Experimental and Numerical Investigation of Sandwich Structures Subject to Blast Loading

Sandwich panels have shown improved air blast performance over more traditional monolithic armor however it is an area of continuous research in order to optimize the beneficial shock mitigative properties of the sandwich structure. To that end a series of shock experiments on various sandwich panels via shock tube with high speed photography and numerical analyses via finite element method were performed to determine their efficacy for shock mitigation. Originally corrugated steel core sandwich panels were investigated varying face sheet thickness, corrugation thickness, boundary conditions, and foam infill. The hierarchy of foam infilling within the core was then iterated on and lastly, the corrugated core was replaced with an open cell foam core entrained with various Newtonian and non-Newtonian fluids to determine their behavior under shock loading. These results found that foam infilling had the greatest impact upon shock performance although the benefit decreased with increased face or corrugation thickness (increasing stiffness). When selectively filling the foam hierarchy within the core it was found that back filled (soft/hard) arrangements to be the most effective but using the foam alternately to attenuate the shock wave was not effective. Lastly, the various non-Newtonian fluid fillings were found to have detrimental effect on the performance of the sandwich structures while often being weighty.

In recent years, a number of micro-architectured materials have been developed to use as cores in sandwich panels. These include pyramidal cores [4][5][6], diamond celled lattice cores [7], corrugated cores [8], hexagonal honeycomb cores [9], foam cores [10], and square honeycomb cores [11]. The benefits of sandwich construction depend on core topology. Core designs that afford simultaneous crushing and stretching resistance are preferred. One of the most preferred practical core topologies in blast resistant sandwich panel construction is the corrugated metallic core. These cores provide manufacturing advantages as well as high strength in both the normal and longitudinal directions of the structures [7, 12, & 13].
Sandwich structures have various energy dissipation mechanisms, such as bending and stretching of the face sheet, as well as compression and shear of the core. This is especially pertinent in the case of impulsive loading, wherein the interstices in the metal cellular core can provide adequate space for the large plastic deformation, which is an efficient mechanism to dissipate the energy produced by blast impact [14][15][16][17]. During blast loading, the cellular solid core can absorb more than one half of the initial kinetic energy imparted to face sheet of the sandwich plate. This is due to crushing in the early stages of deformation, prior to significant overall bending and stretching, which causes a reduction in the separation between the face sheets. The high crushing strength and energy absorption per unit mass of the core is therefore important [18][19][20][21][22].
Different material properties have been suggested to provide blast attenuation.
Depending on the acoustic impedance of the interacting medium, the shock wave will reflect, transmit, and/or dissipate to differing degrees [23]. Zhuang et al. [24] examined the scattering effects of stress waves in layered composite materials. Their experimental results show that due to the scattering effects, shock propagation in the layered composites was dramatically slowed, and that shock speed in composites can be lower than that of either of its components. Wakabayashi et al. conducted experiments that suggest that low-density materials may provide the most effective blast mitigation [25]. In recent years, sandwich structures with strong face sheets and lightweight cores have become central structural components for blast mitigation.
Polymeric foams offer unique structural, impact, thermal and acoustic properties, which make them an excellent choice as core materials to obtain low density blast resistive sandwich structures [2,26]. Based on these ideas, extensive research on blast mitigating layered sandwich structures has been performed in recent years, using foam cores with different wave impedances to minimize shock effect [27][28][29].
Studies on metallic sandwich panels subjected to air blasts [8,17] indicate that sandwich plates with high ductility and high energy absorption capacity per unit areal mass show good performance. Liang et al. [30] and Wei et al. [31] studied the behavior of metallic sandwich cores with varying strengths and found that soft cores (those in which the core is much less stiff then the sandwich panels' faces) reduce the momentum transferred, thus providing better mitigation for blast loading. For metallic structures, energy absorption in metallic lattice cores is through large scale plasticity, shear and compressive buckling, and eventual tearing of core walls and face sheets [26].
Another possible application of structural foams is for use as a filler material inside cellular metallic core sandwich structures. It is possible to obtain a new sandwich structure by combining these two cores' shock absorption advantages and decrease the transmitted shock load due to differing acoustic impedances. Moreover, foam filling stabilizes the core cell walls against buckling and increases the strength of the core. Vaziri et al. [21]  Recently, functionally graded foams (where the material properties vary continuously, or stepwise, within the material itself) have gained attention in improving energy absorbing capabilities of sandwich structures. Gardner et al. [27] found that the higher number of layers of increasing impedance improved performance (up to the maximum studied of four). The increasing material interfaces allow for blast wave scattering/dispersion of through interface variations and stepwise compression of the core. Wang et al. [29] studied stepwise graded foam core composite sandwich plates. Three layer cores arranged via acoustic impedance were used in two different configurations: low-mid-high and mid-high-low. In the first configuration, the properties of the layers gradually increased and obtained better performance than the second configuration. Work has also been done with the stepwise increasing of metallic core corrugation thickness, similar to the foams mentioned above but focusing on increasing stiffness rather than impedance. The 5 findings shows that a gradual increase in core stiffness from front to back results in the most effective formation for blast mitigation similar to the foams [39][40][41].
Shear thickening fluids, also known as dilatants, are non-Newtonian fluids that are characterized by a nonlinear increase in viscosity with increasing shear rate. These fluids have a number of applications but recently have been actively and numerously investigated for their potential use as additives to body armor (typically fibrous body armor) as they allow for smaller more flexible armor (as the relatively slow shear rate from human activity results in low viscosity) with greater ballistic protection (when the high shear rate ballistic impact results in high viscosity which dissipates the impact). This behavior and use of shear thickening fluids has been extensively studied by Wagner [42][43][44][45]. The use of these materials in shock mitigation, however, has been sparsely investigated with a review of the literature revealing only two investigations into the topic, each with a limited focus, and reaching opposite conclusions [46,47].

