-element modelling. Most ballistic experimental studies are accompanied by a numerical model for comparison, which offers quicker analysis and optimization capabilities.

A comparison of two-scale numerical modelling, a microstructural model (MSM) and a continuum-based model, was conducted [11]. The MSM scale modelled the yarns, enabling the simulation of composite delamination and complex intralayer fractures. Good agreement between the ballistic curves of both models and the experimental results was observed. A mixed macro-homogeneous and meso-heterogeneous approach has also been used to model composite plates [12,13], which allows the observation of yarn and matrix interaction in the near-impact area, while saving computational costs using macro-homogeneous modelling in areas further from the impact point, where the yarn and matrix are considered a single homogeneous entity. A similar scale of modelling defines the individual plies in which full three-dimensional stress states are considered [14]. This model allows the analysis of the interactions between the intralayer and interlayer damage mechanisms. Another study compared this numerical model with a simpler model without interlaminar damage modelling, and a good agreement was observed between the numerical and experimental models [15].

Most of the aforementioned studies were based on thin plates, whereas this study concentrated on thick plates. Naik and Doshi [16] discussed the analytical and experimental studies on thick GFRP plates. They developed a wave propagation model to analyse ballistic impacts and found that shear plugging and friction absorbed most of the impact energy, whereas matrix cracking and delamination did not absorb significant impact energy. The experiment was performed using a 7 mm thick GFRP plate with a cylindrical projectile having a flat end diameter of 6.33 mm and a mass of 5.84 g with a ballistic velocity of 174.22 m/s. No penetration was observed.

In this study, the behaviour of CFRP, GFRP, and KFRP composite thick plates under ballistic impact was experimentally investigated. Layered plates of each type were produced using hand lay-up and vacuum bagging techniques. First, 10 mm plates were manufactured using these techniques, producing plates with a thickness of 2 mm. They were then bonded together using epoxy adhesive and hot pressed at 75°C to obtain thick layered plates with a total thickness of approximately 20 mm. The specimens were then fired using bullets in a special firing room according to the National Institute of Justice (NIJ 0108.01) ballistic standard. The resulting impacted plates were further investigated. Current manufacturing methods were intentionally maintained to obtain thick layered composite plates.

The difference between this research and other research is that this research concentrates on layered structures, compared to solid structures before. We could not find any literature based on layered structures. That will affect the failure modes. Naik and Doshi [16], in which shear plugging was the dominant failure modes, found out that it may be that in layered structures, delamination will be the dominant failure modes. The layered structures may change the behaviour of impact energy absorption from shear plugging to delamination. Also, we will use deformable bullets. Therefore, bullet deformation will also absorb impact energy.

The difference between this research and other studies is that this study concentrated on layered rather than solid structures. We could not find any studies based on layered structures that can affect the failure mode. A previous study [16] reported shear plugging to be the dominant failure mode; however, delamination may be the dominant failure mode in layered structures, that is, the layered structures may change the behaviour of impact energy absorption from shear plugging to delamination. In addition, a deformable bullet was used in this study. Therefore, bullet deformation may also result in impact energy absorption.

2 Specimen preparations

The materials used in this experimentation are given in Table 1. First, ten layers of fabric were stacked together, mixed with resin and hardener, vacuumed, and left in room temperature for 24 h until it was fully cured. We made ten plates using this technique. The thickness is around 2 mm per plate. After completion, we bonded the ten plates together using epoxy resin, put it in the hot press, and pressed it at 75°C for 2 h until it was totally bonded and was completed. Therefore, the plates were in the form of layered structures, as seen in Figure 1. The total thickness is around 20 mm.

Figure 1

Side view of the composite plate specimens: (a) 27.1 mm CFRP, (b) 20.6 mm GFRP, and (c) 30.2 mm KFRP.

3 Test standard, set-up, and procedure

3.1 Mechanical properties

Tensile and compression tests were carried out to obtain the mechanical characteristics of the materials used during ballistic tests, which are GFRP, KFRP, and CFRP. The tensile tests were done according to ASTM D3039 [17], while the compression tests were carried out using ASTM D3410 [18]. The stress–strain curves are given in Figure 2. The resulting tensile and compression strengths and modulus of elasticity are given in Table 2.

Figure 2

Tensile and compressive stress–strain of (a) GFRP, (b) CFRP, and (c) KFRP coupon tests.

