High Explosives Group Leader: Joe Shepherd

1
High ExplosivesGroup Leader Joe Shepherd

• ASCI Academic Strategic Alliances Program
• Caltech Site Visit
• October 8-9, 1998

2
Overview

• First Principles
• molecular models of explosives, binders and
  reactions
• reactive flow hydrodynamics with detailed
  chemistry and EoS
• Evolutionary
• engineering models of explosives
• high-resolution (AMR) computations with available
  models
• Goals
• Explosives HMX, TATB , Binder Kel-F
• Equation of State, PVET surfaces, G, Cp,
  Hugoniots, isotherms,
• corner-turning problem

3
Simulation Development Roadmap
PROBLEM SOLVING ENVIRONMENT
Continuum Modeling
Mesh Generation
Solid Modeling
Binder and Grain Interactions
Fracture and Fragmentation
Length and Time Scales
Integration and Parallelism
Integration and Parallelism
Equation of State
MicroMechanics
Reaction Rates
Constitutive Relations
Reaction Pathways
Molecular Processes
Phase Transitions
Materials Science
Fluid Mechanics
Computational Science
Solid Mechanics
4
HE Group Participants

• Materials Science and Chemistry (Goddard)
• Siddharth Dasgupta, Gregg Caldwell, Lu Sun, Rick
  Muller,Tahir Cagin
• High Explosives (Shepherd)
• Eric Morano, Chris Eckett, Patrick Hung, Marco
  Arienti, Sairam Sundaram, Ron Fedkiw
• Solid Mechanics (Ortiz)
• Adrian Lew
• Integration CCS
• Dan Meiron, Michael Aivazis

5
Organization

• Weekly meetings (Thursday, 1 pm, Kaplun Library)
• meetings serve as tutorials and progress reviews
• Strong interactions between MS and HE groups

Thermodynamics Continuum Mechanics
Quantum Mechanics Force Fields Molecular Dynamics
Numerical Simulation High Explosive and Shock Data
6
Accomplishments I

• Equation of state of HMX, TATB, Kel-F
• QM, force fields, molecular dynamics by MP group
  (Dasgupta, Caldwell, Sun)
• thermodynamic analysis (PVE surface fits, shock
  hugoniots, Cp, G) by HE group (Arienti, Hung)
• Integration into continuum models by HE
  (Shepherd, Morano)

7
HMX-b
8
TATB
9
Kel-F
10
Thermodynamic Analysis

• Purpose Extract thermodynamic properties from
  molecular dynamics simulation results
• Fit PVE surfaces Birch-Murnaghan or
  Mie-Grüneisen
• Compute
• Grüneisen coefficient
• coefficient of thermal expansion
• Hugoniots (shock adiabats)
• compressibilities

11
Grüneisen coefficient
12
Thermal Expansion Coefficient
Dobratz
13
Isotherms P(V)
14
Hugoniot (Shock Adiabat) P(V)
15
Hugoniot Temperatures
16
Accomplishments II

• Detailed reaction modeling (Eckett)
• reduced chemistry model for H2-O2
• full chemistry model integrated into patch
  integrator
• 1D shock tube simulations of endothermic and
  exothermic reactions

17
Reduced Reaction Models

• Use partial equilibrium and steady-state
  assumptions
• Reduce 19-reaction, 8-species model to 3-step, 5
  species

18
Detailed Chemistry Solutions

• Time-accurate integration of reacting flow
• Adaptive Mesh Refinement (AMR) on spatial
  gradients
• Resolved reaction zone
• Example shock wave in dissociating oxygen

19
Detailed Chemistry Solutions

• Validated with shock tube computations for O2-O
  system

20
Stable Detonation in H2-O2

• P 1 atm, T 300 K, stoichiometric, D 1.2 DCJ

21
Accomplishments III

• Engineering Simulations of HE
• Mie-Grüneisen Equation of State
• calibrated for high explosives (HMX)
• extended to treat expanded states
• patch integrator for simple rate law
• Simulations
• shock tube, inert reactive
• corner turning, pressure-dependent reactions
  (n5)
• interactions with contact surfaces

22
Numerical solution of Euler Equations

• Amrita computational facility (Quirk)
• Two-dimensional, unsteady compressible flow
• Patch-based adaptive mesh refinement
• Finite-volume discretization
• Glaister version of Roe upwind scheme
• minmod limiter
• one-step reaction

23
Inert HMX Shock Tube
24
Detonating HMX Shock Tube
25
HMX Corner Turning I
26
Adaptive Mesh Refinement
27
HMX Corner Turning - finite rate
28
Wave Propagation in Binders

• S. Sundaram, W. Knauss
• PBX is a mixture of HE and binder
• binder is polymer with complex material response
• at low pressure, binder is very compliant
  compared to HE
• polymers become stiffer with increasing pressure
• pressure-dependent viscoelastic constituitive
  model developed
• finite-element simulations of wave propagation in
  rods and plates

29
Validation Verification

• Comparison of HMX TATB EoS with hydrodynamic
  and shock compression data
• original LJ (6-12) nonbonding potentials are too
  repulsive (hard)
• replaced by exp-6, compressibility improved but
  initial density too low
• unit cells too small for melting materials
• Comparison of fluctuation-dissipation results
  with thermodynamic analysis
• either convergence is too slow or methods are
  inappropriate for molecular crystals

30
Validation Verification

• Full vs. reduced chemistry
• reaction zone lengths reproduced over a wide
  range of composition
• intrinsic uncertainty in existing reaction
  mechanisms/rate sets dominates validation process
• Engineering models
• shock tube simulations are important tools
• limited corner-turning data available
• further improvements in reaction/EoS models
  needed prior to serious validation
• Few benchmarks (experimental or analytical)
  available for unsteady detailed chemistry/HE
  simulations