Introduction
A major consideration in the design of military vehicles is their resistance to explosive blast loading. With the fast development of modern military technology, monolithic plates are continuing to fall behind the desired levels of blast protection.
Sandwich structures with cellular solid cores, such as metallic foams and honey-comb structures, have shown superior weight specific stiffness and strength properties compared to their monolithic counterparts in blast resistant structural applications.
Their cellular microstructure allows them to undergo large deformation at nearly constant nominal stress and thus absorb more energy [1][2][3]. To date, the effect of foam filling on blast mitigation of corrugated core sandwich panels under shock loads has not been fully understood. In this study, shock tube experiments and FEM were used to investigate the influence of foam infill on the blast resistivity of corrugated steel core sandwich panels. In addition, monolithic face sheets and foam core sandwich panels were tested and analyzed to validate the FEM. More studies were numerically conducted to investigate the effect of face sheet thickness and corrugated sheet thickness under two different boundary conditions, namely simply supported and Encastre Supported. In order to see the effect of corrugated core rigidity, soft, medium, and hard core cases were studied numerically utilizing both filled and empty conditions under blast loading.
In recent years, a number of micro-architectured materials have been developed to use as cores in sandwich panels. These include pyramidal cores [4][5][6], diamond celled lattice cores [7], corrugated cores [8], hexagonal honeycomb cores [9], foam cores [10], and square honeycomb cores [11]. The benefits of sandwich construction depend on core topology. Core designs that afford simultaneous crushing and stretching resistance are preferred. One of the most preferred practical core topologies in blast resistant sandwich panel construction is the corrugated metallic core. These cores provide manufacturing advantages as well as high strength in both the normal and longitudinal directions of the structures [7,12,13].
Sandwich structures have various energy dissipation mechanisms, such as bending and stretching of the face sheet, as well as compression and shear of the core.
This is especially pertinent in the case of impulsive loading, wherein the interstices in the metal cellular core can provide adequate space for the large plastic deformation, which is an efficient mechanism to dissipate the energy produced by blast impact [14][15][16][17]. During blast loading, the cellular solid core can absorb more than one half of the initial kinetic energy imparted to face sheet of the sandwich plate. This is due to crushing in the early stages of deformation, prior to significant overall bending and stretching, which causes a reduction in the separation between the face sheets. The high crushing strength and energy absorption per unit mass of the core is therefore important [18][19][20][21][22].
Different material properties have been suggested to provide blast attenuation.
Depending on the acoustic impedance of the interacting medium, the shock wave will reflect, transmit, and/or dissipate to differing degrees [23]. Zhuang et al. [24] examined the scattering effects of stress waves in layered composite materials. Their experimental results show that due to the scattering effects, shock propagation in the layered composites was dramatically slowed, and that shock speed in composites can be lower than that of either of its components.
Wakabayashi et al. conducted experiments that suggest that low-density materials may provide the most effective blast mitigation [25]. In recent years, sandwich structures with strong face sheets and lightweight cores have become central structural components for blast mitigation. Polymeric foams offer unique structural, impact, thermal and acoustic properties, which make them an excellent choice as core materials to obtain low density blast resistive sandwich structures [2,26]. Based on these ideas, extensive research on blast mitigating layered sandwich structures has been performed in recent years, using foam cores with different wave impedances to minimize shock effect [27][28][29].
Studies on metallic sandwich panels subjected to air blasts [17,8] indicate that sandwich plates with high ductility and high energy absorption capacity per unit areal mass show good performance. Liang et al. [30] and Wei et al. [31] studied the behavior of metallic sandwich cores with varying strengths and found that soft cores (those in which the core is much less stiff then the sandwich panels' faces) reduce the momentum transferred, thus providing better mitigation for blast loading. For metallic structures, energy absorption in metallic lattice cores is through large scale plasticity, shear and compressive buckling, and eventual tearing of core walls and face sheets [26].
Another possible application of structural foams is for use as a filler material inside cellular metallic core sandwich structures. It is possible to obtain a new sandwich structure by combining these two cores' shock absorption advantages and decrease the transmitted shock load due to differing acoustic impedances. Moreover, foam filling stabilizes the core cell walls against buckling and increases the strength of the core. Vaziri et al. [21] [34] showed that low modulus elastomer infill in pyramidal lattice truss metallic core increased the impact energy absorption capacity. Other studies have had success improving the impact resistance of honeycomb cores by fully or partially filling the cells of the honeycomb [35][36][37][38].
In this study the influence of face sheet thickness, corrugation thickness, boundary condition and foam filling on shock mitigation is explored. Encastre boundary conditions generally decreased panel deflection. The decrease was more prominent with face thickness change than with core thickness change. Generally soft core structures performed better under shock loading than strong or slapping cores with the one exception that completely foam filled panels were the best core having the least back-face deflection. Foam filling reduced the deflection of the panels in all cases although the degree of improvement decreased with the increase in corrugation and face sheet thickness.

Specimen preparation
Corrugated steel core sandwich structures used in this study were produced with low carbon steel face sheets and galvanized, low carbon steel sinusoidal corrugations in a four-layer match-up. A schematic of the sandwich panels is shown in  average mass was 616.2 g, 630.4 g, and 491.9 g for empty corrugated steel core sandwich panels, foam filled corrugated steel core sandwich panels, and foam core sandwich panels, respectively. All three different sandwich panel configurations (see

Shock loading procedure
A shock tube apparatus was used to generate shock waves with planar wave fronts. A photograph of the shock tube used in these studies can be seen in Fig. 3. A typical pressure profile generated by the shock tube and used in these experiments is shown in Fig. 4. The exit muzzle inner diameter of the shock tube was 38.1 mm (see    A high speed photography system was utilized to capture the motion of the specimens in order to determine their deformation and damage propagation. The lens axis of the camera was set perpendicular to the shock tube as shown in Fig. 5. A Photron SA1 high speed digital camera was used at a framing rate of 20,000 frames per second with an image resolution of 512 x 512 pixels over 3 ms duration.

Numerical procedure
Dynamic explicit 3D FEA analyses of the sandwich panels subjected to a blast load were performed using Abaqus/Explicit finite element software. During analysis, nonlinear deformations were accounted for and simulated for duration of 3 ms.

Finite element model
A model was created to render Simple and Encastre Supported (along the back-face's short edges) corrugated steel core sandwich panels 203 x 50.8 x 25.34 mm, subjected to blast loading. Each layer section was modeled as a homogenous sheet with a prescribed thickness. Front and back face sheets were modeled as shells from their back and top faces respectively while the corrugated layers were also modeled as shells defined by their mid planes (Fig. 6a). Interactions between the corrugated core and the front and back-face sheets were taken as surface-to-surface contacts under the penalty contact method and finite tangential sliding. The shear stress limit for failure was prescribed as 20 MPa.
This shock pressure profile was input into Abaqus as tabular data. It was applied to the specimens' front face as a non-uniform function of area as shown in Fig. 6b. This variation in shock pressure induced by the shock tube was observed experimentally and validated numerically by Kumar et al. [40]. For FEM simulations, the specimen was symmetrically aligned with the center of the shock tube, and the distance between the supports was fixed at 152.4 mm.