Table 2

Mechanical characteristics of GFRP, CFRP, and KFRP specimens

	GFRP	CFRP	KFRP
Tensile modulus (GPa)	20.22	55.51	16.50
Tensile strength (MPa)	547.87	714.42	544.89
Maximum tensile strain (%)	2.87	1.29	3.44
Compressive modulus (GPa)	20.1	52.50	21.39
Compressive strength (MPa)	309.6	261.6	57.6
Maximum compressive strain (%)	1.6	0.5	0.34

3.2 Ballistic tests

The ballistic impact experiment was designed to follow the NIJ 0108.01 standard [19] as it sets the minimum performance requirements and methods of testing for ballistic resistance protective material. Although this ballistic resistance test is not aimed at creating an actual standardized ballistic resistance protective material, NIJ 0108.01 is still used so that the quality of the ballistic performance of each specimen can be compared with a recognizable standard. Table 3 shows the standard, while Figure 3 shows the test set-up according to the NIJ 0108.01 standard.

Table 3

Ballistic performance requirements based on NIJ 0108.01 [16]

Armor type	Test ammunition	Nominal bullet mass	Suggested barrel length	Required bullet velocity	Required hits per armour specimen	Number of allowed failure
I	22 LRHV	2.6 g	15–16.5 cm	320 ± 12 m/s	5	0
	Lead	40 g	6–6.5 in	1,050 ± 40 ft/s
	38 Special RN	10.2 g	15–16.5 cm	259 ± 15 m/s	5	0
	Lead	158 g	6–6.5 in	850 ± 50 ft/s
II-A	357 Magnum	10.2 g	10–12 cm	381 ± 15 m/s	5	0
	JSP	158 g	4–4.75 in	1,250 ± 50 ft/s
	9 mm FMJ	8.0 g	10–12 cm	332 ± 12 m/s	5	0
		124 g	4–4.75 in	1,090 ± 40 ft/s
II	357 Magnum	10.2 g	15–16.5 cm	425 ± 15 m/s	5	0
	JSP	158 g	6–6.5 in	1,395 ± 50 ft/s
	9 mm FMJ	8.0 g	10–12 cm	358 ± 12 m/s	5	0
		124 g	4–4.75 in	1,175 ± 40 ft/s
III-A	44 Magnum	15.55 g	14–16 cm	426 ± 15 m/s	5	0
	Lead SWC Gas	240 g	5.5–6.25 in	1,400 ± 50 ft/s
	Checked
	9 mm FMJ	8.0 g	24–26 cm	426 ± 15 m/s	5	0
		124 g	9.5–10.25 in	1,400 ± 50 ft/s
III	7.62 mm	9.7 g	56 cm	838 ± 15 m/s	5	0
	308 Winchester	150 g	22 in	2,850 ± 50 ft/s
	FMJ
IV	30-06 AP	10.8 g	56 cm	868 ± 15 m/s	1	0
		166 g	22 in	2,850 ± 50 ft/s

AP – armor piercing; FMJ – full metal jacket; JSP – jacketed soft point; LRHV – long rifle high velocity; RN – round nose; SWC – semi-wadcutter.

Figure 3

Ballistic test set-up (a) and real experimental set-up (b) according to NIJ 0108.01 [19].

The specimens will be shot five times with a “fair shot” with a 9 mm FMJ RN moving at the velocity of around 358 m/s. For this research, the 9 mm FMJ used will be a 9 mm × 19 mm MU1-TJ bullet that will be shot with a G2 Elite handgun. The basic specifications of the gun and bullet are shown in Tables 4 and 5, as well as in Figures 4 and 5.

Table 4

G2 elite specifications

Parameter	Value	Unit	Source
Capacity	15	rounds	Manufacturer
Barrel length	127	mm	Manufacturer
Barrel rifling	6	grooves	Manufacturer
Grooves	1:10 right hand twist	—	Manufacturer
Weight	1 ± 0.05	kg	Manufacturer
Overall length	221	mm	Manufacturer
Height	139	mm	Manufacturer
Front sight	Fixed	—	Manufacturer
Rear sight	Adjustable/fixed	—	Manufacturer
Effective range	25	mm	Manufacturer
Caliber	9 × 19	mm	Manufacturer
Operation/action	Single action, semi auto	—	Manufacturer

Table 5

MU1-TJ bullet specification

Parameter	Value	Unit	Source
Cartridge
Weight	12.26	g	Manufacturer
Rim thickness	1.25	mm	Manufacturer
Extractor length	8.8	mm	Manufacturer
Bullet
Length	15.7	mm	Manufacturer
Weight	8	g	Manufacturer
Core material	Lead antimony	—	Manufacturer
Jacket material	Brass 72 (CuZn 28)	—	Manufacturer
Type	Full Metal Jacket Round Nose (FMJ RN)	—	Manufacturer
Case
Material	Brass 72 (CuZn 28)	—	Manufacturer
Type	Rimless, conical, and centrefire	—	Manufacturer
Primer
Type	Non-corrosive, non-mercuric	—	Manufacturer
Propellant
Type	Smokeless powder	—	Manufacturer
Characteristics
Average velocity	380	m/s	Manufacturer

Figure 4

G2 Elite PINDAD.

