Modern armour configurations against 14.5 mm AP Diederen, A.M.1, Broos, J.P.F.1, Peijen, M.C.P.2 1 TNO Prins Maurits Laboratory, P.O. Box 45, NL-2280 AA Rijswijk, The Netherlands 2 Royal Netherlands Army, Directorate of Materiel, P.O. Box 90822, NL-2509 LV The Hague Abstract This paper presents the experimentally established ballistic resistance of armour configurations with modern ballistic materials against Russian 14.5 mm AP-I B32 (steel core) ammunition. The materials involved are ultra hard / high-hardness steel, aluminium, titanium, ceramics and composites. Tests are conducted at both 0º and 60º NATO impact angles. For a number of armour configurations the ballistic limit velocities such as the V99 are established using the Kneubuehl method. These threshold velocities correspond with a certain probability of stopping the ammunition (99% for V99). The different armour configurations can be compared using a single graph per stopping probability (e.g. 99%) which shows the areal density of the tested armour configurations against the minimum shooting distance for which the 14.5 mm AP ammunition will be stopped with the probability indicated. 1. Background The current inventory of military vehicles of The Netherlands as well as of most European countries largely originates from the period of the Cold War, both as to procurement and especially as to development. For peace operations however the threat for passengers and crew has changed dramatically. Vehicles with a well- protected frontal arc or vehicles, which before would operate behind the frontlines, are now exposed to threats from around (including from behind), from above and from beneath. Examples are snipers, mines (a/o. horizontal effect mines) and handheld anti-tank weapons. Moreover, both politics and the public accept own victims during peace operations even less than would be the case when more direct national interests would be at stake. For this reason the protection of passengers and crew of vehicles, especially for the lighter vehicles, often has to be changed drastically. At the same time the accompanying increase in weight has to be limited to prevent the need for adapting parts like the engine, driveline and suspension which would make the vehicle even heavier. At the same time the total spectrum of threats against armoured vehicles is very extensive and diverse. The number of artillery-delivered (canons, rockets and mortars) ‘high performance’ ammunition types alone will increase from 18 to 94 during the next 10 years for the types that will be developed outside the United States [1]. The number of ATGMs (Anti-Tank Guided Missile) will increase from 43 to 68 during the same period [1]. It is clear that the arms race between ammunition and armour is all but finished. In order to limit the weight increase associated with the necessary levels of armour protection there is a strong need for ballistic materials with a high mass Modern armour configurations against 14.5 mm AP Page 2 efficiency (see definition in Appendix A), often leading to the application of lightweight armour materials. It is clear that there is great demand for lightweight armour systems: for upgrading existing vehicles, for add-on armour for both existing and new vehicles and for designing new vehicles, bearing both existing and emerging new threats in mind. 2. Introduction We have chosen to test armour configurations mainly based on aluminium armour representing the hull of a light to medium armoured vehicle. Aluminium has been chosen because it offers sufficient structural stiffness for these types of vehicles and because (depending on the threat and the aluminium type) aluminium armour has a slightly higher mass efficiency than RHA (Rolled Homogeneous Armour). See Appendix A for the definition of mass efficiency. Besides aluminium, also titanium base armour has been used because of its relatively high mass efficiency (1.5 to 1.8 against kinetic energy ammunition according to several sources) combined with its suitability to be used as a hull material with sufficient structural stiffness despite low areal