Material properties
The material properties of the corrugated steel sheets and face sheets as listed in literature were used in this study. These were: modulus of elasticity (E) 205 GPa, Poisson's ratio (ν) 0.29, density (ρ) 7.85 g/cm3, and one-dimensional acoustic wave impedance 3962 x 10 4 kg/m2 s.
A Jonson Cook material model with strain hardening was applied in Abaqus.
The yield stress is, therefore, expressed as where is the yield stress at nonzero strain rate, is the equivalent plastic strain, is the quasi static strain rate, is the equivalent plastic strain rate, and A, B, C, n and m are material constants [41]. is the nondimensional temperature ratio and set to zero in this paper.
Johnson Cook parameters for the material used in this study are given in Table   1 and were obtained from literature [42]. Table 1 Johnson Cook parameters for low carbon steel used in FEM analysis [42].
The foam filled sandwich structures used general purpose humidity cured Polyurethane (PU), of density 0.0446 g/cm3 and elasticity modulus 0.24 MPa. These materials can deform elastically to large strains, up to 90% strain in compression and are intended for finite strain applications The mechanical properties of the PU foam were obtained using quasi static compression tests, and the high strain rate (3000/s) properties were found via Split Hopkinson Pressure Bar experiments [43] and are shown in Fig. 7. Ogden strain energy potential was applied using tabular data from experimentation [41]. In all FE analysis, a Poisson's ratio, m = 0 was used.

Experimental results and discussion
A series of three successful experiments was performed for each geometry to insure repeatability. The applied pressure loading as well as the results obtained were consistent in all the experiments, and are discussed below.

Empty corrugated steel core sandwich specimen response
The side view history of the empty specimen shows different compression behavior over the entirety of the metallic corrugated core (see Fig. 8) during loading.
The beginning of the Back-Face Deflection (BFD) occurred 0.25 ms after the initial Front-Face Deflection (FFD) of the specimen, which implies a coupled response.
Core compression and bending/stretching stages can be clearly observed by using core compression/time curves. In

Foam Core Sandwich Panels
High-speed side view camera images, recording the deflections of foam core sandwich panels, are given in Fig. 10. After the shock wave impinged upon the specimen, the weak foam core did not resist the front face motion enough to decrease its velocity until 1.25 ms. As seen in Fig. 11, the back-face started moving in synchronization with the front face after this time. This behavior is called ''slapping'' [18,19]. The core compresses very rapidly to about 10 mm and then maintains this compression as both faces are moving together. The foam core was subjected to very high (1263/s) strain rate loading in these experiments.

Fig. 10
High-speed images of foam core sandwich specimen during shock loading.

Fig. 11
Experimental results for foam core sandwich panels under shock loading: Front Face Deflection (FFD), Back Face Deflection (BFD), core compression, and strain rate of the core.

Foam Filled Corrugated Steel Core Sandwich Panels
The foam filling caused large changes in specimen behavior when compared to that of the empty and foam core specimens. This behavior can be seen in Fig. 12, which shows high speed side view camera images. Core compression in the foam filled case is decreased when compared to empty and foam core sandwich panels. The BFD starts almost at the same time as the FFD. In Fig. 13a, velocity profiles of both faces are given. The slope of both the front and back-face velocities show the same magnitude, except during two durations between 0.5-1 ms and 2-2.5 ms. The frontface started deflecting initially, but after only 0.25 ms the BFD began. During the first 0.50 ms the first core compression was observed, followed by a brief expansion of the core between 0.50 ms and about 1.00 ms and then a larger, secondary compression lasting until 2.00 ms. Between 1 ms and 2 ms the FFD increased more than the BFD, thus causing an increase in the core compression. After this time, both the BFD and FFD started to decrease, and the core decompressed. This reaction occurred over much less time than the empty corrugated steel core and foam core sandwich panels. In this case, the obtained maximum strain rate in the core was much lower than that of both the empty and foam core cases, and was calculated to be 170/s around 1.25 ms (see Fig. 13b). The core compression percentage at this time was about 5.28%. This means that the foam did not exhibit high strain rate behavior in the foam filled corrugated core sandwich panel under blast loading. It is observed that the foam filling increased bending rigidity and core compression strength.

Fig. 12
High-speed images of fully foam filled corrugated core sandwich specimen during shock loading.

Validation of numerical solutions
To verify the material properties, boundary and contact conditions in the FEM, numerical results were compared to the shock tube experimental results for the face sheet, foam core sandwich panel, empty corrugated steel core sandwich panel, and foam filled corrugated steel core sandwich panels. These comparisons are shown in Fig. 16.
The Pearson's correlation coefficient R 2 was calculated using equations given in [44] as a means to evaluate the model's accuracy. The correlation coefficient is a measure of accuracy of the linear relationship between the experimental and predicted data. The predictability of the finite element model over both the front and back-faces of the four different experimental designs is shown in Table 2.

Fig. 16
Validation of FEM simulations with the experimental results for single face sheet, empty corrugated core sandwich panels, foam core sandwich panels and foam filled sandwich panels.

Investigation of face sheet thickness, corrugated core sheet thickness, and boundary conditions on blast performance
The effect of face sheet thickness, corrugated steel sheet thickness, and boundary condition (see Fig. 17) were investigated using FEM simulations. The values of all the variables are given in Table 3.  Table 4). Thus, unit mass deflections were calculated for comparison to one another. In total, 36 simulations were performed to understand each of the parameter's influences on the behavior of the sandwich panel.