Figure 5

MU1-TJ bullet: (a) front view, (b) side view, (c) back view, and (d) cross-sectional view.

4 Experimental results and analysis

A ballistic impact experiment based on the NIJ 0108.01 standard conducted on three composite plate specimens (CFRP, GFRP, and KFRP) has been successfully done. The CFRP, GFRP, and KFRP specimens were impacted, respectively, by six bullets, five bullets, and seven bullets. Each specimen was impacted by at least five fair hits. The results show that none of the fair hits fully penetrated any of the specimens, only resulting in partial penetrations. In other words, all specimens have successfully stopped ballistic impact from five Type II NIJ 0108.01 standardized fair hits.

Every impact velocity of the bullet has been recorded successfully by the light gates and chronograph. The averaged velocity for each fair hit for the CFRP, GFRP, and KFRP specimens, respectively, are 357.42, 357.80, and 355.72 m/s. All the recorded fair hit velocities conform to the Type II armour specifications for a 9 mm FMJ bullet, which is 358 ± 12 m/s.

The final experiment data sheet is presented in Table 6. The hit coordinate is written in mm, where the most bottom left point of the specimen shall be considered the origin point (0, 0) while the most top right point of the specimen is considered the point (200, 200). Penetration depth is only measured for one impact since each specimen is only cut once. The purpose of this cross-sectional cut is so that the through thickness damage can be observed. For CFRP, GFRP, and KFRP specimens, the line of the cut is aligned to impact hole numbers 3, 1, and 2, respectively (refer to the dotted lines in Figures 6, 8 and 10). Those impact holes are chosen because, in their respective specimen, they are the most isolated penetration holes compared to the other ones. This consideration is taken so that the observed through-thickness damage is mostly caused by the corresponding bullet only and not caused by cumulative damage of several impacts.

Figure 6

CFRP specimen impacted front (the dotted line shows the cutting location).

Another important criterion for the impact problem is specific energy absorption (SEA). SEA is defined as the energy absorbed per unit weight, given in J/kg. Table 6 shows the SEA for different plates, GFRP, CFRP, and KFRP. Table 6 shows that the SEA of GFRP is the highest, with the SEA of 367 J/kg and KFRP is the lowest at 235 J/kg. CFRP is at 350 J/kg. It shows that GFRP is the most efficient in absorbing the ballistic impact for layered thick composite plates.

4.1 Carbon fibre epoxy

The impacted carbon epoxy plate can be seen in Figure 6.

As seen in Figure 6, each impacted hole of the fair hit on the CFRP specimen is very localized, showing a small, damaged area for each hit compared to the bullet impact size itself. No total penetration occurred. From this point of view, each fair hit shows damaged and pulled-out fibre but no apparent delamination. The specimen was then cut by a water jet to show the damaged area through the thickness. The damaged cut is shown in Figure 7.

Figure 7

CFRP specimen sectioned view: (a) left side of the cut, and (b) right side of the cut, mirrored.

Through thickness, damages are more identifiable, as seen in Figure 7. Identifiable damage details have been marked with numbers (correspond the numbers below with the number marks in Figure 7):

1. Delamination of the first two topmost ply groups from impact number 4,

2. Entrance damage of impact number 4,

3. Entrance damage of impact number 3,

4. Delamination of the topmost ply group from impact number 3,

5. Bullet fragments from impact number 3,

6. Ply group delamination from impact number 6,

7. Ply group delamination from poor bonding,

8. Bullet fragments from impact number 4,

9. Delamination of the topmost ply group from impact numbers 3 and 4,

10. Entrance damage of impact number 3,

11. Delamination of a single ply from impact number 3, and

12. Ply group delamination from impact number 6.

From Figure 7, some noticeable remarks about the damages are as follows:

1. Bullet entrance damage (marked 2 and 3) shows clean, brittle, and localized damage. It is also noticeable that there is a slight conical deformation protruding towards the opposite of the bullet’s travel direction.