densities. Among the chosen armour configurations, a thin spall-liner has been used together with titanium base armour for two reasons. The first reason is that titanium shows considerably more spalling than aluminium armour, so a spall-liner is called for to protect the vehicle crew. The second reason is to validate and quantify the effect of a thin spall-liner increasing the threshold velocity of an armour configuration against a specific threat, given that the armour is almost balanced against that specific threat. In order to investigate the ballistic protection of as many modern and new materials as possible for the given budget, besides normal impact only one oblique impact angle has been chosen: 60º NATO. If one can only choose one oblique impact angle, this angle is considered to be the most representative for the vertical angle between the shooting line and the frontal (sloped) vehicle armour as well as for oblique impact of the vehicle side armour when hit from the terrain in front of the vehicle. It is always possible to estimate the influence of other impact angles afterwards without the need for additional experiments by performing computer simulations, e.g. with the hydrocode Autodyn. The experimental results already available will hereby act as validation for these computer simulations. For the experiments covered by this paper which constitutes the first series of experiments in a multi-year research project, the Russian 14.5 mm AP-I B32 (steel core) ammunition has been chosen. This represents the upper part of the threat spectrum for kinetic energy projectiles from non-regular opposing forces. This type of ammunition and the weapons to use it are widely spread (40+ countries) and relatively easy to produce. The armour configurations will be compared to one another based on their V99 (99% stopping probability) because this is much closer to real protection requirements than for instance the V50 (50% stopping probability). As will be shown in this paper, the sequence of performance of armour configurations is different when based on the V99 instead of the V50. Lightweight Armour Systems Symposium 1999 Page 3 3. Armour configurations All armour configurations consist of aluminium or titanium base armour with a hard outer layer of other materials. All metallic outer layers are spaced from the aluminium or titanium, the ceramic containing outer layers are adhesively bonded to the (aluminium) base armour. Table 1 gives an overview of the applied armour configurations, which are depicted in Appendix B. Table 1 Armour Impact 1st outer layer airgap 2nd outer layer airgap Base armour Liner config. angle 1 0º SPS-43 yes SPS-43 yes Al 5083 H321 - 2 0º ARMOX-600S yes ARMOX-600S yes Al 5083 H321 - 3 0º ARMOX-600S yes - - Al 5083 H321 - 4 60º SPS-43 yes - - Al 5083 H321 - 5 60º ARMOX-600S yes - - Al 5083 H321 - 6 0º ARMOX-600S yes - - Ti-6Al-4V - 7 0º ARMOX-600S yes - - Ti-6Al-4V Dyneema 8 60º SPS-43 yes - - Ti-6Al-4V - 9 0º LIBA - - - Al 5083 H321 - 10 60º LIBA - - - Al 5083 H321 - 11 0º MARS-300 perf - MARS-300 perf yes Al 5083 H321 - 12 60º MARS-300 perf yes - - Al 5083 H321 - 13 0º DIMOX-AS - - - Al 5083 H321 - The lateral target dimensions are 500 x 500 mm, except for the Dyneema spall-liners that have lateral dimensions of 460 x 460 mm. The (apparent) density ‘ρ’ (see text below) of the armour plates has been determined by measuring the dimensions and weighing the plates. Three types of very hard steel with a Brinell hardness of around 500 HB (SPS-43) and around 600 HB (ARMOX-600S and MARS-300 perforated) have been chosen: − ARMOX-600S, manufactured by Swedish Steel, ρ = 7.84 g/cm3. − SPS-43, manufactured by Special Materials, St. Petersburg, Russia, ρ = 7.63 to 7.70 g/cm3, dependent on plate thickness. − MARS-300 perforated, manufactured by Creusot-Loire of France, ρ = 4.09 to 5.25 g/cm3, dependent on the hole size which varies with the plate thickness. For titanium, the customary alloy Ti-6Al-4V has been chosen. The titanium armour plates for target configurations 6, 7 and 8 (see table 1) have been cut from a large plate of ‘Tikrutan LT31’ with a Brinell hardness of around 300 HB, manufactured by Deutsche Titan of Germany. The measured density is 4.45 g/cm3. LIBA (Light Improved Ballistic Armour) consists of very hard ceramic pellets in a matrix of a polyurethane rubber / resin mixture. It is an Israeli invention [2] and is marketed for Europe by Ten Cate Advanced Composites of The Netherlands. Modern armour configurations against 14.5 mm AP Page 4 The LIBA was clamped directly to the aluminium base armour using screw clamps and quick-acting clamps. The measured density is 3.00 g/cm3 and 2.97 g/cm3 for the normal and oblique LIBA target plates respectively. DIMOX-AS Type 112 is a so-called CMC (Ceramic Matrix Composite) and consists of SiC (siliciumcarbide) particles reinforcing a matrix of Al O (alumina) and an 2 3 interconnected network of aluminium alloy. This CMC combines (part of) the very high hardness of ceramic armour with (part of) the multihit capacity of aluminium armour. Manufacturer is Lanxide Armor Products, Newark, United States of America. The measured density is 3.25 g/cm3. The DIMOX tiles (lateral dimensions of 100 x 100 mm) are attached to the aluminium base armour using the flexible polyurethane adhesive SIKAFLEX 228. To save material cost, each DIMOX tile is laterally confined by the much cheaper Al O tiles 2 3 of similar thickness. The tile pattern is like a brick wall, see figure 1. The Dyneema polyethylene fibre spall-liner, type UD-HB2, is manufactured by DSM High Performance Fibers of The Netherlands. The measured density is 0.92 g/cm3. The Dyneema is mounted on the titanium base armour of configuration 7 by means of 5 bolts (4 at the corners and one at the centre). The aluminium base armour (Al 5083 H321) has a Brinell hardness of around 95 HB and a density of 2.65 g/cm3. Most aluminium armour plates were manufactured by British Aluminium. Apart from the armour configurations mentioned in table 1 also target configurations comprising SiC tiles and Al O tiles on aluminium Al 7039 base armour have been 2 3 tested. However, insufficient experiments have been performed so far to establish their threshold velocities. Both SiC and Al O tiles (lateral dimensions of 100 x 100 mm) are attached to the 2 3 aluminium base armour using the flexible polyurethane adhesive SIKAFLEX 228. Again to save material cost, each SiC tile is laterally confined by Al O tiles of equal 2 3 thickness, see the tile pattern of figure 1. The lateral dimensions of the underlying aluminium base armour (ρ=2.74 g/cm3) are 800 x 300 mm. For SiC ρ=3.12 g/cm3, for Al O (96% alumina content) ρ=3.69 g/cm3. The ceramic tiles were manufactured by 2 3 ETEC of Germany. SiC Al O 2 3 Fig. 1. Tile pattern for SiC tiles, confined by Al O tiles. 2 3 Lightweight Armour Systems Symposium 1999 Page 5 4. Test set-up The targets are mounted on a frame (0º frame or 60º frame) using screw clamps and quick-acting clamps. The armour plates are spaced from one another using square tubular sections at the perimeter of the armour plates. The weapon is placed at a fixed position; the target is shifted before each shot so an undamaged part can be impacted. The distance between weapon and target is 28.7 meters. The impact velocity of the projectile is registered just in front of the target by means of 2 light screens (see figure 2). For a number of experiments high-speed photography recordings have been made using an Imacon 468 camera (see figure 2). Velocity Interferometer Target rack Opto electronic High speed camera Light screens Continuous wave Doppler Radar Light screens Weapon mount Fig. 2. Small-calibre test range at TNO Prins Maurits Laboratory. In order to determine the threshold velocity (V50, V90, and so on) as a function of the impact velocity, a number of shots have to perforate the target and a number of shots have to be stopped by the target. For this reason for a large number of shots the amount of gunpowder in the cartridge has to be diminished to establish a lower impact velocity (corresponding with a larger shooting distance). For a limited number of shots the amount of gunpowder has been slightly increased to establish a sufficient number of perforations. 