Back-face Deflections (BFDs)
All of the obtained BFDs from FEM simulations can be seen in Fig. 19. The figure shows BFD differences between empty corrugated steel core sandwich panels and foam filled corrugated steel core sandwich panels under shock loading. In

Conclusions
The main objective of this study was to develop a sandwich structure with improved performance under shock loading at room temperatures. To achieve this purpose, metallic corrugated core sandwich panels with polymeric foam filling were developed.
The results obtained from this study are summarized as follows: (1) The corrugations prior to plastic deformation showed elastic buckling and bending. To increase the buckling resistivity and bending rigidity of the corrugations, foam fillings are applied between cells. The foams increased the buckling and bending rigidity of the core, and both experimental and FEM results show that the foam filling generally increased the blast resistivity of the sandwich panels unless the combined factors caused the panels to become too stiff.
(2) For the experimental results front face and back-face deflections were reduced by foam filling by more than 50% around maximum deflection time, while increasing the mass of the panel by only 2.30%.
(3) The experiments showed hard, soft, and slapping core collapses, and of these three the hard core unexpectedly had the least back-face deflection.    Table 3. Investigated parameters in FEM simulations of empty and foam filled corrugated steel core sandwich panels Table 4.Calculated masses of FEM models for the empty corrugated steel core sandwich panels

INTRODUCTION
The search for light weight blast mitigative armor to protect structures from highintesity dynamic loads, created by explosions, has stimulated interest in the mechanical response of metallic core sandwich plates. Sandwich structures were recently studied with different face sheets and cores in an attempt to meet this need.
To date, many metallic core topologies have been developed for use in the sandwich panels [1][2][3][4][5]. One of the most common core topologies in blast resistant sandwich panel construction is the corrugated metallic core which provides manufacturing advantages, and high strength in the normal and longitidinal directions [2,6,7].
Studies on metallic sandwich panels subjected to dynamic air pressure shock (herein refered to as blasts) [3,8], indicate that sandwich plates with high-ductility, and high energy absorption capacity, perform well. Liang et al. [9] and Wei et al. [10] studied the behavior of metallic sandwich cores with varying strengths and found that "soft" cores reduce the momentum transferred, thus providing better mitigation for blast loading. For metallic sandwich structures, energy absorption in metallic lattice cores is through large scale plasticity, shear and compressive buckling, and eventual tearing of core walls and facesheets [11]. In sandwich panels, scattering due to interfaces between dissimilar materials plays an important role in shock wave dissipation, dispersion, and ultimately, mitigation. Depending on the acoustic impedance of the interacting medium, the shock wave will reflect, transmit, and dissipate to different degrees [12]. Zhuang et al. [13] examined the scattering effects of stress waves in layered composite materials. Their experimental results show that the scattering effects dramatically slows shock propagation in the layered composites and can lower the shock speed in composites below that of either of its components.
Wakabayashi et al. [14] experiments' suggest that low-density materials may provide the most effective blast mitigation. In recent years, sandwich structures with strong facesheets and lightweight cores have become central structural components for blast mitigation. Polymeric foams offer unique structural, impact, thermal and acoustic properties, which make them an excellent choice as core materials to obtain low density blast resistive sandwich structures [11,15]. These polymeric foams can be used as filler material inside the interstices of cellular metallic core sandwich structures. It is possible to obtain a new sandwich structure, combining these two cores' shock absorption advantages and decrease the transmitted shock load by their differing acoustic impedences [16][17][18][19][20]. Moreover, foam filling stabilizes the core cell walls against buckling and increases the strength of the core. Foam filling interstices of the metallic core sandwich structures also ensure some multifunctional advantages such as acoustic and thermal insulation [21].
In recent years, functionally graded foams (where the material properties vary continuously, or stepwise, within the material itself) have gained attention in improving energy absorbing capabilities of sandwich structures. Gardner et al. [22] found that the higher the number of layers of increasing impedance improved performance (up to the maximum studied of four). The increasing material interfaces, allows for blast wave scattering/dispersion of through interface variations and stepwise compression of the core. Wang et al. [23] studied stepwise graded foam core composite sandwich plates. Three layer cores arranged via acoustic impedance were used in two different configurations: low-mid-high and mid-low-high. In the first configuration, the properties of the layers gradually increased and obtained better performance than the second configuration.
Work has also been done with the stepwise increasing of metalic core corrugation thickness, similar to the foams mentioned above but focusing on increaseing stiffness rather than impedance. The finding shows that a gradual increase in core stiffness from front to back results in the most effective formation for blast mitigation [24,25].
The study presented focuses on the blast resistance and energy absorption of foam filled corrugated steel core sandwich structures. These structures had various core configurations that were obtained by different filling strategies and then experimentally and numerically subjected to shock wave loading. The experimental and numerical results show that, soft/hard arrangements (front to back) are the most effective structures for blast resistivity.

Specimen Preparation
Corrugated steel core sandwich structures were produced with different foam filling configurations. Low carbon steel face sheets were used with sinusoidal corrugated steel for the core in a four layer match up. The corrugated sheet was galvanized low carbon steel; a diagram of the sandwich panel without filling can be seen in Fig. 1(a) illustrating the five interstitial layers (when a layer is later discussed in this document it is in reference to these layers). The corrugated sheets are 6.35 mm high and 31.75 mm peak to peak.

Each sheet (including face sheets) is bonded by the epoxy adhesive G/Flex
(West System Inc., Bristol, RI), which has a tensile adhesion strength of 20 MPa.
Select interstices layers within the corrugated steel core were filled with low density polyurethane (PU) foam of varied configurations. These layers between steel sheets are coded in order from front-face to back-face one through five. To distinguish between configurations when an empty layer is filled with foam this layer is coded with an F prefix with each layer filled listed in series (example: configuration F1-F5 would be a panel configuration in which the first and last layers are filled as seen by the first configuration in Fig. 1(b)). Fig. 1(b) shows the section of six different experimentally shock loaded sandwich configurations with a diagram of layer labeling. These were tested under blast loading with simply supported boundary conditions.

Shock Load Procedure
A shock tube apparatus was used to load specimens with a planar shock wave. A typical pressure profile generated by the shock tube is shown in Fig. 2. The muzzle inner diameter of the shock tube used was 38.1 mm (see Fig. 3) [26]. Two pressure transducers (PCB102A) were mounted at the end of the muzzle section to record the incident and reflected pressure profiles. The first pressure sensor was mounted 20 mm away from the muzzle, and the second was mounted 180 mm away (160 mm separation from the first pressure sensor).
The incident peak pressure of the shock wave was chosen to be 1.1 MPa and a reflected peak pressure of approximately 5.5 MPa was obtained for the panels (Fig. 2).
The specimens were placed onto a simply supported fixture with a 152.4 mm span. The flat front-face of the specimen was set normal to the axis of the shock tube with the face completely covering the muzzle. A diagram of this set up can be seen in Figure 3. At least three specimens of each type were shock loaded to ensure repeatability.
A high-speed photography system was utilized to determine their deformation and damage propagation by capturing the motion of the specimens. The lens axis of the camera was set perpendicular to the shock tube as shown in Fig. 3. A Photron SA1 high-speed digital camera was used at a frame rate of 20,000 frames per second with an image resolution of 512x512 pixels over 3 ms duration. Dynamic explicit analysis was performed with three-dimensional nonlinear geometry with a 3 ms run time.

