A Facility for Simulating the Dynamic Response of Materials - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

A Facility for Simulating the Dynamic Response of Materials

Description:

A Facility for Simulating the Dynamic Response of Materials. Steering ... variety of advection schemes. Eulerian formulation. Parallel Implementation. MPI based ... – PowerPoint PPT presentation

Number of Views:37
Avg rating:3.0/5.0
Slides: 22
Provided by: danme
Category:

less

Transcript and Presenter's Notes

Title: A Facility for Simulating the Dynamic Response of Materials


1
A Facility for Simulating the Dynamic Response of
Materials
  • Steering Committee Meeting
  • Wednesday, April 12, 2000

2
VTF Software architecture
Python scripting layer
Materials properties
Solid Mechanics
Solid-fluid coupling
CFD solvers
Adaptive Mesh Refinement
Parallel adaptive meshing
MPI
Scalable I/O
Nexus/Globus
3
Overview of virtual facility (VTF)
  • Python scripting interface drives all simulations
  • Computational Engines
  • CFD (Eulerian)
  • High explosives
  • Compressible turbulence
  • Solid mechanics (Lagrangian)
  • Fluid-solid coupling
  • Materials properties database
  • Facilities for high performance computing
  • Facilities for visualization, I/O etc.

4
Status of VTF as of 10/99
  • Integrated Simulations
  • coupled simulation of HE fluid dynamics with
    Lagrangian solid dynamics
  • serial CFD
  • serial solid dynamics
  • achieved via 2-D research code
  • coupled simulation of strongly compressible
    perfect gas CFD with Lagrangian solid dynamics
  • parallel CFD
  • serial solid dynamics
  • VTF software architecture implemented
  • MPI aware Python interpreter

5
CFD solver
  • Solves Euler equations
  • perfect gas
  • 2-D, axisymmetric or 3-D
  • variety of advection schemes
  • Eulerian formulation
  • Parallel Implementation
  • MPI based
  • Implemented on variety of platforms
  • Cray T3E
  • ASCI Blue Mountain
  • ASCI Blue Pacific
  • Linux Beowulf

6
Results of site review 10/5/99
  • Overall the center is making excellent progress
  • research is well focused
  • appropriate levels of technical depth and breadth
    by researchers
  • good interdisciplinary collaborations
  • research goals and milestones are in general
    being met
  • interactions with DP lab personnel are positive
  • Recommendations
  • roadmaps should contain risk assessment
  • provide clear vision of software architecture
  • Computational science still an issue
  • at present will not satisfy ASCI goals

7
What we will have by years end
  • Integrated Simulations
  • 3-D fully coupled simulations of HE detonation
    impinging on solid targets
  • parallel 3-D Lagrangian solid mechanics
  • parallel multi species CFD with adaptive mesh
    refinement
  • fully functional problem solving environment
  • materials properties database
  • Supported code base

8
VTF capability combinatorics
eulerian
lagrangian
coupled
2D
axisymmetric
3D
serial
parallel
no AMR
AMR
stand alone
Pyre
Globus
Visualization
SIO
UMP
Constitutive models
Geometrical configurations
9
VTF capabilities at the end of FY1999
eulerian
lagrangian
coupled
2D
axisymmetric
3D
serial
parallel
no AMR
AMR
stand alone
Pyre
Globus
Visualization
SIO
UMP
Constitutive models
Geometrical configurations
10
Status at the half-way mark for FY 00
  • CFD solver
  • integration of AMR in 2-D
  • integration of HE engineering model in parallel
  • demonstration of coupling in 3-d with spring
    solver
  • integration with CPT algorithm in 3-D
  • integration with adlib (in progress)
  • Solid dynamics solver
  • parallel meshing
  • parallel mechanics
  • parallel contact (in progress)
  • interface with 3-D CPT algorithm
  • integration with RM3D (in progress)

11
What is left to complete?
  • Full parallel coupling of adlib and RM3d
  • adlib and CFD run in separate parallel process
    spaces
  • solid boundary broadcast to all processes
  • parallel CPT algorithm used to compute level set
  • Integration into Pyre
  • We are a few weeks away from achieving this
  • Computational science goals
  • Integration with Globus
  • Visualization
  • Scalable I/O
  • Scalability on ASCI platforms

12
The deliverable for FY2000
eulerian
lagrangian
coupled
2D
axisymmetric
3D
serial
parallel
no AMR
AMR
stand alone
Pyre
Globus
Visualization
SIO
UMP
Constitutive models
Geometrical configurations
13
Implementation plan milestones for FY2000
14
Revised milestones for FY2000
15
Milestone status
Pyre
2D
coupled
engineering HE model
parallel
1Q00
  • Engineering model incorporated into CFD code
  • adlib has been made parallel
  • Milestone redefined and pushed to 3rd quarter
  • Folded into 3rd quarter milestone and extended
    to 3-D

16
Milestone status
Pyre
3D
lagrangian
engineering HE model
parallel
Ta
2Q00
  • All simulation capabilities in place
  • Awaiting completion of full adlib scripting into
    Pyre
  • Expect completion by middle of Q3 00

17
Milestone status for FY2000
noPyre
3D
eulerian
engineering HE model
parallel
AMR
3Q00
  • GrACE used to enable AMR capability in 2-d CFD
    solver
  • HE model solver
  • Perfect gas
  • Scalability studies are in progress (S. Brunett)
  • Milestone is on schedule

18
Milestone status for FY2000
  • Awaiting full scripting capability for adlib
  • Awaiting completion of interface to uniform
    materials properties database

19
Milestone status for FY2000
  • Requires interaction with CS group to finalize
    interfaces for SIO and VIZ
  • Expect completion on schedule

20
Milestone status for FY2000
  • Requires interaction with CS group to finalize
    interfaces for SIO and VIZ
  • Expect completion on schedule

21
Milestone status for FY2000
  • LES capability has been formulated
  • ENO/WENO solver completed
  • Milestone is on schedule
Write a Comment
User Comments (0)
About PowerShow.com