DFSPH-2D

A Python implementation of a 2D divergence-free SPH (DFSPH) solver. This solver is based on the method described by Koschier and Bender (2017) for fast incompressible flow simulation.

Warning

The gif above feels like it's a real time computation. Unfortunately it's an acceleration of the capture of a simulation, which takes likes 2 minutes to be computed.

Installation

Ensure you have Python 3.8+ installed. Then, install the package in editable mode using the provided pyproject.toml:

pip install -e .

Usage

Run the simulation via the main entry point:

python main.py [options]

You can customize parameters using the following options:

• -r, --support_radius
  SPH support radius (default: 0.2742/2).

• -dt, --timestep
  Time step for the simulation (default: 0.01).

• -s, --steps
  Number of simulation steps (default: 1000). (Used in headless mode.)

Box & Grid Parameters

• --box_origin
  Origin of the box for particle initialization (default: [-0.8, -0.8]).

• --box_size
  Size of the box for particle initialization (default: [1.4, 1.4]).

• --grid_size
  Grid dimensions as (width, height) (default: [2, 2]).

• --grid_origin
  Position of the grid in simulation space (default: [-1, -1]).

Physical Parameters

• --rest_density
  Rest density of the fluid (default: 1027.0).

Visualization

• -v, --visualize
  Enable real-time visualization (default: enabled).

Export/Import

• -e, --export_results
  File name to export particle data (default: empty).

• -i, --import_results
  File name to import particle data for visualization (default: empty).

• -ii, --import_init
  File name to import initial particle configuration (default: empty).

Reference

Koschier, M., & Bender, J. (2017). DFSPH: Fast Divergence-Free SPH for Incompressible Flows. Link to paper