The effects of blast-induced fragments on cellular materials

International Journal of Impact Engineering 92 (2016) 50–65

Contents lists available at ScienceDirect

International Journal of Impact Engineering j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j i m p e n g

The effects of blast-induced fragments on cellular materials N. Ranwaha, S. Chung Kim Yuen * Blast Impact and Survivability Research Unit (BISRU), Department of Mechanical Engineering, University of Cape Town, Private Bag, Rondebosch 7701, South Africa

A R T I C L E

I N F O

Article history: Received 15 July 2014 Received in revised form 23 November 2015 Accepted 9 December 2015 Available online 17 December 2015 Keywords: Blast load Fragmentation Impact Cellular materials Energy absorbers

A B S T R A C T

This paper presents the results of an experimental study to characterise blast-induced fragments and to understand how different cellular materials alleviate the damage caused by the blast-induced fragment. In the experimental arrangement, a front plate, 106 mm in diameter, is subjected to a localised blast load to generate a cap fragment (Mode IIc failure) to impact a rear plate of similar dimensions, located parallel and offset from the front plate. Different charge diameters and masses are used to create fragments of different sizes and masses (4.1 g to 12.5 g) propelled at different speeds (244 m/s to 741 m/s). Various cellular materials (aluminium foam, aluminium honeycomb, balsa wood, Corecell M80 foam, Divinycell H200 PVC foam and polyurethane 200 foam) of thicknesses 40 mm and 60 mm are placed in front of the rear plate to act as energy absorbers. The damage caused by the fragment and the protective performance of the cellular materials are quantified by means of the maximum deflection of the rear plate. The results indicate that the cellular materials alleviate the damage incurred to the rear plate, with different materials absorbing different amounts of impact energy. For the range of experiments carried out and foam investigated, the Divinycell foam provided the best protection while Corecell foam offered the least resistance to damage. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Fragments released upon the detonation of either improvised explosive devices (IED’s) or ordnance or explosion can be classified as either primary fragments that are generated by the casings or containers which surround the explosive source or secondary fragments that are formed by nearby pieces of structures where the explosion occurred. Over the past few years, the number of global IED incidents has risen [1,2], whereby casualties and damage to structures are caused by the release and impact of blast-induced fragments. These fragments are typically generated and propelled in an unpredictable and uncontrollable manner from the source. A number of studies have been conducted on the phenomenon of fragmentation to provide guidance for the prediction of fragment loadings on structures as a result of accidental explosions in or near the structures based on open field tests, for example References 3,4. Arnold and Rottenkolber [5] presented experimental data on the fragmentation of thin aluminium and steel spherical shells subjected to internal blasts by studying the speed distribution of the generated fragments. Between 2800 and 5000 fragments travelling between speeds of 2700 m/s and 3000 m/s were ob-

* Corresponding author. Blast Impact and Survivability Research Unit (BISRU), Department of Mechanical Engineering, University of Cape Town, Private Bag, Rondebosch 7701, South Africa. Tel.: +27 21 650 4809; Fax: +27 21 650 4808. E-mail address: [email protected] (S. Chung Kim Yuen). http://dx.doi.org/10.1016/j.ijimpeng.2015.12.003 0734-743X/© 2015 Elsevier Ltd. All rights reserved.

served for the various sizes of the spherical charges (66 mm– 180 mm diameter range). With refined physical and mathematical models of fracture, several authors [5–9] have developed and improved numerical simulation procedures to predict the initial speeds and size distribution of fragments generated by the shattering of munition casing under the blast wave propagated by the explosive it contains. In some cases, close correlation was found between the predictive model and the actual experimental findings. Generally, fragments generated from objects in contact with the detonating explosives differ in mass, size, shape and velocity. Although irregular in geometry, there have been numerous attempts to statistically classify the ‘uncontrolled’ fragments in terms of mass distribution and velocity [3,10,11]. The generation of blast-induced fragments, however, can be “designed” to be more controlled in terms of the size and mass of the fragments, based on the quantity and size of the explosive. Nurick et al. [12,13] undertook investigations on the response of circular and square thin mild steel plates to blast loads over the entire area of the plate. Disc fragments were generated from the exposed area of the plate by means of Mode II (tensile tearing at the boundary) and Mode III (transverse shearing at the boundary) failure modes. Subsequently, when subjected to localised blast loads, fragmentation in the form of capping (Mode IIc failure mode – the ejection of a cap fragment from the central region of the plate where the load was applied) was attained [14–16]. The damage caused by these “capped” fragments was studied by Nurick et al. [17,18] whereby mild steel square tubes were subjected to localised blast loading. The opposite faces of the

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Nomenclature

δ d df εd E Abs Ek mf η ρ SEA σ0 σ pl vf

(rear) plate maximum deflection plate diameter fragment diameter onset strain of densification energy absorbed in quasi-static loading (fragment) kinetic energy fragment mass energy absorption efficiency density specific energy absorbed static yield stress plateau stress fragment speed

51

comb sandwich panels subjected to out-of-plane projectile impacts were reported on by Hoo Fatt and Park [25]. It was found that during perforation, the incoming projectile sheared off a compressed plug of the sandwich panel (front sheet, core and back sheet). Despite the numerous studies on the blast loading of structures reporting on capping, there is a need to gain insights for better understanding of the generation of capping as a method of fragmentation, its characteristics, the damage it may cause and how the damage can be alleviated. This paper presents the experimental results on the characteristics of blast-fragments which are blastinduced by explosive charges of different sizes and masses. The damage caused by the fragment is quantified by assessing the maximum deflection of a second witness target plate. The damage mitigation effectiveness of different energy absorbing cellular materials was also investigated via the comparisons of the deformation of the target plate after fragment impact. 2. Material characterisation