2. Bullets are extremely deformed and fragmented. They are clearly stuck between delaminated ply groups (marked 5 and 8). Delamination of ply groups creates ample space for bullet fragments to spread and get stuck under.

3. There is delamination of ply groups closer to the back surface of the specimen (marked 6, 7, and 12). It must be noted that they are not directly impacted by the bullets, yet the delamination is still present. It is probable that the reason for this phenomenon is poor bonding between ply groups which then is triggered by the wave propagation and vibrations from several impacts.

4. There is one apparent single-ply delamination (marked 11). However, observing the rest of the specimen, ply group delamination is more prominent. The lack of single-ply delamination in contrast to ply group delamination must be caused by the fact that the bonding and curing process of each ply group is more thorough since it uses the vacuum bagging process while bonding between ply groups is only done with a hot press machine and resin impregnation within fibres are not possible.

4.2 Glass fibre composites

The resulting impacted glass fibre composites are given in Figure 8.

Figure 8

GFRP specimen impacted front (the dotted line shows the cutting location).

Figure 8 shows the impacted front of the GFRP specimen, which was impacted by five fair hits. Consistent with the CFRP specimen, each penetration hole is localized. However, what is prominent with the GFRP specimen is that the damaged area can be seen more clearly around the penetration hole. This is shown by the lighter-coloured circles surrounding each impacted hole, signifying a delamination area on the topmost layers nearest the impacted front. However, this observable delamination damage is more pronounced since glass fibres are generally quite transparent in the first place. Again, no total penetration happened. Figure 9 shows the thickness damage area of the specimens.

Figure 9

GFRP specimen sectioned view: (a) left side of the cut, and (b) right side of the cut, mirrored.

Through thickness, damages of the GFRP specimen are shown in Figure 9. Identifiable damage details have been marked with numbers (correspond the numbers below with the number marks in Figure 9):

1. Damage and delamination from impact number 4,

2. Ply group delamination and bullet fragments from impact number 5,

3. Bullet fragments from impact number 1,

4. Entrance damage of impact number 1,

5. Bullet fragments from impact number 3,

6. Ply group delamination from poor bonding,

7. Damage and delamination from impact number 4,

8. Ply group delamination from impact number 5,

9. Delamination of a single ply from impact number 1,

10. Entrance damage of impact number 1, and

11. Ply group delamination from impact number 1.

Figure 9 shows damages that are consistent with the remarks of Figure 8. Bullet entrance damage shows localized damage and there was an apparent conical deformation towards the opposite of the bullet’s travel direction. Bullets are also extremely deformed and stuck in the delamination space between ply groups. There are ply groups delamination which are not directly connected to the bullet’s path, and single-ply delamination is also present. It is consistent how most of the damages are dominated by ply group delamination.

4.3 Kevlar fibre composites

Figure 10 shows the impacted Kevlar fibre composites.

Figure 10

KFRP specimen impacted front (the dotted line shows the cutting location).

Figure 10 shows the impacted front of the KFRP specimen, which was impacted by five fair hits and two invalid hits. All impacted holes are localized, and damaged fibres are only present very tightly near the bullet entrances. There are several closely placed impacts, which would not count as a fair hit if full penetration occurred. However, even when there were several closely packed impacts, there were still no full penetrations. This shows the ability of KFRP to adequately absorb the ballistic impacts. Figure 11 shows the thickness damaged area of the Kevlar composites.

Figure 11

KFRP specimen sectioned view: (a) left side of the cut, and (b) right side of the cut, mirrored.

Through thickness, damages of the KFRP specimen are shown in Figure 11. Identifiable damage details have been marked with numbers (correspond the numbers below with the number marks in Figure 11):

1. Bullet fragments from impact number 1,

2. Ply group delamination from poor bonding,

3. Ply group delamination from impact number 1,

4. Entrance damage of impact number 2,

5. Bullet fragments and delamination of ply group from impact number 2,

6. Delamination of a single ply from impact number 1 and 7,

7. Bullet fragments from impact number 1 and 7 stuck between ply groups,

8. Delamination of ply groups from impact number 1 and 7,

9. Delamination of ply groups from impact number 2,

10. Delamination of a single ply from impact number 2,

11. Entrance damage of impact number 2,

12. Delamination of a single ply from poor bonding, and

13. Delamination of a single ply from impact number 2.

Figure 11 shows similar damage results compared to Figures 7 and 9. What is most noticeably different for the KFRP specimen is that single-ply delamination is more prominent. Considering that seams between ply groups for KFRP are also the most noticeable, it can be safely stated that KFRP has the least thorough resin impregnation for inner ply bonding and bonding between ply groups. However, even when this is the case, it cannot be directly stated that this phenomenon causes a lack of ballistic performance since the KFRP plate still absorbed every bullet’s kinetic energy.

