Flow3D

Python wrapper for running and managing simulations in FLOW-3D

Getting Started

1. Installation

pip install Flow3D

2. In the environment with Flow3D create a new Workspace:

flow3d-manage workspace_init example

• This command can also create a huggingface dataset:

  2.1 Create Workspace HuggingFace Dataset

  flow3d-manage workspace_init example upload=True

  Which runs runs the following command to initialize a dataset for the workspace (defaults to HuggingFace)

  python manage.py workspace_dataset_init

3. Navigate to created workspace directory and use package with manage.py file.

cd out/example

Usage for individual simulations

1. Initialize Simulation folders and simulation.yml file.

python manage.py simulation_init example_simulation

2. Edit generated simulation.yml file to your specific configurations

# If using scientific notation, make sure to format like 1.0e+4 or 3.5e-3
# https://stackoverflow.com/a/77563328/10521456

# Global
simulation_finish_time: 0.001    # 0.001 Seconds
velocity: 0.0

# Units here are in (m)
mesh:
  size: 1.0e-5      # 10 µm resolution
  x_start: 0      # 0 µm
  x_end: 1.0e-3     # 1000 µm
  y_start: 0      # 0 µm
  y_end: 1.0e-3     # 1000 µm
  z_start: 0      # 0 µm
  z_end: 3.0e-3     # 3000 µm

fluid_region:
  x_start: 0      # 0 µm
  x_end: 1.0e-3     # 1000 µm
  y_start: 0      # 0 µm
  y_end: 1.0e-3     # 1000 µm
  z_start: 0      # 0 µm
  z_end: 2.8e-3     # 2800 µm

beam:
  diameter: 1.0e-4  # 100 µm (not explicity in prepin file)
  x: 5.0e-4         # Starting Location of Laser Beam 500 µm (0.05 cm)
  y: 5.0e-4         # Starting Location of Laser Beam 500 µm (0.05 cm)
  z: 0.01         # Starting Location of Laser Beam 10,000 µm (1.00 cm)

3. Generate prepin file from simulation.yml

python manage.py simulation_generate_prepin example_simulation

4. Run Simulation (and all other postprocessing and visualization steps)

python manage.py simulation_run example_simulation

• This also runs the following with default arguments:

  4.1. (postprocess = True) Postprocess Simulations

  python manage.py simulation_postprocess example_simulation

• Following arguments only work if postprocess is set to True.

  4.2. (visualize = True)Visualize Simulations

  python manage.py simulation_visualize example_simulation num_proc=4

  4.3. (upload = True) Generate Simulation Dataset

  python manage.py simulation_generate_dataset example_simulation

  4.4. (upload = True) Upload Simulation and source files. (defaults to HuggingFace)

  python manage.py simulation_upload_dataset example_simulation

Usage for batch simulations

1. Create an initialization script (initialize.py) within the workspace directory.

import os

from flow3d import Workspace

# Sets current directory path of `manage.py` to the workspace.
workspace_path = os.path.dirname(__file__)

workspace_filename = os.path.basename(workspace_path)

workspace = Workspace(
    workspace_path=workspace_path,
    filename=workspace_filename,
)

step = 100
start = 100
end = 1000

config = {
    "simulation_finish_time": 0.001,
    "power": 100,
    "velocity": 0.0,
    "temperature_initial": 299.15,
    "evaporation": 0,
    "lens_radius": 5.0e-5,
    "spot_radius": 5.0e-5,
    "mesh": {
        "size": 2.0e-5,
        "x_start": 0,
        "x_end": 5.0e-4,
        "y_start": 0,
        "y_end": 5.0e-4,
        "z_start": 0,
        "z_end": 1.0e-3
    },
    "fluid_region": {
        "x_start": 0,
        "x_end": 5.0e-4,
        "y_start": 0,
        "y_end": 5.0e-4,
        "z_start": 0,
        "z_end": 8.0e-4
    },
    "beam": {
        "diameter": 1.0e-4,
        "x": 2.5e-4,
        "y": 2.5e-4,
        "z": 0.01
    }
}

for power in range(start, end + step, step):
    config["power"] = power
    simulation_folder_string = f"{power}".zfill(4)
    simulation_name = f"{simulation_folder_string}_watts"
    workspace.simulation_initialize(simulation_name, config)

2. Run script to initialize simulations

python initialize.py

3. Build all simulations

python manage.py simulation_build_all

4. Run all simulation

python manage.py simulation_run_all upload=True num_proc=4