square tubes were assumed to behave as two parallel rectangular plates. The charge was applied on one side of the tube and the damage as a result of the generated fragment was assessed by the deformation of the opposite side of the tube. The observations in the experimental study [17] showed that the fragment caused inelastic deflection on the opposite face, while fusion of the fragment onto the opposite face was observed at higher impulses. In the validated numerical model [18], the speeds of the fragments were predicted to be up to 835 m/s. With the threat posed by high velocity fragments to nearby structures and development of new materials, there are increased interests to seek ways to mitigate the impact that fragments may cause. The energy absorbing or impact resisting ability of different materials when subjected to impact by fragments or other projectiles has been widely reported. The response of natural fibre composites; comprised of flax, hemp and jute fabric – reinforced polypropylene composites subjected to impact by fragment-simulating projectiles was studied by Wambua et al. [19]. The dominant failure modes of the composites include fibre fracture, delamination and shear cutout. In terms of energy absorption, the composite hybrid structures performed better than mild steel and the plain composites. Ceramic armour system dissipated impact energy by cracking as reported by López-Puente et al. [20]. In an experimental study by Yungwirth et al. [21], pyramidal micro-architectured lattice (trusses) were used as cores of steel and aluminium alloy sandwich panels under projectile impact. The findings indicated that the panels exhibited similar failure patterns. The front face sheet failed by ductile hole enlargement while the rear plate failed by petalling as the projectile perforates through the material. The cores did not dissipate any significant amount of energy. The most common energy absorbing materials used to alleviate damage from projectile impact are cellular materials in the form of foams (metallic and polymeric), wood or honeycomb. These materials with inherent cellular structures are often used as cores of sandwich panels to dissipate impact energy. Hou et al. [22] reported on investigations on aluminium foam sandwich panels under impact by flat, hemispherical and conical nosed projectiles. It was found that the front face sheet fails by circular hole formation with insignificant deflection, while the core is subjected to tunnelling during partial or complete perforation. Flexible polyurethane foam was subjected to projectile impact in an investigation carried out by Zaretsky et al. [23]. Beyond a projectile speed of 43 m/s, the polyurethane foam turned into powder form upon impact. Atas and Sevim [24] subjected sandwich panels with either polyvinylchloride (PVC) foam cores or balsa wood cores to impact from drop weights. Balsa wood was observed to be stiffer than PVC in impact but debonded from the face sheets easier than the PVC. Investigations on aluminium honey-

In the characterisation of cellular materials, the plateau stress and onset strain of densification on the compressive stress vs. strain curve are prime indicators of the energy absorbing ability of a material. While the plateau stress can be inferred from the stressstrain curve, there are different methods to determine the onset strain of densification as reported by Li et al. [26]. A more consistent approach to determining the onset strain of densification, as suggested by Li et al. [26], is based on the energy efficiency method (Eq. 1) that was proposed by Avalle et al. [27].

η=

E Abs σ (ε )

(1)

η is the energy absorbing efficiency, σ (ε ) is stress at strain ‘ ε ’ and E Abs is energy absorbed from a strain of 0 to ‘ ε ’. The energy absorbed is essentially the area under the stress-strain curve from the original unstrained position to a generic strain ‘ ε ’, as described by Eq. 2. The onset strain of densification is thus determined as the strain at which the energy absorbing efficiency is at its maximum. ε

E Abs = ∫ σ (ε )d ε

(2)

0

In this study, six materials, as shown in Fig. 1: aluminium foam (relative density 8.5%), aluminium honeycomb (relative density 3.8%), balsa wood, Corecell M80 foam, Divinycell H200 foam (PVC foam)

Fig. 1. Photograph of all the energy absorbing materials used in the study.

52

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 2. Quasi-static compressive stress vs. strain for the materials used in the blast tests (cross-head speed = 10 mm/min).

and polyurethane 200 foam, are investigated as energy absorbing materials to mitigate the damage caused by a blast-induced fragment. These materials were selected based on availability. The results of the characteristics of these materials under quasi-static compression tests are shown in Fig. 2, with the results summarised in Table 1. Divinycell foam exhibited the highest plateau stress (5 MPa) while balsa wood showed the lowest plateau stress (0.6 MPa). The energy efficiency method (as discussed by Avalle et al. [27] and Li et al. [26]) is used to determine the onset strain of densification (the strain at which the energy absorption efficiency is maximum). Based on the energy efficiency method, balsa wood has the highest onset strain of densification at 65%. The polyurethane foam is from two separate batches, hence ‘A’ and ‘B’ 3. Blast testing 3.1. Experimental setup Two plates are used in each experiment – a front plate that is exposed to a localised blast load to generate a fragment and a rear plate to quantify the damage caused by the fragment and assess the protective performance of the cellular materials. Both plates, made from mild steel, are nominally 1.6 mm thick with a circular exposed

area of diameter 106 mm. The experimental arrangements, shown in Fig. 3(a and b), consist of two pairs of clamping plates to secure the front and rear plates at the outer boundary, spacer rods to provide a constant separation between the two plates, a speed trap assembly to determine the speed of the fragment and a containment tube to hold and locate the energy absorbing materials in front of the rear plate. The test rig is mounted onto a ballistic pendulum used to determine the impulse imparted to the test plate upon detonation of the explosive. The front plate is loaded by detonating plastic explosive (PE4) shaped into circular disc of prescribed diameter (27 mm, 36 mm and 43 mm) and placed onto a polystyrene pad located on the front plate. The polystyrene pad is assumed to burn upon detonation and has no significant outcome on the response of the plate. A 1 g leader charge is used to attach the detonator to the main charge. The charge masses are varied between 7 g and 11 g (which includes the 1 g leader charge) to provide a range of speeds of the fragment. 3.2. Experimental plan and execution All blast tests are undertaken with two mild steel plates of an exposed circular region of 106 mm diameter and a thickness of

Table 1 Summarised results of the quasi-static compression tests on the materials. Material

Aluminium foam Aluminium honeycomb Balsa wood Corecell M-80 foam Divinycell H200 foam Polyurethane 200 foam A Polyurethane 200 foam B

Mean density

Mean mass

Mean size

Onset strain of densification

Plateau stress

Energy absorbed until onset of densification

ρ (kg/m3)

m (g)

B × L × H (mmxmmxmm)

εd (%)

σ pl (MPa)

SEA/Mass

SEA/Volumetric

(MJ/kg)

(GJ/m3)

163.0 1194.4 251.7 702.6 510.6 531.3 447

34.33 127.1 19.8 62.4 95.2 101.1 81.9

229.7 103.4 77.3 85.8 185.6 191.0 191.3

3.72 0.84 1.37 0.85 5.96 2.96 2.86

25.3 × 25.4 × 25.2 25 × 25 × 13 30 × 30 × 19.7 25.5 × 25.4 × 15.3 35.4 × 36 × 25.2 24.7 × 24.7 × 24.4 24.8 × 24.7 × 24.4

54 59 65 54 60 53 52

1.04 1.9 0.55 1.1 5.0 3.0 2.4

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

53

Fig. 3. Sectioned schematic of experimental set-ups (a) without energy absorbers and (b) with energy absorbers.