Ultimately, all the manufactured impact targets were adequately performed when impacted with a Type II NIJ 0108.01 ballistic threat, which is the most common in the context of firearm threats. Furthermore, the penetration depth for observed impact is only 11.10% (CFRP), 9.95% (GFRP), and 22.22% (KFRP). This shows that these targets could still absorb more kinetic energy from a ballistic impact, considering that there is no noticeable back-face deformation for all of them.

The resulting comparisons between the specimens are shown in Table 7, which shows conclusively a comprehensive visual observation comparison between all specimens before and after the experiment. It should be noted that while the GFRP plate resulted in the best relative ballistic performance in terms of bullet penetration percentage, it is also the heaviest. While KFRP is the lightest of them all, it is the most expensive, and the relative penetration is the largest. On the other end, CFRP has the middle of the pack density, with only 11% bullet penetration.

Table 7

Ballistic impact specimen and experiment results comparison

Parameter	Specimen
	CFRP	GFRP	KFRP
Specimen similarities
Impact area dimensions	200 mm × 200 mm
Number of ply groups	10
Total number of plies	100
Fibre type	Woven
Matrix type	Epoxy resin EPR 174
Manufacturing method	Hand-layup + vacuum bag + hot press machine
Specimen comparable differences
Thickness	27.12 mm	20.6 mm	30.2 mm
Weight	1.46 kg	1.51 kg	1.39 kg
Fibre volume fraction	50.00%	56.90%	56.50%
Fibre mass fraction	58.70%	74.30%	58.20%
Density	1.36 × 10−6 kg/mm3	1.88 × 10−6 kg/mm3	1.17 × 10−6 kg/mm3
Weave type	Twill weave	Plain weave	Plain weave
Experiment result similarities
No. of penetrating fair hits	0 out of 5
Damage characteristics	Localized damage near bullet entry point with fibre breakages
	Conical deformation facing opposite of bullet’s directory on impact surface
	Extremely deformed and fragmented bullets
	Prominent amount of delamination between ply groups
	No noticeable back surface deformation or damage
	Indirect impact delamination (due to vibration or poor bonding)
Experiment result differences
Single ply delamination	Visible	Visible	Prominent
Impact surface delamination	Indiscernible	Prominent	Indiscernible
Ply group bonding seams	Indiscernible	Visible	Prominent
Penetration depth	3.01 mm (11.10%)	2.05 mm (9.95%)	6.71 mm (22.22%)

The experiments show that there was no total penetration during the ballistic impact. All specimens were able to withstand the ballistic impact. Delamination was the major damage failure mode on the plates. It seems that the layered structure of the plate makes the delamination the prominent major failure mode, compared to shear plugging as given by Naik and Doshi [16]. Therefore, delamination should be included in the numerical analysis of ballistic impact on the composite layered thick plate.

5 Conclusion

Ballistic impacts on CFRP, GFRP, and KFRP thick plates have been carried out. The total thickness of the plates was 20 mm, comprising 100 layers of carbon, glass, and Kevlar fabric cloths with epoxy resins. The plates were manufactured by hand lay-up and vacuum methods. The manufacturing process was chosen so that the plates were in the form of thick layered plates. The plates were then fired with a bullet according to the NIJ 0108.01 standard in the special firing room. The bullet was MU1-TJ bullet 9 × 19 calibre, firing with a G2 Elite PINDAD pistol. The bullet mass is 8 mg with a firing speed of 358 m/s. The results were that no total penetration happened. All plates survived the fire tests. The penetration depth was 3.01, 2.05, and 6.71 mm for CFRP, GFRP, and KFRP, respectively. The SEA was in the order of 367, 350, and 235 J/kg for GFRP, CFRP, and KFRP, respectively. Therefore, in this experiment, the GFRP performs better compared to the other two plates. The delamination was the major failure mode during the tests. Therefore, in the case of layered structures, delamination absorbs substantial impact energy compared to shear plugging in the case of solid structures. Further research should be done on these important findings.