project-computational-science

A simple lattice-Boltzmann implementation for 2D flow simulations with particle tracking

Contents

This LBM code includes:

• D2Q9 lattice
• BGK collision
• Bounce back
• Aerosol tracking
• Infection simulation

Run simulation

Note: Some simulations can take a while to run. Also be warned that when using many iterations it may use a decent amount of disk space. About 0.1MB per iteration.

Using experiment.py in the repository we can simulate the following experiments:

• Lid driven cavity cavity
• Karman vortex karman
• Validation using lid driven cavity validation
• Own experiment, with our own map created in maps/concept.txt using mapeditor.py. Setting up behaviour for the inlets and outlets in the function "own_experiment" in experiment.py will result in different simulations.
• ...

Specify in experiment.py the initial conditions by changing model_params. After the simulation is complete a folder /frames folder along with a .html can be used to view the 2d animation.

Benchmarks

Lid-driven cavity

A simple lid driven cavity test in a 100x100 unit square. Below are the computed velocity norms at Reynolds number = 100, after 10000 iterations.

A comparison of uy = f(y) and vx = f(x) at the center of the domain with reference data from "U. Ghia, K. N. Ghia, C. T. Shin, High-Re solutions for incompressible flow using Navier-Stokes equations and multigrid method" (placed in validation/cavity_vx_ref.txt). MSE errors for vx and uy are given such that comparison between different LBM can be made.

Examples

An animation of the lid driven cavity (python3 experiments.py cavity) :

And one for a Karman vortex (python3 experiments.py karman):

Notes

Map editor

Using the map editor we can create our own environment with the following map attributes:

• AIR -- Acts as a medium for the fluid.
• WALL -- Causes fluid to bounce back.
• INLET -- Creates a rightward flow of fluid. Velocity is set according to u_lb.
• OUTLET -- Causes fluid to 'exit' the domain.
• INFECTED -- Infected person where particles spawn.
• SUSCEPTIBLE -- Susceptible person can get infected with particles. (don't become INFECTED in simulation)

Authors

• Jeroen van den Berg (jeroenvdb@live.nl)
• Erencan Tatar (erencantatar@gmail.com)
• Robbie Koevoets (robbie.koevoets@live.nl)

Dependencies

To install the required dependencies, simply run pip install -r requirements.txt

matplotlib==3.3.2
numpy==1.19.2
opencv_contrib_python==4.5.1.48
pandas==1.1.3
scipy==1.5.2