1.6 mm. The material is found to have a static yield stress of 288.2 MPa and a fracture strain of 34.1%. The blast propagated from the charge placed in close proximity to the front plate generates a fragment from the centre of the front plate which impacts the rear plate. Table 2 lists the input parameters of all the blast tests conducted in the study. In Series 1, no energy absorber is used to protect the rear plate while the plate separation distance is 150 mm and 190 mm; with cylindrical PE4 charges of diameters 27 mm, 36 mm and 43 mm and increasing in mass from 7 g to 11 g. The energy absorbers, shown in Fig. 1, are cut into discs of 106 mm diameter to fit snugly into the containment tube. In the subsequent series of tests (Series 2–5), all the six energy absorbing materials are tested at 40 mm thickness. A fixed charge diameter of 36 mm is used and the charge mass increases from 7 g to 11 g. Interim analyses on the relationship between kinetic energy of the fragment and maximum rear plate deflection suggest that Divinycell foam provides the best mitigation of damage to the rear plate. Balsa wood appears to be at the mean region of all tested energy absorbing materials while polyurethane foam shows a different trend as it exhibits a high damage alleviation at lower fragment kinetic energies and deteriorating at the highest fragment kinetic energy. Consequently, these three materials are tested in a series of tests at a thickness of 60 mm, with the charge diameter of 36 mm and the charge mass increasing from 7 g to 11 g. In Series 4, a single 40 mm thick disc or two 20 mm thick discs or four 10 mm thick discs of polyurethane foam is used to inves-

tigate the effects of varying the layering arrangement on damage mitigation. A final parametric series of tests (Series 5) is conducted with balsa wood, Divinycell foam and polyurethane foam to investigate the effects of subjecting a protected plate to fragments of different sizes (in diameter and mass) but of similar impact kinetic energies. A total of 107 blast tests are conducted when all the input parameters are taken into account. 4. Experimental results and discussions 4.1. Impulse from the blast and front plate failure For the particular experimental arrangements as described in this study, the minimum charge mass at which the front plate fails by mode IIc (capping) is 7 g. As impulse is imparted from 7 g charge mass to the front plate, it is noted that for all charge diameters, a fragment is ejected from the central region of the plate, leaving inelastic deflection with little to no additional tearing in the radial direction. As the charge mass is increased, petalling on the front plate is observed, in addition to capping and inelastic deflection. A typical evolution of front plate failure is shown in Fig. 4. Fig. 5 shows the impulse of the blast plotted against increasing charge masses for all three charge diameters. In general and collectively, the impulse is observed to increase linearly with increasing charge mass and the relationship is independent of the charge diameter. Similar observations were made by Nurick and Radford [14]

Table 2 Summary of the experimental plan of the blast tests. Test series

Test parameters

Energy absorbing materials used

Series 1

No energy absorbers Charge diameter: 27 mm, 36 mm and 43 mm Charge mass 7 g–11 g 40 mm thick energy absorbers Charge diameter: 36 mm Charge mass 7 g–11 g 60 mm thick energy absorbers Charge diameter 36 mm Charge mass 7 g–11 g 40 mm thick polyurethane foam (1 × 40 mm, 2 × 20 mm and 4 × 10 mm) Charge diameter 36 mm Charge mass 7 g–11 g 40 mm thick energy absorbers Charge diameter : 27 mm, 36 mm and 43 mm Charge mass 7 g–11 g

None

Series 2

Series 3

Series 4

Series 5

Aluminium foam, aluminium honeycomb, Balsa wood, Corecell foam, Divinycell foam, polyurethane foam Balsa wood, Divinycell foam, polyurethane foam

Polyurethane foam

Balsa wood, Divinycell foam, polyurethane foam

54

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 4. Photograph showing the failure of the front plates when charge mass increases for 43 mm charge diameter.

who reported on four different load diameters (18 mm, 25 mm, 33 mm and 40 mm) of PE4 charges for localised loading on mild steel circular plates. 4.2. Fragment characteristics and kinetics In general, the diameter of the fragment was found to be proportionally less than the load diameter, as also shown in Reference 14. The average fragment diameter, mass, minimum and maximum speed for the different charge diameters are listed in Table 3. After each test, the fragment is observed to be hot when it is recovered. For post-impact states of the fragments, some of the fragments are observed to maintain a geometry which is similar to when they are released from the front plate – albeit being slightly flattened at the point on the face of the fragment with which impact on the rear plate occurs (state A in Fig. 6). This occurs as a fragment impacts the rear plate with its face and rebounds from the plate or if it fails to perforate the energy absorber. In some cases, the skewed or spinning orienta-

tion of the motion of the fragment is such that it crumples after impact (State B). The fragment is also observed to be released from the front plate as multiple fragments (State C). In some conditions, the fragment gets fused to a section of the rear plate (State D) and in other cases it is observed that the fragment gets slightly shredded or bent outwards from the edge (State E). It is observed that the measured fragment speeds show considerable variation, and hence unpredictability, for a given charge diameter and charge mass. Fig. 7 shows the speed of the released fragments at increasing charge masses, for tests conducted at 36 mm charge diameter (79 out of the 107 tests in the study are conducted at 36 mm diameter). The dotted linear bounds are the least squares trends running through the minimum and maximum values of fragment speeds for the charge masses 7 g, 8 g, 9 g, 10 g and 11 g. Even though there is considerable experimental variation, there is a clear increasing trend in the measured fragment speed as charge mass is increased from 7 g to 11 g. The range of fragment speeds for each charge mass increases with increasing charge mass. At a charge mass of 7 g, the range in fragment speed is 96 m/s, with a

Fig. 5. Impulse against varying charge mass for tests conducted using all three charge diameters.

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Table 3 Summarised details of the released fragments from the blast tests. Charge diameter (mm) Average fragment diameter d f Average fragment mass m f Minimum fragment speed (v f )min Maximum fragment speed (v f )max

(mm) (g) (m/s) (m/s)

27

36

43

25.73 4.11 322.6 666.7

34.04 8.41 312.5 740.7

40.08 12.59 243.9 500.0

mean speed of 371 m/s and variance of 8.2%, and the range increases to 215 m/s at 11 g charge mass, with a mean speed of 630 m/s and variance of 9.6%. The plot of fragment kinetic energy against charge mass for tests conducted at 36 mm charge diameter is shown in Fig. 8. With the experimental variation of fragment speed observed, it is

55

expected that the resultantly determined kinetic energy also shows substantial variation following the equation of fragment kinetic energy:

1 E k = m f v f2 2

(3)

where E k is the kinetic energy, m f is the mass of the fragment and v f is the speed of the fragment. The variation of fragment speed and the fact that the speed is squared in determining the kinetic energy, a resulting wide range in the values of the energy for each charge mass is observed in Fig. 8. At the minimum charge mass of 7 g, the range of kinetic energy is 256 J within a mean variation of 14.6%, and this range increases to 1072 J at a charge mass of 11 g with a mean variation of 20.5%. A quadratic trend is chosen due to the square relation between speed

Fig. 6. Various states of the fragments after impact with the rear plates.