5. Kneubuehl method The threshold velocities (V50, V90, and so on) are determined according to the Kneubuehl method [3]. This requires a minimum number of 12 shots. The difference between this method and the V50-determination according to STANAG 2920 is that the Kneubuehl method takes the standard deviation into account. By so doing, the threshold velocity is determined as a function of impact velocity instead of determining only one specific threshold velocity (V50: 50% stopping probability). By using the Kneubuehl method, not only the V50 is established but also the sensitivity for decreasing or increasing the impact velocity (shooting distance). This is important, because an armour which is in favour of another armour based only on the V50 (see solid line opposed to dashed line in example of figure 3) can perform worse than the Modern armour configurations against 14.5 mm AP Page 6 other armour at a lower impact velocity (in this example a lower V90 than the other armour). Armour A Armour B Fig. 3. Example of threshold velocities as a function of impact velocity (vertical axis: stopping probability; horizontal axis: impact velocity). Given the V50 and the standard deviation ‘σ’ of an armour configuration, the stopping probability ‘P’ for any impact velocity ‘V’ can be estimated using the t formula [3]: P(V ) =1− 1 (cid:1)Vte−12(V−σV50)2 dV t 2π⋅σ −∞ 6. Results For each armour configuration except no. 13 (see table 1) 12 to 20 shots have been performed (12 is the required minimum for the Kneubuehl method). The impact velocities have been chosen in such a way that both stops and perforations have been realised so that the threshold velocities (see table 2) could be established. The V50 corresponds with an estimated stopping probability of 50%, the V90 corresponds with an estimated stopping probability of 90%, and so on. The areal densities (kg/m2) are given relative to RHA required to stop the threat (14.5 mm AP-I B32). For target configuration 13 only a V50 could be established using the method according to STANAG 2920. Lightweight Armour Systems Symposium 1999 Page 7 Table 2 Armour Impact V50 [m/sec] V90 [m/sec] V99 [m/sec] Standard Areal density config. angle deviation relative to RHA [m/sec] (RHA=100%) 1 0º 899 872 850 21.1 60.0 2 0º 833 782 741 39.8 52.2 3 0º 906 877 853 22.7 58.5 4 60º 788 773 761 11.2 51.1 5 60º 909 889 873 15.5 61.1 6 0º 905 889 875 12.8 50.4 7 0º 994 980 969 10.7 52.0 8 60º 880 856 836 18.7 49.8 9 0º 878 874 872 2.5 44.1 10 60º 933 863 805 55.2 55.5 11 0º 936 924 913 9.7 56.6 12 60º 904 856 816 38.1 55.1 13 0º 851 *) 14 *) 38.1 *) established according to STANAG 2920 Figure 4 is a graphical representation of the results (V50 and V90) for normal impact. It shows the influence of the standard deviation or spread in the results on the relative performance of the armour configurations. 14,5 mm AP-I B32 Normal impact 1000 7 7 950 11 11 3 6 ] 900 1 ec 6 V50 s 9 m/ 3 [ 9 1 V90 V 850 13 2 800 2 750 35 40 45 50 55 60 65 areal density [RHA=100%] Fig.4. V50 and V90 as a function of areal density. Modern armour configurations against 14.5 mm AP Page 8 Figure 5 is another way of representing the results from table 2, using the formula according to Kneubuehl (see chapter 5) and showing that the relative performance of armour configurations is different when based on the V99 instead of the V50. A clear example is the marked difference between the LIBA targets for normal impact (armour configuration 9) and for oblique impact (armour configuration 10). The V50 for armour configuration 10 is higher than for armour configuration 9, but due to the large spread in experimental results (the large standard deviation) its V99 is lower than for armour configuration 9. estimated stopping probability as a function of impact velocity 100 90 %] 80 [ y 70 bilit 60 a 9 ob 50 pr 10 g 40 n pi 30 p o 20 t s 10 0 700 750 800 850 900 950 1000 1050 1100 impact velocity [m/sec] Fig.5. Estimated stopping probability as a function of impact velocity. Figure 6 gives the required areal density (relative to