MicroStructPy - Microstructure Mesh Generation in Python
Repository
Documentation
PyPI
MicroStructPy is a microstructure mesh generator written in Python.
Features of MicroStructPy include:
- 2D and 3D microstructures
- Periodic microstructures
- Grain size, shape, orientation, and position control
- Polycrystals, amorphous phases, and voids
- Unstructured and raster meshes
- Mesh verification
- Visualizations
- Output to common file formats
- Customizable workflow

The three steps to creating a microstructure are:
1) seed the domain with particles,
2) create a Voronoi power diagram, and
3) convert the diagram into an unstructured mesh.
Download & Installation
To install MicroStructPy, download it from PyPI using:
pip install microstructpy
If there is an error with the install, try pip install pybind11 first,
then install MicroStructPy.
When upgrading from an earlier version, run pip uninstall pyvoro-mmalahe
first: MicroStructPy now uses pyvoro-rimoli, which installs the same
pyvoro module.
MicroStructPy can also be installed from source:
git clone https://github.com/kip-hart/MicroStructPy.git
pip install -e MicroStructPy/
Installing MicroStructPy creates the command line program microstructpy and
the Python package microstructpy.
The command line program executes a standard workflow on XML input files,
while the package exposes classes and functions for a customized workflow.
Run a Demo
MicroStructPy includes several demo and example files to help new users get
started with the program.
A full list of examples is available online at
https://docs.microstructpy.org/examples.html.
Here is minimal example input file:
<?xml version="1.0" encoding="UTF-8"?>
<input>
<material>
<shape> circle </shape>
<size> 0.15 </size>
</material>
<domain>
<shape> square </shape>
</domain>
</input>
This example can be run from the command line by excuting:
microstructpy --demo=minimal.xml
Alternatively, you can copy the text to a file such as
my_input.xml and run microstructpy my_input.xml.
The same output can be obtained from using the package in a script:
import matplotlib.pyplot as plt
import microstructpy as msp
phase = {'shape': 'circle', 'size': 0.15}
domain = msp.geometry.Square()
# Unpositioned list of seeds
seeds = msp.seeding.SeedList.from_info(phase, domain.area)
# Position seeds in domain
seeds.position(domain)
# Create polygonal mesh
polygon_mesh = msp.meshing.PolyMesh.from_seeds(seeds, domain)
# Create triangular mesh
triangle_mesh = msp.meshing.TriMesh.from_polymesh(polygon_mesh)
# Plot outputs
for output in [seeds, polygon_mesh, triangle_mesh]:
plt.figure()
output.plot(edgecolor='k')
plt.axis('image')
plt.axis([-0.5, 0.5, -0.5, 0.5])
plt.show()
Documentation
MicroStructPy documentation is available online at
https://docs.microstructpy.org.
To build a local copy of the documentation, execute the following from the
top-level directory of the MicroStructPy repository:
pip install tox
tox -e docs
Once built, the documentation will be in docs/build/.
Publications
If you use MicroStructPy in you work, please consider including these citations
in your bibliography:
K. A. Hart and J. J. Rimoli, Generation of statistically representative
microstructures with direct grain geometry control,
Computer Methods in Applied Mechanics and Engineering, 370 (2020), 113242.
(BibTeX)
(DOI)
K. A. Hart and J. J. Rimoli, MicroStructPy: A statistical microstructure mesh
generator in Python, SoftwareX, 12 (2020), 100595.
(BibTeX)
(DOI)
The news article AE Doctoral Student Kenneth A. Hart Presents MicroStructPy to the World,
written by the School of Aerospace Engineering at Georgia Tech,
describes MicroStructPy for a general audience.
Contributing
Contributions to the project are welcome.
Please use the GitHub pull request and issue submission features.
See the Contributing Guidelines for more details.
License and Attributions
MicroStructPy is open source and freely available.
Copyright for MicroStructPy is held by Georgia Tech Research Corporation.
MicroStructPy is a major part of Kenneth (Kip) Hart's doctoral thesis,
advised by Prof. Julian Rimoli.