John E. Reaugh

Defect evolution and pore collapse in crystalline energetic materials

This work examines the use of crystal based continuum mechanics in the context of dynamic loading. In particular, we examine model forms and simulations which are relevant to pore collapse in crystalline energetic materials. Strain localization and the associated generation of heat are important for the initiation of chemical reactions in this context. The crystal mechanics based model serves as a convenient testbed for the interactions among wave motion, slip kinetics, defect generation kinetics and physical length scale. After calibration to available molecular dynamics and single crystal gas gun data for HMX (octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine), the model is used to predict behaviors for the collapse of pores under various conditions. Implications for experimental observations are discussed.

Direct numerical simulation of shear localization and decomposition reactions in shock-loaded HMX crystal

A numerical model is developed to study the shock wave ignition of octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX) crystal. The model accounts for the coupling between crystal thermal/mechanical responses and chemical reactions that are driven by the temperature field. This allows for the direct numerical simulation of decomposition reactions in the hot spots formed by mechanical loading. The model is used to simulate intragranular pore collapse under shock wave loading. In a reference case: (i) shear-enabled micro-jetting is responsible for a modest extent of reaction in the pore collapse region, and (ii) shear banding is found to be an important mode of localization. The shear bands, which are filled with molten HMX, grow out of the pore collapse region and serve as potential ignition sites. The model predictions of shear banding and reactivity are found to be quite sensitive to the respective flow strengths of the solid and liquid phases. In this regard, it is shown that reasonable assumptions o...

Impact studies of five ceramic materials and pyrex

Summary We measured the ballistic performance of five ceramic materials (alumina, silicon carbide, boron carbide, aluminum nitride, and titanium diboride) and Pyrex, when they are backed by thick steel plates. The projectile for all tests was a right-circular cylinder of tungsten sinter-alloy W2 with length 25.4 mm and diameter 6.35 mm, fired at velocities from 1.35 to 2.65 km/s. For this threat we determined the minimum areal density of each material that is needed to keep the projectile from penetrating the backup steel. For all of the facing materials studied here, this performance measure increases approximately linearly with projectile velocity. However, the rate of increase is significantly lower for aluminum nitride than for the other materials studied. Indeed, aluminum nitride is a poor performer at the lowest velocity tested, but is clearly the best at the highest velocity. Our computer simulations show the significant influence of the backing material on ceramic performance, manifested by a transition region extending two projectile diameters upstream from the material interface. Experiments with multiple material layers show that this influence also manifests itself through a significant dependence of ballistic performance on the ordering of the material layers.

Computer simulations to study the explosive consolidation of powders into rods

In the direct method of explosive consolidation, metal or ceramic powder to be consolidated is placed in a metal tube, which in turn is surrounded by a concentric cylinder of explosive. The explosive is detonated at one end so that the detonation front moves axially along the powder‐filled tube. The desired result is a crack‐free, uniformly consolidated rod with the density of the starting material. At present, the wall thickness of the hollow cylinder of explosive that best produces the desired result is determined by experiment. We present the results of computer simulations revealing details of this dynamic process that are difficult to obtain experimentally. The simulations show that the best thickness depends on the equations of state of both explosive and powder. Such simulations can guide changes to the experimental geometry, even when approximate equations of state are used, although the best explosive and its thickness must ultimately be determined by experiment.

Mesoscale modeling of deflagration-induced deconsolidation in polymer-bonded explosives

Initially undamaged polymer-bonded explosives can transition from conductive burning to more violent convective burning via rapid deconsolidation at higher pressures. The pressure-dependent infiltration of cracks and pores, i.e., damage, by product gases at the burn-front is a key step in the transition to convective burning. However, the relative influence of pre-existing damage and the evolution of deflagration-induced damage during the transition to convective burning is not well understood. The objective of this study is to investigate the role of microstructure and initial pressurization on deconsolidation. We performed simulations using the multi-physics hydrocode, ALE3D. HMX-Viton A served as our model explosive. A Prout-Tompkins chemical kinetic model, Vielles Law pressure-dependent burning, Gruneisen equation-of-state, and simplified strength model were used for the HMX. The propensity for deconsolidation increased with increasing defect size and decreasing initial pressurization, as measured by the increase in burning surface area. These studies are important because they enable the development of continuum-scale damage models and the design of inherently safer explosives.