Fig. 7. Speed of the fragment, plotted against charge mass for tests conducted at 36 mm charge diameter.

56

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 8. Fragment kinetic energy against charge mass for tests conducted at 36 mm charge diameter.

and kinetic energy and the R2 correlation value of the regression analysis is 0.74. 4.3. Rear plate damage 4.3.1. Impact of rear plates without energy absorbers (series 1) The results of this series of tests, where the rear plates are not protected by any energy absorbers, are listed in Table 4. It is noted that in most of the cases, partial tearing is observed on the rear plate. For 36 mm charge diameter, all plates show some level of

tearing, although in some, the tearing is small and thus, the deflection can be measured. For large or full rear plate tearing, the deflection is not measured. Due to the fairly inconsistent material properties of mild steel and due to the fact that the fragment does not always fly towards the rear plate in translation, but with spinning in some cases, the charge mass required for tearing the rear plate is not always at the higher end of the charge mass range. For charge diameters of 27 mm and 36 mm, full tearing is observed on the rear plate at the charge masses of 10 g and 11 g respectively, with the fragment fusing to the section of the rear

Table 4 Summary of results for tests in which no energy absorbers are used.

27 mm charge diameter

36 mm charge diameter

43 mm charge diameter

Test description

Charge mass (g)

Impulse (Ns)

Fragment speed (m/s)

Fragment kinetic energy (J)

Rear plate deflection (mm)

NA00-2707-T1 NA00-2707-01 NA00-2707-02 NA00-2707-03 NA00-2707-04 NA00-2708-01 NA00-2708-02 NA00-2709-01 NA00-2710-01 NA00-3607-01 NA00-3607-02 NA00-3608-T1 NA00-3608-01 NA00-3609-01 NA00-3609-02 NA00-3609-03 NA00-3610-01 NA00-3611-01 NA00-3611-02 NA00-4307-01 NA00-4307-02 NA00-4308-01 NA00-4308-02 NA00-4309-01 NA00-4310-01 NA00-4311-01

7 7 7 7 7 8 8 9 10 7 7 8 8 9 9 9 10 11 11 7 7 8 8 9 10 11

15.7 14.4 14.4 13.9 15.4 15.5 16.4 16.0 17.4 13.9 14.7 16.0 15.3 16.7 17.7 18.9 19.8 19.6 22.0 13.9 14.0 16.9 16.8 18.0 19.5 21.9

444.4 370.4 434.8 476.2 322.6 625.0 666.7 666.7 588.2 400.0 322.6 465.1 434.8 N/A 476.2 408.2 555.6 645.2 526.3 N/A 243.9 322.6 281.7 377.4 476.2 500.0

414.8 267.5 378.1 419.5 169.1 878.9 844.4 844.4 726.6 637.6 430.3 887.0 820.4 N/A 1020.4 676.4 1125.5 1560.9 1260.4 N/A 356.9 619.1 521.7 904.2 1428.6 1612.5

12.31 8.67 10.28 Partial tear* 7.45 Large tear Large tear Large tear Secondary cap 15.7 10.38 Large tear 19.24 Large tear Large tear 14.52 Large tear Secondary cap Large tear No impact Partial tear* 10.61 5.73 16.45 Large tear Large tear

* These partial tears are not accompanied by noticeable inelastic deflection as they occur when the fragment slices the rear plate near the boundary of the exposed area and simply induces a localised opening on the plate.

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

57

Fig. 9. Rear plate deflection against fragment kinetic energy with no energy absorber – showing plate tearing points.

plate which is fully torn. The observation of fusion of the fragment to the rear plate is reminiscent of the observations by Nurick and Bryant [17] on square tubes subjected to localised loading. The phenomenon of the fragment fully tearing a section of the rear plate is referred to as ‘secondary capping’; ‘primary’ capping refers to front plate fragmentation as a result of the localised blast. Fig. 9 shows rear plate deflection against fragment kinetic energy whenno energy absorbers are present. The two plate separation distances are shown and for each of the charge diameters, a linear least squares trend is fitted through points for both

the plate separations with ± 1 plate thickness deflection zone. For a charge diameter of 27 mm, the highest fragment kinetic energy at which rear plate deflection can be measured is 415 J; 820 J for 36 mm charge diameter and for the tests conducted with charge diameter at 43 mm, 930 J is the maximum fragment kinetic energy at which rear plate deflection can be measured. Fragments impacting the rear plate with kinetic energies which are higher than these limits are observed to result in the fragment perforating the rear plate and consequently subjecting it to large or complete tearing such that it is not possible to quantify the plate damage by deflection.

Fig. 10. Entry, exit and sectioned views of the 40 mm thick energy absorbers subjected to impacts from fragment released by 7 g charge masses at 36 mm charge diameter.

58

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 11. Entry, exit and sectioned views of the 40 mm thick energy absorbers subjected to impact from fragments released by 11 g charge masses at 36 mm charge diameter.

4.4. Tests with energy absorbers protecting the rear plate For all the materials, the collective observation is that the introduction of the energy absorbing cellular materials does alleviate damage from the fragment by reducing the rear plate deflection and delaying or preventing any tearing on the rear plate. It is also observed that the general size of the entry opening (ballistic entry) is not affected by the charge mass but by the diameter of the fragment. The exit opening generally gets wider and the deformation on the exit face gets more extensive as charge mass is increased. The damage observed in the rear plate is similar to circular plates subjected to blast load; in some cases, characterised by large in-

elastic deformation, and in other cases, capping and tearing associated with large inelastic deformation. 4.4.1. Tests with 40 mm thick energy absorbers protecting the rear plate (series 2) Figs. 10 and 11 are photographs of the entry, exit and sectioned views of all the materials in the blast tests at charge masses of 7 g and 11 g respectively. Fig. 12 is a comparative plot of rear plate deflection against fragment kinetic energy for all the 40 mm thick energy absorbers in the second series of tests. The summary of the results of the Series 2 of tests conducted with 40 mm thick absorbers is listed in Table 5.

Fig. 12. Comparative plot of rear plate deflection against fragment kinetic energy for all energy absorbers of 40 mm thickness.

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

59

Table 5 Summary of results for the second test series in which 40 mm thick energy absorbers are used.