Finite Element Model
An explicit finite element model was created to replicate the corrugated steel core sandwich panel's experimental shock loading with identical conditions (geometry, boundary conditions, and loading). The steel plates and corrugated sheets were modeled as homogenous shells with the prescribed thickness. The front and back-face plates were modeled from their back and top faces respectively while the corrugated layers were defined by their mid planes ( Fig. 4(a)). Interactions between Muzzl the corrugated core and the face plates were designated as surface-to-surface contacts under the penalty contact method with finite tangential sliding. The shear stress limit to failure was set to the adhesive limit of 20 MPa. The foam filling is modeled as 3D brick elements rigidly tied to the surrounding shells using the tie constraint, surface to surface discretization method. All elements utilize reduced integration and hourglass control.
An experimentally obtained shock pressure profile of average impulse, discussed below, (Fig. 5(a)) was imported into Abaqus as tabular pressure data. The shock pressure history imparted onto the FE model is a nonlinear function of the pressure profile and area to be loaded to mimic experimental conditions. This shock pressure distribution is described in literature to be a combination of uniform and nonuniform parts [27,28]. A uniform pressure profile is imparted upon the specimen directly in front of the muzzle's inner diameter while the non-uniform shock pressure profile tappers linearly to zero from the edge of the muzzle inner diameter to a range of an additional one-third of the diameter outwards, as shown in Fig. 4(b).
Karagiozova et al [27] suggested the non-uniform distributed shock pressure profile had an exponential decay function, however, this study's variation is assumed to be a linear decay function for simplification.
As the experimental specimens were held by simply supported knife edges, to

Material Properties
The material properties of the corrugated steel sheets and face sheets were taken from literature to be: modulus of elasticity (E) 205 GPa, Poisson's ratio (ν) 0.29, density (ρ)ρ) 7.85 g/cm 3 , and one-dimensional acoustic wave impedance 3,962 x 10 4 kg/m 2 s (Table 1). A Johnson-Cook material model with strain hardening was also used in this FEM for steel to capture strain rate effects ( Table 2). The yield stress is for a Johnson-Cook material model is expressed as [29,30]: where ̅ is the yield stress at nonzero strain rate, ̅ is the equivalent plastic strain, 0 is the quasi static strain rate, ̅ is the equivalent plastic strain rate, A, B, C, n and m are material constants. ̂ is the non dimensional temperature ratio and set to zero in this paper.

Applied Pressure and impulses
Three, 10 mil (0.254 mm) Mylar sheets were used as a rupturing diaphragm to create shock waves. Subramaniam et al. [32] showed that the pressure subjected to a movable surface, like the front-face sheet of the sandwich structure, can be accepted as the same as that applied to a fixed rigid wall found by using measured reflected pressure profiles. Real-time measured pressure profiles from the closest pressure sensor (reflected) can be seen in Fig. 5(a). The average peak value of the reflected shock pressure was 5.18 ± 0.10 MPa. The specimens' masses were measured to be 600.2 ±2.35g and their thicknesses to be 28.85 ± 1.13 mm.
From the reflected pressure profile captured by the transducer closest to the specimen the impulse imparted onto that specimen can be calculated (see Fig. 5(b)).
These pressure profiles can be considered to be the same as the pressure applied to the specimens. Since the cross-sectional area of the muzzle is known the pressure impulse applied on the specimens can be calculated as: where I is the impulse, is the reflected pressure and 0 is the atmospheric pressure. Specific Impulses were calculated for comparison with each other by dividing specimen masses. The impulses were obtained up to 3 ms, with a 2.46 % maximum standard variation. Since the difference in total impulse imparted upon each specimen (regardless of thickness, weight, or filling) was so small, face deflections are considered without normalizations and all numerical simulations were loaded assuming the same average reflected pressure profile.

High Speed Photography and Experimental Results
The real-time observations of the deformation of each type of specimen are shown in Fig. 7. On the right side of each image is the shock tube: the shock wave impinges upon the front-face of the panels from this side.
In all cases, local collapse of the front-face around the center axes occurs from the elastic deflection of the corrugated core soon after the shock wave impinges upon the specimen. As the specimens are loaded over the 3 ms, the shock load is transferred from front to back-face via the corrugated sheet layers. In each layer, the shock load is   Table 2.
Pearson's coefficient equation is given below: where is the simulated transient response and is the experimental transient response.
The coefficient of determination in all cases is 0.96 or higher indicating that the trends of the experimental results and the finite element model are well aligned. A Russell error equal to or less than 0.15 is considered excellent while a Russell error between 0.15 and 0.28 is deemed acceptable, and anything greater than 0.28 is poor [35]. All Russell Error measurements are within the excellent range as seen in Table 2 all FE simulations are accepted as having good agreement with experimental results.

Investigation of filling hierarchy effect by FEM
FE analyses were performed using ABAQUS/Explicit 6.10.1 commercial software. Different filling configurations (front-face, back-face, middle, both-face, and alternate layer filled) were analyzed to observe panel responses to blast loading.
Graphical illustrations of these configurations are given in Fig. 11. Figures 12-16 illustrate the BFD and FFD history of the models separated into these five configurations groups. Fully filled and empty cases are shown in all figures for comparison. Figure 17 shows the boundary conditions used in these numerical  Figure 12 shows the BFD and FFD response of front filled panels by the number of the foam filled layers increasing from front side to back side. By increasing the number of foam filled layers from front to back in the sandwich panels the deflection on the front-face is reduced as the number of layers filled is increased but deflection is increased on the back-face, even more so than the empty case. The increase in BFD is low; the maximum variation in BFD at 3 ms is 2.95%. However, in FFD, the number of foam filled layers decreased the deflection to a maximum of almost 24% at 3 ms. The FFD is observed to be more sensitive than BFD to the number of front filled layers.
The back side filling effect on deflections of front and back-face sheets of the sandwich panels is shown in Figure 13. The results show that increasing the number of the foam filled layers from back to front reduced both the FFD and BFD. At 3 ms, this variation was 17.22% and 33.6% for FFD and BFD, respectively. Each succesive layer filled decreases deflection on each face by a smaller margine. In addition, the four layer filled panel has slightly less BFD than the fully filled panel. Figure 14 shows the deflection history of the center filled panels. Although the number of filled layers increased there is almost no change in BFD (at 3 ms 0.31%) with respect to one another however, both configurations experienced more BFD than the empty case. A 17.9% decrease in FFD occurred as the filled middle layers increased from one to three. Figure 15 shows panels filled near both faces. The most interesting result is that there is no difference between the two cases in BFD until 2.4 ms, after which the more filled panel exhibits slightly less deflection. However, in FFD a 23.9% drop in deflection occurred by the end of 3 ms as the filling layers increased.
The alternately filled cases were investigated for the effect of Filled (F)/Empty (E) layer alteration. The F/E/F/E/F and E/F/E/F/E cases were analyzed (Fig. 16) and the results show that beginning and ending with a filled layer is better than the filling scenario where the first and last layer is empty. At 3 ms, BFD deflection of F/E/F/E/F is less than E/F/E/F/E by about 13.3% and FFD deflection is 22.19% less. With the number of layers studied the influence of filling the layers next to the faces versus the influence of the number of alternate fillings on face deflection cannot be directly separated.