31
Role of Computational Science

• Current simulations use Amrita environment
• Future simulations require
• new environment
• 3D AMR flow solvers
• porting to ASCI platforms
• Tools for interactive exchange of data and
  analysis
• EoS data
• thermodynamic analysis
• numerical simulations of HE

32
Milestones I

• Chemistry
• TATB molecular structure
• level 0 1 force fields
• Molecular dynamics of explosives
• PVT computations for HMX, TATB, KELF
• Reaction mechanisms for CHNO
• Deferred
• Equations of State
• thermodynamic analyses

33
Milestones II

• Reaction Zone Modeling
• reduced reaction model for H2-O2
• implemented detailed model into unsteady solver
• detailed chemistry simulations for H2-O2 in 1D
  started
• 2D simulations deferred
• Engineering models of HE
• Generic issues with MG EoS
• Implement JTF model (partial)
• Implemented MG model for HE detonation
• shock tube and corner turning simulations
• compare hot spot models with grain-binder
  interactions (deferred)

34
Milestones III

• Micromechanics of Granular Material
• pressure-dependent nonlinear viscoelastic
  material
• 1D wave mechanics under shear and compressive
  loads
• 2D simulations in progress
• 2D Prototype of Virtual Facility
• Detonation-shock wave mixing computations
  demonstrated
• Defined interface coupling scheme (ghostfluid
  method)
• Explored CAVEAT and CFDLIB
• AMR still method of choice for detonation/compress
  ible turbulence

35
Future Directions I.

• Atomistic and Molecular studies
• refine force fields treat other binders and
  explosives
• reactive force fields
• initiation of reaction behind shock fronts
• individual molecules
• MD of molecular assemblies
• product equations of state
• Detailed gas-phase chemistry
• 1D unstable simulations with full mechanisms
• 2D simulations with reduced reaction mechanisms
• mechanisms for gaseous nitramine (NM, RDX, HMX)

36
Future Directions II.

• Evolve engineering models
• more realistic models for product EoS (JWLT)
• reaction rate models (JTF)
• Grain-binder interactions
• 2D simulations of binder layers
• Integrated simulation fluid-solid coupling
• flexural waves in shock/detonation tube
• analytic solutions and data available for elastic
  case

37
Interactions

• Pier Tang - XNH, Los Alamos , 3 visits, seminars,
  transfer EoS reaction model code.
• Extended visit by Chris Craddock and Andrew
  McGhee of U. Queensland.
• Ported ASUM-DV flow solver to Amrita.
• JES, S. Dasgupta, Lu Sun, E. Morano attended
  Gordon conference on energetic materials, 2
  posters.
• JES, S. Dasgupta attended 11th Detonation
  Symposium, presented invited paper.
• Recruited two students (Hung, Arienti) and
  postdoc (Fedkiw) to project

38
Progress Toward Integrated Simulation
Prototype of integrated simulation with 2 of 3
elements
39
High Explosives - Summary of Milestones Progress

• 1.Chemistry
• TATB molecular structure level 0 1 force
  fields
• Molecular dynamics of explosives PVT computations
  for HMX, TATB, KELF
• Reaction mechanisms for CHNO
• Equations of State thermodynamic analyses G, CP,
  a, KT
• 2. Reaction Zone Modeling
• reduced reaction model for H2-O2
• implemented detailed model into unsteady solver
• detailed chemistry simulations for H2-O2 in 1D
• 2D simulations
• 3. Engineering models of HE
• Generic issues with MG EoS
• Implement JTF model
• Implemented MG model for HE detonation
• shock tube and corner turning simulations
• compare hot spot models with grain-binder
  interactions
• 4. Micromechanics of Granular Material
• model binder as pressure-dependent linear
  viscoelastic material
• 1D wave mechanics under shear and compressive
  loads

40
2-D Prototype Plan (FY99)

• Goal Demonstrate integrated simulation of
  Eulerian fluid dynamics with deforming boundary
  coupled to Lagrangean solid dynamics.
• Steps
• Define test problem
• Identify simulation components and algorithms
• Write/test proof-of-principle software
• Demonstrate loose coupling computation of test
  problem with simplified geometry (tube) and
  physics (MG HE model, elastic solid)
• Deliverable 2D Protoype of VTF (end of FY99)

41
Test Problem

• Detonation inside of a tube (planar or
  axisymmetric)

Plastic deformation
Elastic deformation
gas
HE
42
Components Algorithms

• HE Compressible Turbulence
• Godunov solver, 2D, fixed mesh or AMR
• engineering models for HE gases
• existing software (Amrita legacy code)
• needs to have boundary algorithm incorporated
• Solid Mechanics
• free Lagrange, continuous rezone FEM
• multiphysics including plasticity, fracture
• existing software (Adlib)
• Boundary coupler algorithm(s)
• level set tracks boundary in Euler code,
  intrinsic in Lagrange method
• ghost fluid method to exchange forces, positions,
  velocity
• time step coordination module

43
Proof-of-Principle

• Single program (no splitting), various levels of
  refinement
• compressible fluid dynamics
• 1D
• 2D
• elastic solid
• lumped mass structural model
• thin-shell model
• 2D elastic wave solution
• 2D FEM model
• boundary tracker
• level set propagation, constrain to SM boundary
• ghostfluid exchange of boundary values

44
Final Steps to Deliverable

• develop proof-of-principle code as modular
  elements
• divorce solid and fluid modules into separate
  programs
• communication socket to exchange boundary
  information
• composition with Python
• time step broker module(s) to coordinate updates
• full-up simulation in serial
• verification against analytic solutions for
  elastic waves
• validation against CIT laboratory experimental
  data for shock detonation waves
• parallel implementation and scaling studies