Sandia National Laboratories
Michael L. Parks
Research
Publications
Presentations
Software
Peridynamics in LAMMPS
Krylov Subspace Recycling
Curriculum Vitae
Home

SOFTWARE

PDLAMMPS (Peridynamics-in-LAMMPS)

Peridynamics is a nonlocal extension of classical continuum mechanics, and is principally used for simulations involving fracture, failure, and fragmentation. Some examples are presented below.

A particular discretization of the peridynamic model has the same computational structure as classical molecular dynamics. I am the principal author of the peridynamic model implemented within Sandia's massively parallel molecular dynamics code, LAMMPS. Credit also goes to Steve Plimpton for helping integrate peridynamics into LAMMPS, and to Pablo Seleson for writing the routine to compute bond damage.

Reports

PDLAMMPS Software

  • PDLAMMPS is distributed as part of the LAMMPS molecular dynamics simulator.
  • Please read the user guide for installation and usage instructions.
  • Please contact me for support issues.

Example Simulations with PDLAMMPS

If you generate any interesting pictures/movies of PDLAMMPS simulations, please e-mail a snapshot and description and I'll post it here.


This is a replication of an experiment described in (Silling, 2005) and also presented in (Parks, 2008). It simulates the impact of a rigid sphere on a homogeneous block of brittle material. The sphere has diameter 0.01 m and velocity of 100 m/s directed normal to the surface of the target. The target material has density 2200 kg/m3 and bulk modulus 14.9 GPa. The target is a cylinder of diameter 7.4 cm and thickness 0.25 cm. It was discretized as as a 3D cubic lattice of particles with lattice constant 0.5 mm, and contains 103,110 particles.

Click on images for movies
Click for movie Click for movie
Top monolayer of brittle target showing fragmentation Cut view of target during impact by projectile

Krylov Subspace Recycling

Many problems in engineering and physics require the solution of a large sequence of linear systems. We can reduce the cost of solving subsequent systems in the sequence by recycling information from previous systems. I develop a family of solvers based upon a technique known as "Krylov Subspace Recycling". Belos currently contains a recycling GMRES solver (GCRODR) and a recycling CG solver (RCG). For some problems, the iteration count required to solve a linear system can be cut by a factor of two.

Report

Recycling GMRES (GCRODR) -- Matlab

  • Download
  • Includes example sequence of linear systems from a finite element fracture mechanics problem constructed by Philippe H. Geubelle and Spandan Maiti.
  • Warning : This is a research code, not a production code! Should you notice any "strange behavior" (abnormal termination, extremely poor convergence, etc.) please contact me.

Recycling GMRES (GCRODR) -- Trilinos

  • GCRODR has been released as part of the Belos package in Trilinos.
  • Please contact me for support issues.

Recycling Conjugate Gradients (RCG) -- Trilinos

  • RCG has been released as part of the Belos package in Trilinos.
  • Please contact me for support issues.

Top of page

Image placeholder

Michael L. Parks


Contact
E-mail: Michael L. Parks
(505)845-0512 (Phone)
(505)845-7442 (Fax)

Mailing address
Sandia National Laboratories
P.O. Box 5800, MS 1320
Albuquerque, NM 87185-1320

UPS, FedEx, etc.
Sandia National Laboratories
1515 Eubank Ave.
Albuquerque, NM 87123