RHA) for the armour configurations impacted at 0º NATO to realise an estimated stopping probability of 99%. Figure 7 gives these results for the armour configurations impacted at 60º NATO. At the right of figures 6 and 7 the shooting distance corresponding with the impact velocity along the vertical axis is given, based on MIL-Std-662E, Issue 94-04. Lightweight Armour Systems Symposium 1999 Page 9 14,5 mm AP-I B32 Normal impact 1000 0 m 7 m] 950 100 m e [ c n a 11 200 m st 900 di ec] ng m/s 9 6 300 m ooti 99 [ 850 3 1 d sh V e 400 m at ul m 800 500 m Si 750 600 m 2 700 35 40 45 50 55 60 65 areal density [RHA=100%] Fig. 6. Required areal densities for an estimated stopping probability of 99%, 0º NATO impact. 14,5 mm AP-I B32 60° NATO 1000 0 m m] 950 100 m e [ c n a 200 m st 900 di ec] 5 ng m/s 300 m ooti 99 [ 850 8 d sh V 12 400 m ate ul 800 10 m 500 m Si 4 750 600 m 700 35 40 45 50 55 60 65 areal density [RHA=100%] Fig. 7. Required areal densities for an estimated stopping probability of 99%, 60º NATO impact. Modern armour configurations against 14.5 mm AP Page 10 Figure 6 shows a decrease in areal density of the armour configurations from very hard steel (1, 2 and 3) and perforated very hard steel (11) via very hard steel in front of titanium base armour (6 and 7) to ceramic pellets in a rubber matrix (9) as outer layer. For armour configuration 13 no V99 has been established, but figure 4 shows that this armour configuration (ceramic matrix composite as outer layer) has a favourable ratio of threshold velocity (V50) and areal density. Based on figure 6 and looking only at the ballistic protection, it can be concluded that armour configurations 11, 7, 6 and 9 are clearly preferable to armour configurations 1 and 3 because of their higher V99s at lower areal densities. Looking at armour configuration 2 compared to 1 and 3 (all very hard steel plus aluminium) one could sketch an estimated trendline from 2 to 1 or 3 showing the effect of increasing the V99 with increasing areal density of these armour configurations of the same kind. So armour configuration 2 is not necessarily better or worse than armour configurations 1 or 3. The same consideration is valid for armour configuration 13 in figure 4. By increasing its areal density the V50 of armour configuration 13 may well match the V50 of armour configuration 9 and still possess a lower areal density. In the ideal situation with sufficient funding and capacity, one would procure additional armour materials and perform additional tests. Alternatively, a trendline showing the increase in V50 or V99 at higher areal densities can be estimated by computer simulations, based on the experimental results already available as a reference. Figure 7 shows the results for oblique impact (60º NATO). Bearing in mind that a certain armour configuration will have a higher V99 when its thickness and areal density is increased, armour configurations 4, 10, 12 and 5 have a ratio between V99 and areal density not too much different from one another. So only a clear distinction can be made between armour configuration 8 (very hard steel plus titanium) with the best ratio of V99 and areal density, and the other armour configurations (4, 10, 12 and 5). For normal impact the areal density ranges from less than 45% (ceramic pellets in a rubber matrix, backed by aluminium) to around 60% (very hard steel plus aluminium) of the areal density of RHA required to stop the threat (14.5 mm AP-I B32). For the target configurations tested at 60º NATO impact angle, the areal density ranges from around 50% (very hard steel plus titanium) to over 60% (very hard steel plus aluminium). SPS-43 and ARMOX-600S (armour configurations 1 to 5) Some of the very hard SPS-43 steel armour plates (around 500 HB) exhibit cracking, both for the 0º NATO and the 60º NATO impact angle experiments. The clamping of these plates at their corners may cause this, but the even harder ARMOX-600S steel armour plates (around 600 HB) did not crack a single time. It should be noted however that the SPS-43 plates were much thinner than the used ARMOX-600S plates.