Comparisons
BFD and FFD graphs of the numerical results are rearranged in Fig. 18 based upon the number of foam filled layers rather than their configurations. Three numerical studies of four foam filled layer configurations were performed (front side filled-F1F2F3F4, back side filled-F2F3F4F5, both side filled-F1F2F4F5). All else being equal, this leaves one layer of the five to take the majority of the core collapse.
The back-face fill experienced the least deformation in both FFD and BFD, both sides filled performed the second best, and front-faced filled had the most deflection.
Having the empty layer closest to the shock wave ensures more collapse of that layer, absorbing more energy, and leaving less shock pressure to bend the stiffend back-face causing less plastic strain and collapse overall. In the two layer filling, four-two layer configurations are studied: front side (F1F2) filled, back side filled (F4F5), both sides filled (F1F5), and alternate filled (F2F4). The back side filled case experiences the least deflection again in both faces followed by both sides, alternate fill, and finally front filled for FFD. The same holds true for BFD except that the alternate and front filled experience almost the same BFD as in the three layer filled cases. As less filling is applied to the panels it can be noted that the front and alternately filled panels begin their deflection history with less deflection than both the back filled and both face filled cases up to about 1.5 ms.
Three one layer configurations were also investigated: front side filling (F1), mid side filling (F3), and back side filling (F5). The results show that back side filling results in the least deflection for the front face (FF) and back face (BF) under shock loading followed by mid filled and front filled, respectively, for FFD. Interestingly, following the trend seen in more layer filled scenarios these last two perform the best early on and are reversed in BFD performance where the mid layer filled panel now experiences the most deflection. This is due to a more localized collapse seen in the mid filled layer case versus front filled layer case where the core is more evenly compressed over its length incresing the area of deformation while decreasing the maximum deflection.

Dynamic Collapse Mechanisms
The core filling hierarchy involves three primary shock wave mitigating mechanisms: (1) The scattering of the stress waves due to different interface impedances, (2) the splitting of stress wave such that it transmits to different parts of the structure at different times due to impedance mismatch (or geometry), and (3) mechanical energy absorptions (i.e., plasticity and hysteresis in the foam). The first two mechanisms are for the most part independent of the filling hierarchy (unless the right delays in the wave front from multiple paths eventually coalesce) instead they are dependant upon the number of filled layers and the impedance mismatch in materials (type of foam and metal). Thus, each hierarchy with the same number of filled layers should experience equal benefits from these two mechanisms so any difference will be due to the third, mechanical energy absorption.
The overall pattern of sandwich panel collapse (mechanical energy absorption) is described by Xue and Hutchinson [36,37]

Conclusions
In this study, the effects of preferentially foam filled metallic corrugated core sandwich panels subjected to blast loading were investigated. A series of shock tube experiments and FEM simulations were performed. Front side filling, back side filling, mid filling, alternating filling and both sides filling are compared with one another.
Comparing the different filling patterns to the baseline non-filled corrugated panel, the following observations can be made: 1. The both side-filled cases decrease FFD as more layers are filled and also decrease BFD.
2. The front-filled cases decrease FFD as more layers are filled but increase BFD.
3. The back-filled cases decrease FFD as more layers are filled and also decrease BFD.
4. The middle-filled cases decrease FFD as more layers are filled but increase BFD.

5.
The alternate-filled cases decrease FFD as more layers are filled but behave differently for BFD. The F2F4 fill behaves like the middle-filled cases and increase BFD while the F1F3F5 case behaves like the side-filled cases and decrease BFD.
6. The front-filled, middle-filled, and one alternate-filled (F2F4) cases increases the amount of BFD over the baseline empty case.
7. Comparing the foam cases to one another shows that the back-filled cases are the most effective at decreasing BFD per filled layer.
Additional results obtained in this research can be concluded as follows: 8. Filling hierarchy changes the deformation history and deforming layer sequence. In all cases, empty layers deformed first followed by foam filled layers. Most layer deformation began with bending or buckling of empty layer cell walls followed by plastic deformation of cells and ultimately collapse.
9. Shear force components are more effective in the empty cells. International Post-Doctoral Research Fellowship Programme.