Computer Simulations to Study the High‐Pressure Deflagration of HMX

The accepted micro‐mechanical picture of the build‐up of detonation in solid explosives from a shock is that imperfections are a source of hot spots. The hot spots ignite and link up in the reaction zone by high‐pressure deflagration. Although the deflagration is subsonic, there are so many ignition sites that the pressure build‐up is rapid enough to strengthen the initial shock. Quantitative advances in this research require a detailed understanding of deflagration at the high pressure, 1 to 50 GPa, which is present in the reaction zone. We performed direct numerical simulations of high‐pressure deflagrations using a simplified global (3‐reaction) chemical kinetics scheme. We used ALE‐3D to calculate coupled chemical reactions, heat transfer, and hydrodynamic flow for finite‐difference zones comprising a mixture of reactants and products at pressure and temperature equilibrium. The speed of isobaric deflagrations depends on the pressure and initial temperature. We show how this dependence changes with ki...

The role of viscosity in TATB hot spot ignition

The role of dissipative effects, such as viscosity, in the ignition of high explosive pores is investigated using a coupled chemical, thermal, and hydrodynamic model. Chemical reactions are tracked with the Cheetah thermochemical code coupled to the ALE3D hydrodynamic code. We perform molecular dynamics simulations to determine the viscosity of liquid TATB. We also analyze shock wave experiments to obtain an estimate for the shock viscosity of TATB. Using the lower bound liquid-like viscosities, we find that the pore collapse is hydrodynamic in nature. Using the upper bound viscosity from shock wave experiments, we find that the pore collapse is closest to the viscous limit.

First Results of Reaction Propagation Rates in HMX at High Pressure

We have measured the reaction propagation rate (RPR) in octahydro‐1,3,5,7‐tetranitro‐1,3,5,7‐tetrazocine (HMX) powder in a diamond anvil cell over the pressure range 0.7–35 GPa. In order to have a cross‐comparison of our experiments, we conducted RPR experiments on nitromethane (NM) up to 15 GPa. Our results on NM are indistinguishable from previous measurements of Rice and Foltz. In comparison to high‐pressure NM, the burn rates in solid HMX are 5–10 times faster at pressures above 10 GPa. Numerical simulations of the burn rate of pressurized HMX were also performed for comparison to the results obtained. The simulated burn rates closely approximate the observed rates at pressures up to 3 GPa. However, further refinement to the computational model is required for the calculated burn rates to approach those observed at higher pressures.

Modeling violent reaction following low speed impact on confined explosives

To ensure the safe storage and deployment of explosives it is important to understand the mechanisms that give rise to ignition and reaction growth in low speed impacts. The High Explosive Response to Mechanical Stimulus (HERMES) material model, integrated in the Lagrangian code LSDYNA, has been developed to model the progress of the reaction after such an impact. The low speed impact characteristics of an HMX based formulation have been examined using the AWE Steven Test. Axisymmetric simulations of an HMX explosive in the AWE Steven Test have been performed. A sensitivity study included the influence of friction, mesh resolution, and confinement. By comparing the experimental and calculated results, key model parameters which determine the explosives response in this configuration have been identified. The model qualitatively predicts the point of ignition within the vehicle. Future refinements are discussed.

Computational studies of the skid test: Evaluation of the non-shock ignition of LX-10 using HERMES

We perform computational studies to evaluate the non-shock ignition response of LX-10 (95wt.% HMX, 5wt.% Viton A) during a skid impact test. We employ the HERMES (High Explosive Response to MEchanical Stimuli) model for LX-10 for Skid test calculations investigating the influence of drop height and angle on the pressure, strain-rate, strain, and ignition states. While grit is typically present in skid tests, it was not considered in these continuum-scale calculations. We found that the incident angle has a much more significant influence on pressure, strain-rate, strain, and ignition states than drop height. The peak HERMES ignition parameter value, Ign, is nearly one order of magnitude higher for an incident angle of 45° than for 14°. Peak Ign values occur at the contact patch where shear deformation is highest and is a result of the shear-dependence in the ignition criteria. While peak Ign values for Steven Tests were approximately 60, the skid test had a much smaller value < 2 for the scenarios considered in this study. The discrepancy in ignition values suggest that grit-explosive interactions play a significant role in skid test response. Since the peak ignition values are much less for the 14° impact angles, the role of the grit may be more important at lower incident angles. Future work should include meso-scale calculations to resolve the localized grit interactions that underpins these shear ignition mechanisms.