Al Foam

Al honeycomb

Balsa wood

Corecell foam

Divinycell foam

Polyurethane foam

Test description

Charge mass (g)

Impulse (Ns)

Fragment speed (m/s)

Fragment kinetic energy (J)

Rear plate deflection (mm)

Material lost from absorber (%)

AL40-3607-01 AL40-3608-01 AL40-3609-01 AL40-3610-01 AL40-3611-01 AH40-3607-01 AH40-367.5-01 AH40-3608-01 AH40-3609-01 AH40-3610-01 AH40-3611-01 BA40-3607-01 BA40-367.5-01 BA40-3608-01 BA40-3609-01 BA40-3610-01 BA40-3610-02 BA40-3610.5-01 BA40-3610.5-02 BA40-3611-01 CM40-3607-01 CM40-3608-01 CM40-3608-02 CM40-3608-03 CM40-3608-04 CM40-3609-01 CM40-3610-01 CM40-3611-01 DV40-3607-01 DV40-3608-01 DV40-3608-02 DV40-368.5-01 DV40-3609-01 DV40-3610-01 DV40-3610-02 DV40-3611-01 PU40-3607-01 PU40-3608-01 PU40-3609-01 PU40-3610-01 PU40-3611-01 PU40-3611-02

7 8 9 10 11 7 7.5 8 9 10 11 7 7.5 8 9 10 10 10.5 10.5 11 7 8 8 8 8 9 10 11 7 8 8 8.5 9 10 10 11 7 8 9 10 11 11

13.6 15.1 16.6 18.9 21.7 13.9 15.1 15.7 15.5 20.5 20.9 13.3 15.0 15.1 17.7 18.2 19.3 20.2 20.6 21.6 N/A 14.4 15.8 16.4 16.0 18.1 19.5 21.4 13.6 15.6 16.7 16.4 16.5 19.4 18.5 21.2 13.9 15.8 17.1 20.6 21.3 21.8

392.2 434.8 512.8 606.1 666.7 370.4 370.4 512.8 526.3 571.4 645.2 363.6 444.4 512.8 512.8 N/A* 689.7 588.2 571.4 740.7 400 476.2 487.8 506.3 408.2 540.5 606.1 666.7 408.2 571.4 370.4 476.2 540.5 645.2 666.7 714.3 370.4 444.4 487.8 526.3 571.4 606.1

645.9 765.6 1117.7 1506.0 1955.6 541.8 530.2 1078.2 1191.1 1306.1 1644.1 489.3 811.9 1038.8 1091.4 N/A* 1997.6 1522.5 1523.3 2332.0 624.0 997.7 1035.1 1097.3 687.2 1110.3 1322.3 2222.2 641.4 1322.4 610.4 968.3 1329.4 1956.3 1911.1 2168.4 576.1 870.1 1069.6 1214.7 1312.7 1506.0

10.18 11.11 14.17 16.37 17.14 11.62 13.11 15.32 15.97 17.42 Secondary cap 9.46 13.66 14.49 16.13 18.49 19.40 Large tearing Large tearing Large tearing 12.68 17.93 15.95 Large tearing 13.73 16.07 16.44 Large tearing 8.57 13.77 10.35 12.75 15.16 16.53 13.75 Large tearing 9.36 11.78 14.72 16.11 17.15 16.84

4.3 5.1 12.6 0.4 6.3 7.0 4.8 7.5 7.8 8.7 9.4 19.9 8.8 4.3 27.3 30.2 17.9 10.9 13.9 25.4 7.7 5.8 3.7 8.9 5.6 11.1 6.0 13.8 2.5 3.7 2.1 3.1 5.3 7.0 2.9 9.0 19.3 12.6 19.0 20.0 22.8 21.6

* Data was not properly captured.

For aluminium foam, it is generally observed that as the fragment perforates through, a compressed foam plug is sheared off the path of the fragment. At all the charge masses, no tearing is observed in the rear plate when aluminium foam protects the rear plate. An average of 5.7% of the original mass of aluminium foam is unrecoverable during perforation (material is lost/detached from the original disc as a result of the fragmentation damage). For aluminium honeycomb, which comprises of three stacked layers of an approximate thickness of 13 mm, it is generally observed that the layers bend as the fragment perforates through, with the largest opening at the layer closest to the rear plate. At the highest charge mass (11 g), aluminium honeycomb shows extensive bending of the layers (see Fig. 11) and the rear plate is observed to fail by secondary capping. An average material fraction of 7.5% is unrecoverable for the honeycomb during perforation by the fragment. At a charge mass of 7 g (see Fig. 10), balsa wood is observed to be at an onset of being perforated by the incoming fragment and the exit face shows cracks in the general direction of the grain of the wood. As the charge mass is progressively increased, small chunks of wood are more extensively lost at the exit face during failure, in addition to the material lost as the penetrating fragment creates the tunnel in the absorber. Above 10 g charge mass, large tearing is observed in the rear plate protected by the wood. An average mass loss of 17.6% is noticed for balsa wood. Corecell foam

Fig. 13. Entry, exit and sectioned views of 60 mm thick energy absorbers subjected to fragment impact from 7 g charge mass.

60

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 14. Entry, exit and sectioned views of 60 mm thick energy absorbers subjected to fragment impact from 11 g charge mass. Fig. 15. Change in rear plate deflection against kinetic energy when balsa wood thickness increases to 60 mm.

shows visible soot deposit, and its texture and colour, after the fragment passes through the foam, suggest that it is partially scorched by the hot fragment. This change of texture becomes more evident as the charge mass increases. The rear plate shows large tearing at 8 g but does not tear again until the charge mass is increased to 11 g. An average of 7.8% of the mass of Corecell foam is unrecoverable after fragment impact. Divinycell foam, from which the least material mass fraction is unrecovered (an average of 4.5%), exhibits cross-cracking after fragment penetration which originate from the entry surface of the foam and surround the tunnel and propagate through the thickness of each absorber disc at an oblique angle, confirming observations made by Kepler [28] on Divinycell foam subjected to indenting impact. Some fragments are found tightly wedged in the Divinycell foam after impact and there appears to be slight densification of the foam material as there is hardening which is not accompanied by signs of scorching. At a charge mass of 11 g, large tearing of the rear plate is observed when it is protected by 40 mm thick Divinycell foam. Polyurethane foam is observed to fail by crushing of the foam material to powder. Similar observations on flexible polyurethane foam subjected to projectile impact were made by Zaretsky et al. [23] at impact speed of above 43 m/s. Rebound motion of the fragment appears unrestrained by the foam. A mean fraction of 19.2% of the

mass of polyurethane foam is unrecoverable during fragment impact. No tearing is observed on rear plates protected by 40 mm thick polyurethane foam. From the comparative plot in Fig. 12, it is clear that the 40 mm thick energy absorbers decrease the rear plate damage. Corecell foam, showing the highest deflection against fragment kinetic energy, appears to alleviate the damage the least while Divinycell foam decreases the rear plate deflection by the biggest margin. The damage alleviation of polyurethane appears to be similar to the Divinycell foam up to fragment kinetic energies of about 700 J, beyond which it deteriorates until damage alleviation is akin to that of Corecell foam for fragments with kinetic energies of about 1400 J. 4.4.2. Tests with 60 mm thick energy absorbers protecting the rear plate (series 3) The entry, exit and sectioned views of balsa wood, Divinycell foam and polyurethane foam after fragment impact when the charge masses are 7 g and 11 g are shown in Fig. 13 and Fig. 14 respectively. The summary of all tests which occur at the series where the absorber thickness is 60 mm is listed in Table 6. It is generally observed that the rear plate deflection decreases (compared to the 40 mm thickness test series).