Introduction
The proliferation of terrorist attacks in the past two decades has caused considerable damage to infrastructure, injuries, and loss of life [1,2]. The majority of these attacks is explosive in nature, according to the U.S. Department of State, and demonstrates a need for effective blast mitigation to protect structures that are known to be in harm's way. This need has renewed interest in a type of composite plate: sandwich panels, as a means of protection. Sandwich structures have low areal density and show improved performance against shock over more traditional monolithic plates. This improved protection comes primarily from the behavior of the core as they are optimized to absorb energy and mitigate the transmitted impulse into the infrastructure of concern [3][4][5]. A great deal of work both experimental and numerical has been conducted into characterizing sandwich panels under shock. As well as further refining their efficacy in mitigating blast using a variety of concepts such as cellular solids (metallic foam or polymeric foams are common) [6][7][8], impedance mismatching [9][10][11][12][13], architectured structures [14][15][16][17], and increasing stiffness designs [18].
In addition, the incorporation of non-Newtonian fluids, specifically shear thickening fluids (STFs), also known as dilatants, into body armor has been an area of increasing research. STFs are characterized by a nonlinear increase in viscosity with at a critical shear rate [19][20][21][22]. STF are composed of nano-particles dispersed within a fluid. The behavior of the STF arises from the behavior of the particles in suspension which as shear rate increases often initially causes the fluid to undergo shear thinning as the particles form layers which slip over one another until a critical shear rate value at which point the particles will cluster together, inhibiting flow and increasing viscosity. Two theories have been proposed to explain this phenomenon: orderdisorder theory and hydrodynamic clustering theory which are discussed at length in [23][24][25][26][27][28][29][30][31]. The magnitude of increases in viscosity and critical shear rate when this viscosity jump occurs is determined by a number of factors, primarily by the concentration of the nano-particles within the fluid. Other contributing factors include particle length (anisotropy), particle size, particles stiffness, particle surface energy, and temperature. [32][33][34] These fluids have a number of applications but when used as additives to body armor (typically fibrous body armor) they allow for thinner more flexible armor (as the relatively slow shear rate from human activity results in low viscosity) with greater ballistic protection (when the high shear rate ballistic impact results in high viscosity which dissipates the impact). It has been found that incorporating STFs into ballistic fiber armor increases both the ballistic and stab protection offered by the armor; work pioneered by Norman Wagner. Although significant research has been conducted on both STFs and its use in armor [35][36][37][38][39][40][41] relatively little research is available on its effect on explosive induced air-blast loading referred to in the rest of this article as shock loading. Tan et al [42] found that STF impregnated fibers reduced peak pressure and rate of pressure rise relative to unimpregnated fibers. M.A. Dawson [43] numerical analyzed fluid filled armor for both Newtonian fluids (with high and low viscosity) and STFs for sandwich panel-like constructs concluding that STFs do not improve shock resistance but highly viscous fluids do.
In this study, shock tube experiments and rheological tests were used to investigate the influence of Newtonian and non-Newtonian fluids infused in foam sandwich panels loaded via dynamic air pressure shock.

STF Synthesis
Three STFs reported in literature where created for their potential use in a fluid infused sandwich panel: silica with polyethylene glycol (PEG), calcium carbide with PEG, and corn starch with water.

Silicon and PEG 200
Fumed silica (12 nm diameter Aerosil RD) was mixed with PEG 200 (200 molecular weight) and ethanol in a mixture ratio of 13:18:193 respectively. The mixture was hand stirred to an even consistency then sonicated using a QSonica sonication machine at 20 KHz, 125 W/cm2 at 50% amplitude, 30 s on 10 s off for 5 hrs. The excess ethanol was then removed by heating the solution to 100 C until the ethanol was completely removed resulting in a 58% w/w mixture.

Calcium Carbide and PEG 200
Calcium Carbide was mixed with PEG 200 at a 50% Volume fraction. The mixture was hand stirred to an even consistency then sonicated using a QSonica sonication machine at 20 KHz, 125 W/cm2 at 50% amplitude, 30 s on 10 s off for 5 hrs.

Corn Starch and Water
Corn starch (Argo) and water where hand mixed at a ratio of roughly 53% w/w until even in consistency.

Specimen Preparation
Fluid filled sandwich panels were produced with different fluid infills. Low

Shock Load Procedure
A shock tube apparatus was used to load specimens with a planar shock wave.
A typical pressure profile generated by the shock tube is shown in Fig. 2. The muzzle inner diameter of the shock tube used was 38.1 mm (see Fig. 3) [26]. Two pressure transducers (PCB102A) were mounted at the end of the muzzle section to record the incident and reflected pressure profiles. The first pressure sensor was mounted 20 mm away from the muzzle, and the second was mounted 180 mm away (160 mm separation from the first pressure sensor). The incident peak pressure of the shock wave was chosen to be 1.0 MPa and a reflected peak pressure of approximately 4.4 MPa was obtained for the panels.    water, and corn starch with water (all in appropriate concentrations) are known STFs and were examined for the effect shear rate has on their viscosity in order to determine an optimal candidate for shock testing.

Applied Pressure and Impulses
Two, 10 mil (0.254 mm) Mylar sheets were used as a rupturing diaphragm to create shock waves. Subramaniam et al. [44] showed that the pressure subjected to a movable surface, like the front-face sheet of the sandwich structure, can be accepted as the same as that applied to a fixed rigid wall found by using measured reflected pressure profiles. Real-time measured pressure profiles from the closest pressure sensor (reflected) can be seen in Fig. 5. The average peak value of the reflected shock pressure was 4.42 ± 0.17 MPa.
From the reflected pressure profile captured by the transducer closest to the specimen the impulse imparted onto that specimen can be calculated (Fig. 6). These pressure profiles can be considered to be the same as the pressure applied to the specimens. Since the cross-sectional area of the muzzle is known the pressure impulse applied on the specimens can be calculated as: where I is the impulse, is the reflected pressure and 0 is the atmospheric pressure. The impulses were obtained up to 10 ms (generally reaching their peak at 8.5 ms).  Table 1. The fluids have nearly identical applied impulses once areal density is accounted for with the exception of air in foam which is much higher and the Oobleck which is lower.

Rheologic Results
Rheologic tests were performed on glycerin, silicon oil, PEG 200, water, calcium carbonate with PEG 200 (50% w/w), fumed silica with PEG 200 (58% w/w), water, and corn starch with water (53% w/w). Table 2 below is a synopsis of the density and quasi static viscosity of the pertinent liquids.

Viscoelastic Response of Silicone Oil
The complex moduli (G' and G'', storage and loss moduli respectively) over a range of frequencies and the effect of shear rate on viscosity of Glycerin and Silicon Oil is shown by Fig. 7 and Fig. 8 respectively. In Fig. 7 one can see that with nearly identical elastic (storage) moduli between the silicon oil and glycerin the silicon oil has a much larger loss modulus (energy lost to heat) and is thus considered a viscoelastic fluid. In Fig. 8 we see that both fluids have relatively stable viscosities with respect to shear rate unlike the next three fluids (note the other three fluids use log scales for viscosity while Fig. 8 is linear).

Fumed Silica and PEG 200
Silica in PEG 200 (58% w/w) did not have observable non-Newtonian (shear thickening) responses during hand mixing, even after excess ethanol had been removed. In addition, no shear thickening behavior was found in rheologic testing over a range of shear rates as seen in Fig. 9. The viscosity of this mixture drops steadily from over 11 Pa*s to less than 0.5 Pa*s from 3 /s to 1150/s showing shear thinning properties but no shear thickening. rheologic testing over a range of shear rates as seen in Fig. 10. The viscosity of this mixture drops steadily from over 400 Pa*s to less than 1 Pa*s from 0.001 /s to 1000/s showing shear thinning properties but no shear thickening.

