PyGran is an open-source toolkit primarily designed for analyzing DEM simulation data. In addition to performing basic and custom post-processing, PyGran enables running DEM simulation with LIGGGHTS in Python. It's recommended to use PyGran with Python 3.7 or higher versions. If you wish to use PyGran's mesh capabilities via VTK, Python<=3.8 is recommended.

The main features of PyGran:

• Interactive DEM simulation and/or analysis using Python
• Parallel "multiple parameter, single script" simulation for parametrization and sensitivity analysis
• Intuitive syntax for particle manipulation and analysis (e.g. slicing, concatenating, etc.)
• Post-processing coupled particle-mesh DEM simulation with VTK
• Quick and easy plotting of DEM data with matplotlib
• Support for high-performance computing with MPI

The core modules in PyGran utilize the following stand-alone packages:

• simulation: provides APIs for running DEM simulation based on the pygran_sim package.
• analysis: provides routines for post-processing DEM data based on the pygran_analysis package.

If you find PyGran useful in your research, please consider citing the following paper:

@article{aam2019pygran,
  title={PyGran: An object-oriented library for DEM simulation and analysis},
  author={Abi-Mansour, Andrew},
  journal={SoftwareX},
  volume={9},
  pages={168--174},
  year={2019},
  publisher={Elsevier},
  doi={10.1016/j.softx.2019.01.016}
}

Installing PyGran is quite straight forward on a Unix/Unix-like machine. Just fire up a terminal and then use pip to install PyGran:

pip install pygran

This will install a basic version of pygran. For a full installation (that includes mpi4py, vtk, etc.), run:

pip install pygran[extra]

PyGran also provides an interface for running DEM simulation with LIGGGHTS. This is achieved by importing the simulation module as shown in the script below for simulating granular flow in a hopper.

from pygran import simulation, params

# Create a DEM object for simulation
sim = simulation.DEM(
    boundary=("f", "f", "f"),
    box=(-1e-3, 1e-3, -1e-3, 1e-3, 0, 4e-3),
    species=(
        {
            "material": params.stearicAcid, 
            "radius": ("constant", 5e-5)},
        ),
    gravity=(9.81, 0, 0, -1),
    mesh={
        "hopper": { # arbitrary mesh name
            "file": "silo.stl", # mesh filename
            "mtype": "mesh/surface", # mesh type
            "material": params.steel, # mesh material
        }
    },
)

# Insert 1000 particles for species 1 (stearic acid)
insert = sim.insert(species=1, value=1000)

# Evolve the system in time
sim.run(nsteps=1e6, dt=1e-7)

Using PyGran for doing post-analysis is also quite straight forward. Computing particle overlaps shown below for instance can be done in few lines of code:

from pygran import analysis

# Instantiate a System class from a dump file
sys = analysis.System(Particles='granular.dump')

# Instantiate a nearest-neighbors class
nns = analysis.Neighbors(Particles=sys.Particles)
overlaps = nns.overlaps

For more examples on using PyGran for running DEM simulation, check out the examples page.

For reporting bugs or suggesting new features/improvements to the code, please open an issue.