Table 6 Summary of results for tests conducted on balsa wood, Divinycell foam and polyurethane foam of 60 mm thickness.

Balsa wood

Divinycell foam

Polyurethane foam

Test description

Charge mass (g)

Impulse (Ns)

Fragment speed (m/s)

Fragment kinetic energy (J)

Rear plate deflection (mm)

Material loss from absorber (%)

BA60-3607-01 BA60-3608-01 BA60-3609-01 BA60-3610-01 BA60-3611-01 BA60-3611-02 DV60-3607-01 DV60-3608-01 DV60-3609-01 DV60-3610-01 DV60-3611-01 PU60-3607-01 PU60-3607-02 PU60-3608-01 PU60-3609-01 PU60-3610-01 PU60-3611-01

7 8 9 10 11 11 7 8 9 10 11 7 7 8 9 10 11

14.6 16.6 18.3 19.3 22.4 21.4 14.6 16.8 18.6 20.1 22.0 14.6 15.2 16.4 19.1 20.2 21.4

357.1 425.5 487.8 526.3 606.1 571.4 339.0 392.2 425.5 512.8 540.5 408.2 312.5 408.2 444.4 526.3 588.2

548.5 766.9 1070.8 1205.0 1395.8 1583.7 488.4 730.5 805.8 1196.6 1387.9 661.4 405.3 742.2 937.3 1113.6 1417.0

4.74 9.86 14.17 16.25 15.45 14.99 1.67 2.48 5.46 9.98 13.06 3.72 1.25 8.34 10.45 11.42 13.79

31.3 20.7 14.9 24.1 13.0 18.1 1.1 4.7 ~0 4.7 4.6 10.3 5.2 10.5 18.2 13.8 14.4

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 16. Change in rear plate deflection against kinetic energy when Divinycell foam thickness increases to 60 mm.

At a charge mass of 7 g, balsa wood is observed to not be completely perforated and, as similarly observed for 40 mm thick balsa wood, the cracking failure on the exit face, in particular, is observed in the general orientation of the grain of the wood. As the charge mass is increased, there is less extensive damage on the exit face, compared to the 40 mm thickness. The average unrecoverable material for the wood at 60 mm thicknessis 20.4%. Divinycell foam exhibits a lower material loss (a mean loss of 3%) at 60 mm thickness after penetration by the fragment; and at the minimum charge mass, no perforation occurs, but the transient deflection of the ‘exit’ face appears to cause the rear plate deflection. The fragment remains wedged in the absorber across all charge masses, but at the highest charge mass, it is loosely wedged in and is removed by the sectioning process. The cross-crack failure observed in the 40 mm thick Divinycell foam is also observed at 60 mm thickness. Polyurethane foam also prevents the fragment from perforation at a charge mass of 7 g, whereas evidence of perforation

61

Fig. 17. Change in rear plate deflection against kinetic energy when polyurethane foam thickness increases to 60 mm.

is observed in the subsequent charge masses. The powdering failure is also noticed for the foam, with a mean material loss of 12.1%. Figs. 15–17 are comparative plots showing how the rear plate deflection against fragment kinetic energy changes from 40 mm thickness to 60 mm thickness for balsa wood, Divinycell foam and polyurethane foam respectively. As shown in Fig. 15, at the lowest kinetic energy regime, 60 mm thick balsa wood provides better damage alleviation than 40 mm balsa. At the highest fragment kinetic energy that the 60 mm thick balsa wood is subjected to about 1500 J, its alleviation of damage is observed to have decreased to be on par with the 40 mm thick energy absorber. A similar trend is observed for Divinycell foam (see Fig. 16) as the additional damage alleviation provided by 60 mm thick foam at the lower kinetic energy regime deteriorates until the 60 mm thick foam provides a similar protection of the rear plate as 40 mm thick foam at a fragment kinetic energy of approximately 1400 J. For polyurethane foam in the comparative plot in Fig. 17, it is evident that

Fig. 18. Comparative plot of rear plate deflection against kinetic energy for the 60 mm thick energy absorbers.

62

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Table 7 Summary of results for tests conducted on 40 mm thick single, double and quadruple-layered polyurethane foam.

Single

Double

Quadruple

Test description

Charge mass (g)

Impulse (Ns)

Fragment speed (m/s)

Fragment kinetic energy (J)

Rear plate deflection (mm)

Material lost from absorber (%)

PU40S-3607-01 PU40S-3608-01 PU40S-3609-01 PU40S-3610-01 PU40S-3611-01 PU40-3607-01 PU40-3608-01 PU40-3609-01 PU40-3610-01 PU40-3611-01 PU40-3611-02 PU40Q-3607-01 PU40Q-3608-01 PU40Q-3609-01 PU40Q-3610-01 PU40Q-3611-01

7 8 9 10 11 7 8 9 10 11 11 7 8 9 10 11

14.42 16.16 17.77 20.35 22.12 13.9 15.8 17.1 20.6 21.3 21.8 14.29 16.02 18.31 19.67 21.44

370.4 408.2 540.5 555.6 625.0 370.4 444.4 487.8 526.3 571.4 606.1 384.6 454.5 425.5 555.6 625.0

556.2 681.4 1152.7 1267.0 1666.0 576.1 870.1 1069.6 1214.7 1312.7 1506.0 576.2 846.1 707.1 1376.5 1644.5

9.38 10.54 15.16 16.17 17.6 9.36 11.78 14.72 16.11 17.15 16.84 10.39 12.44 11.17 16.67 18.2

