Dr. Scott Cook
Tarleton State University
Dept of Math
Implements billiard dynamics of many spherical particles in arbitrary dimension with standard and non-standard particle-wall and particle-particle collision maps using either standard CPU serial mode or accelerated parallel mode using NVIDIA GPU with numba cuda.
Designed for use with Google Colab notebook below, though can be used with other setups.
https://drive.google.com/file/d/1-Sj2OfLQunmWG-I5wPey1QWwMNKK2k9g/view?usp=sharing
Currently supports the standard specular collision and the no-slip collision law introduced by Broomhead and Gutkin and extended by Chris Cox and Renato Feres. See:
- http://www.sciencedirect.com/science/article/pii/016727899390205F?via%3Dihub
- https://arxiv.org/abs/1602.01490
- https://arxiv.org/abs/1612.03355
This simulation should work in any dimension, but has been tested only in 2D and 3D. Note that all particles are assumed to be spherical, but their mass distribution can be controlled via the gamma parameter.
Dependencies
- Python 3 (written and tested with 3.6, but should be backward compatible)
- scientific stack (numpy, matplotlib, etc)
- FFMEG or equivalent (for creating animations)
- Numba cuda for parallel mode on NVIDIA GPU
- Pytables for file i/o
Future developments
- Add random billiard collision laws (both thermally passive and thermally active)
- Improve 3D animations to show spinning body like 2D animations do now.
- Develop interactive widget to change physical parameters and re-run experiment from a GUI