Corn Starch and Water
Corn starch in water as with Calcium Carbide in PEG 200 had observable non-Newtonian responses during hand mixing. This was in part confirmed by rheologic testing over a range of shear rates as seen in Fig. 11. After a plateau of roughly 7,500,000 Pa*s a typical shear thinning regime is observed between the shear rates of 0.00025 /s and 0.15000 /s down to 80,000 Pa*s (all shown in green in Fig. 11 forcing it into a slower shear rate and eventually taking more force to shear than available to the machine.

Fig. 11 Effect of shear rate on the viscosity of corn starch and water
Given the lack of a clear shear thickening response both manually and via testing from the other two STF candidates the corn starch in water was chosen as the STF for further shock experimentation.

High Speed Photography and Experimental Results.
The real-time observations of the deformation of each type of specimen are shown in Fig. 12. On the right side of each image is the shock tube: the shock wave impinges upon the front-face of the panels from this side. The side view images of the five types of panels are shown by Fig. 12, in which they are being exposed to a shock event over 2.5 ms. Note that the silicon enclosing the fluids in the water and oobleck panel cases shears away allowing the fluid within to be violently excreted beginning around 1.0 ms. 10  glycerin having similar deflection curves. The silicon oil undergoes slightly more deflection than the glycerin and has a steeper rise once BFD begins. Lastly, the Oobleck, 15.5mm, deflects slightly more on average than the water, 14.0 mm, (both of which are less than glycerin and silicon oil) and has a BFD rise similar to that of silicon oil while water with the lowest BFD has the lowest BFD slope. Fig. 13 BFD of sandwich panels exposed to shock wave

Time (ms)
Air filled Glycerin Silicon Oil Water Oobleck

Core Compression
As the photographs of the cameras are synchronized the compression of the core can be calculated along the midline of the panels by subtracting the FFD by the BFD as seen in Fig. 13 and Fig. 14. From this the strain and strain rate of the bulk core can be obtained from the following: where, l original is the original thickness of the core and Δl/dt is the rate of deformation. The FFD, BFD, core compression and strain rate of the core for each panel type is shown by Fig. 15 through Fig. 19. The air filed panels show the greatest core compression, 23.7 mm, (roughly 93% at 1.5 ms) with a peak core strain rate of approximately 1000 /s. Over the course of 7.5 ms the core can be seen recovering its original thickness in an oscillating manner. The glycerin filled panel reached a peak core compression of 13.0 mm (51%) at 2.0 ms after shock impact with a peak core strain rate of 800 /s. The silicon oil filled panel reached a peak core compression of 17.2 mm (68%) at 1.9 ms after shock impact with a peak core strain rate of 650 /s. The water filled panel reached a peak core compression of 18.6 mm (73%) at 2.3 ms after shock impact with a peak core strain rate of 750 /s. The oobleck filled panel reached a peak core compression of 18.6 mm (73%) at 2.3 ms after shock impact with a peak core strain rate of 650 /s. Note that both the water and oobleck filled panels do not experience core recovery over the observed period most likely due to lack of restoring pressure from the enclosing silicon. The two liquids' similar core compression is most likely due to the liquids expulsion from the core during the shock load. The elbow in the core compression history of the oobleck filled panels, where the strain rate drops to nearly 0 /s before rising again for an additional millisecond peaking at roughly 400 /s gives some indication that the core may have provided some temporary increased resistance to compression.

Post Mortem Analysis
Two damage patterns are evident after the specimens have been shock loaded.
The first is global bending of the specimen, to various degrees; with possible localized debonding (or tearing) of the silicon barrier with the metallic face sheets at the mid 6. The starch and water (oobleck) filled panels reached maximum BFD sooner than with simply water filled.
8. Glycerin (less viscous) filled panels experienced less FFD and BFD than silicon oil (more viscous) filled panels.
9. Glycerin had the least core compression of the fluids. 10. Viscoelasticity does not seem to play a role in shock mitigation

Chapter 5: Conclusions
In these studies the effects of face plate thickness, corrugation thickness, preferentially foam filled cells, boundary conditions, and fluid infill of sandwich armor to shock load were investigated. Many of the insights found were possible due to an experimentally validated FEM. Foam filling proved a low density solution to increase bending rigidity, buckling resistance, as well as decreasing back face deflection (BFD) and front face deflection (FFD). The minimization of plastic deformation due to the foam also suggests that these panels would perform better for repeated loadings although this was not investigated. Increasing the stiffness of the panel through increasing the face plates, corrugation thickness, or foam infill proved effective as long as the core continued to behave in a soft-core manner. Increasing the stiffness of any of these components reduced the effectiveness of the foam as a shock mitigater however the stiffness of the foam was not varied as an additional parameter so its effectiveness may also scale as its stiffness is matched to that of the surrounding structure.
Preferentialy filling the foam within the corrugation allows for a more nuanced control of the responses of the sandwich panel to shock. Generally, increasing the amount of foam decreases the deflection of the front face of the panels but will increase the deflection of the back face (even over unfilled panels) unless some of the filling is nearest the back face in which the BFD can be greatly reduced. Filling the majority of the back face levels can provide better BFD than even entirely filled specimens although they would have increased FFD as universally the empty layers would compress, shear, and buckle before the foam filled layers would. When filling both sides, the panels behave as thickened face sheeted sandwich panels. So, foam filling can be used as a method to increase face sheet thickness behavior.
While the investigation into filling the sandwich panels with liquids is not yet complete, with the data at hand both the Newtonian and non-Newtonian fillings provide no advantage over previous corrugation and foam filled specimens once their areal density is taken into account. Rheologic testing of shear thickening fluids (STFs) where inconsistent with published results and necessitates further investigation. The two non-Newtonian fluids (silicon oilviscoelastic, and oobleck -STF) performed worse than their Newtonian counterparts although why has not yet been definitively shown. The water and oobleck filled sandwich panels were characterized by extreme failures in bonding of their components likely due to corrosion of the mild steel by water. This in turn changed the panels from an enclosed system whose response would be dominated by the fluids energy and momentum transferal properties to more of an ablative style armor in which the manner in which these fluids and/or core are expended determine the response of the panels. Counter intuitively the specimens which absorbed more impulse, with the exception of the air filled panels, tended to perform better. Additional investigation with improved bonding or non-reactive STFs needs to be conducted for additional insight.