11.68 14.44 16.90 16.75 16.57 19.3 12.6 19.0 20.0 22.8 21.6 10.26 10.26 12.34 11.15 17.40

increasing the thickness from 40 mm to 60 mm increases damage alleviation. It is also observed that the gradient of the plot of the fit of 60 mm thick foam is slightly higher than for the 40 mm thick foam. The comparative graph in Fig. 18 shows the three materials tested at 60 mm thickness plotted along each other for rear plate deflection against fragment kinetic energy. The rear plate deflects the least when Divinycell foam is providing the damage mitigation, whereas the highest rear plate deflection is observed when protection of the rear plate is provided by balsa wood. The rear plate deflection against fragment kinetic energy from the polyurethane foam appears to be the average of the three materials in terms of damage. 4.4.3. Test with energy absorbers of varying layering arrangements (series 4) The results of the fourth series of tests on 40 mm thick polyurethane investigating the effect of layering – one disc of 40 mm thickness, two 20 mm thick discs and four 10 mm thick discs –

are listed in Table 7. Figs. 19 and 20 show the entry, exit and sectioned views of 40 mm thick foam of varying layering arrangements tested with charge masses of 7 g and 11 respectively. Irrespective of the layering arrangements, perforation by the fragment is observed at all charge masses as the polyurethane foam fails in powdering foam upon impact. The exit opening gets wider with increasing charge masses. For multiple layers (double and quadruple), it appears that the subsequent layer has a wider opening than the previous, particularly as charge mass is increased. The collective plot of rear plate deflection against fragment kinetic energy for the three layering arrangements of 40 mm thick polyurethane foam is shown in Fig. 21. The plot shows the three arrangements as part of one system of axes and as it can be observed from the graph, a single linear least squares fit goes through all the points with a good R2 correlation of 0.95 and all the points fall within an experimental variation of 90%. The layering arrangements do not appear to have any significant effect on the residual damage of the rear plate.

Fig. 19. Entry, exit and sectioned views of 40 mm thick single, double and quadruplelayered polyurethane foam subjected to 7 g charge mass at 36 mm charge diameter.

Fig. 20. Entry, exit and sectioned views of 40 mm thick single, double and quadruplelayered polyurethane foam subjected to 11 g charge mass at 36 mm charge diameter.

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

63

Fig. 21. Plot showing the rear plate deflections against fragment kinetic energy, of the various polyurethane layering arrangements.

4.4.4. Tests with fragments of varying sizes with similar kinetic energies (series 5) The results of the impact of balsa wood, Divinycell foam and polyurethane foam subjected to impact by different sized fragments for similar kinetic energy analysis are listed in Table 8. The graph showing the analysis of the three materials with regard to rear plate deflection against fragment kinetic energy is shown in Fig. 22. Due to the experimental variation, particularly in the speed of the fragment, a wide range of kinetic energies make it difficult to synchronise the kinetic energies of the fragments to one value. For balsa wood, the mean kinetic energy of the three different sized fragments is 489.9 J with a mean percentage variation of 10.2%. The smallest fragment (released by a charge of 27 mm diameter) is observed to be travelling at 488 m/s and imparts a rear plate deflection of 11.1 mm after perforating through the balsa wood; whereas the fragment released from 43 mm diameter charge impact the balsa wood at a speed of 270 m/s and almost perforates through the wood, causing a rear plate deflection of 7.8 mm. For Divinycell foam, the smallest fragment manages to perforate through the foam and causes a rear plate deflection 10.9 mm from an initial speed of 541 m/s just before

impacting the foam. The largest fragment impacts Divinycell foam at 294.1 m/s and imparts a rear plate deflection of less than 2 mm. The mean kinetic energy for the three fragments striking a Divinycellfoam-protected rear plate is 611.4 J with a mean error of 7.5%. For polyurethane foam, which still fails by powdering, the fragment released by a 27 mm diameter charge strikes and perforates the foam from an initial speed of 571 m/s and a rear plate deflection of 10.4 mm is observed on the rear plate. The fragment ejected by a 43 mm diameter charge strikes the foam at a speed of 307 m/s and imparts a rear plate deflection of 7.9 mm, even though full perforation is not observed. The mean kinetic energy for the three fragments impacted against polyurethane foam is 588.6 J with an error of 4.5%. It can thus be inferred from Fig. 22 for all three materials that, if the kinetic energy between different size fragments is maintained constant within a close proximity, the smallest fragment, which travels at the highest speed imparts the highest damage on the rear plate. Conversely, the largest fragment, travelling at the lowest speed of the considered fragments, imparts the lowest damage on the rear plate.

Table 8 Summarised results for the test series where balsa wood, Divinycell foam and polyurethane foam are subjected to impact from different sized fragments (for investigations on close kinetic energies).

Tests used for the parametric analysis are highlighted in grey.

64

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

Fig. 22. Graph showing rear plate deflection under impact from different size fragments with similar kinetic energies.

5. Concluding remarks Characteristics of a fragment generated by localised blast loading of a mild steel plate, how such a fragment damages a parallel mild steel plate and the use of cellular materials to alleviate damage from the fragment have been presented in this paper. The minimum charge mass at which capping on the front is observed for all the charge diameters (27 mm, 36 mm and 43 mm) is 7 g. At a charge mass of 10 g for 27 mm charge diameter and 11 g for 36 mm and 43 mm charge diameters, extensive petalling is observed in addition to capping. The impulse from the blast increases linearly with increasing charge mass and its magnitude is independent of the charge diameter. In the study, the released fragments have a range of release speeds between 244 m/s and 741 m/s. This method of fragmentation by means of capping presents considerable experimental variation. The complex blast loading condition, inconsistent materials properties, and the varying released mechanism of the fragments from the front plate (translation and translation with spinning) can all be attributed as parts of the experimental variation of speed of the fragment. When the rear plate is not protected by an energy absorber, tearing – partial tearing, large tearing or secondary capping – is almost always observed in the rear plate. The fragment kinetic energies beyond which large tearing or secondary capping are observed on the rear plates are 414.8 J, 820.4 J and 904.2 J for the respective charge diameters of 27 mm, 36 mm and 43 mm (in the absence of energy absorbers). In observations of how the different absorbers fail after fragment impact, the entry opening seems to be independent of the charge mass, but the exit opening (during perforation) gets wider as charge mass increases. Aluminium foam generally gets perforated with a compressed foam plug being sheared off the material. The layers of aluminium honeycomb are observed to bend during perforation. Balsa wood loses small chucks of wood particularly on the exit face as charge mass increases and any cracking is observed along the grain of the wood. Corecell foam shows scorch marks, in addition to soot deposit from the blast, during perforation by the fragment. When subjected to penetration or perforation,

Divinycell foam undergoes cross-cracking around the perforation tunnel and propagated obliquely through the thickness of the foam layers. In post-impact analyses, the fragment is sometimes found tightly wedged in the Divinycell foam absorber. Polyurethane foam generally crushes into powder as it is penetrated or perforated through by a fragment. Divinycell foam provides the best damage alleviation while Corecell foam is the most vulnerable to fragment impact. Increasing the thickness of the absorber from 40 mm to 60 mm improves damage alleviation at lower fragment kinetic energies but the improvement generally deteriorates with increasing kinetic energy. Varying the layering arrangement within one thickness of polyurethane foam does not affect the performance and damage alleviation of the foam as the peak stress is almost equal to the plateau stress. When different sized fragments are released within a narrow band of kinetic energy, the smallest fragment (having the highest speed) imparts the highest damage to the rear plate while the largest fragment – which has the lowest speed – imparts the least damage. Acknowledgements The authors would like to acknowledge Peter Slattery of Aerothane Applications for the selfless provision of polyurethane foam for the study. The financial assistance of the University of Cape Town and South Africa’s National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed and conclusions arrived at, are those of the author and are not necessarily to be attributed to the NRF. The authors also thank the staff at the Mechanical Engineering Workshop, University of Cape Town, for manufacturing the test rig. References [1] JIEDDO COIC MID. Global IED monthly summary report, Joint Improvised Explosive Device Defeat Organization, USA, 2012. [2] Austin LJ III, Dunford JF Jr, Improvised explosive devices: contracts for culvert denial systems, Special Inspector General for Afghanistan Reconstruction (SIGAR), Arlington, USA, SIGAR SP-13-8, 2013.

N. Ranwaha, S. Chung Kim Yuen / International Journal of Impact Engineering 92 (2016) 50–65

[3] Kulesz JJ, Wilbeck JS, Cox PA, Westine PS, Baker WE. Manual for the prediction of blast and fragment loadings on structures. Texas, USA: Southwest Research Institute; 1981. [4] Whitney MG, Friesenhahn GJ. Blast and fragment loading on containment structures – a manual overview. Texas, USA: Southwest Research Institute; 1984. p. 1071–88. AD-P004 869. [5] Arnold W, Rottenkolber E. Fragment mass distribution of metal cased explosive charges. Int J Impact Eng 2008;35(12):1393–8. doi:10.1016/j.ijimpeng .2008.07.049. [6] World Academy of Science, Engineering and Technology. International Journal of Mathematical, Computational, Physical, Electrical and Computer Engineering 2011;5(11):1770–5. . [7] Tanapornraweekit G, Kulsirikasem W. Effects of material properties of warhead casing on natural fragmentation performance of high explosive (HE) warhead. World Acad Sci, Eng Technol 2011;5:11–20. [8] Ugrcic M. Numerical simulation of the fragmentation process of high explosive projectiles. Sci Tech Rev 2013;63(2):47–57. [9] Moxnes JF, Prytz AK, Froyland O, Klokkehaug S, Skriudalen S, Friis E, et al. Experimental and numerical study of the fragmentation of expanding warhead casings by using different numerical codes and solution techniques. Def Technol 2014;(0). [10] de Bejar LA, Simmons L, Davis JL. Standoff-mortar fragment velocity characterization before and after perforating conventional building walls. Int J Impact Eng 2008;35(9):1043–52. doi:10.1016/j.ijimpeng.2007 .06.002. [11] Elek P, Jaramaz S. Fragment mass distribution of naturally fragmenting warheads. FME Trans 2009;37:129–35. [12] Teeling-Smith RG, Nurick GN. The deformation and tearing of thin circular plates subjected to impulsive loads. Int J Impact Eng 1991;11(1):77–91. [13] Nurick GN, Shave GC. The deformation and tearing of thin square plates subjected to impulsive loads-an experimental study. Int J Impact Eng 1996;18(1):99–116. [14] Nurick GN, Radford AM. Deformation and tearing of clamped circular plates subjected to localised central blast loads, Recent developments in computational and applied mechanics: a volume in honour of John B. Martin, International Centre for Numerical Methods in Engineering (CIMNE), Barcelona, Spain, pp. 276–301, 1997.

65

[15] Wiehahn MA, Nurick GN, Bowles HC. Some insights into the mechanism of the deformation and tearing of thin plates at high strain rates incorporating temperature dependent material properties, Structures Under Shock and Impact VI (SUSI VI), Computational Mechanics Publications, UK, pp. 207–20, 2000. [16] Jacob N, Chung Kim Yuen S, Nurick GN, Bonorchis D, Desai SA, Tait D. Scaling aspects of quadrangular plates subjected to localised blast loads – experiments and predictions. Int J Impact Eng 2004;30(8–9):1179–208. [17] Nurick GN, Bryant MW. Fragmentation damage as a result of an explosion, Int Symp Plasticity and Impact Mech, New Delhi, India, pp. 484–98, 1996. [18] Balden VH, Nurick GN. Numerical simulation of the post failure motion of steel plates subjected to blast loading. Int J Impact Eng 2005;32(1–4):14–34. [19] Wambua P, Vangrimde B, Lomov S, Verpoest I. The response of natural fibre composites to ballistic impact by fragment simulating projectiles. Compos Struc 2007;77(2):232–40. [20] López-Puente J, Arias A, Zaera R, Navarro C. The effect of the thickness of the adhesive layer on the ballistic limit of ceramic/metal armours. An experimental and numerical study. Int J Impact Eng 2005;32(1–4):321–36. [21] Yungwirth CJ, Wadley HNG, O’Connor JH, Zakraysek AJ, Deshpande VS. Impact response of sandwich plates with a pyramidal lattice core. Int J Impact Eng 2008;35(8):920–36. [22] Hou W, Zhu F, Lu G, Fang DN. Ballistic impact experiments of metallic sandwich panels with aluminium foam core. Int J Impact Eng 2010;37(10):1045–55. [23] Zaretsky E, Asaf Z, Ran E, Aizik F. Impact response of high density flexible polyurethane foam. Int J Impact Eng 2012;39(1):1–7. [24] Atas C, Sevim C. On the impact response of sandwich composites with cores of balsa wood and PVC foam. Compos Struc 2010;93(1):40–8. [25] Hoo Fatt MS, Park KS. Perforation of honeycomb sandwich plates by projectiles. Compos Part A Appl Sci Manuf 2000;31(8):889–99. [26] Li QM, Magkiriadis I, Harrigan JJ. Compressive strain at the onset of densification of cellular solids. J Cell Plast 2006;42(5):371–92. [27] Avalle M, Belingardi G, Montanini R. Characterization of polymeric structural foams under compressive impact loading by means of energy-absorption diagram. Int J Impact Eng 2001;25(5):455–72. [28] Kepler J. Impact penetration of sandwich panels at different velocities – an experimental parameter study: part I – parameters and results. J Sandw Struct Mater 2004;6(4):357–74.