FluidVerse/2D_SABW_OOOO

2D Shock-Induced Air Bubble Collapse in Water with Open BC

Description

The dataset captures the time-evolving behavior of 2D cylindrical air bubbles subjected to an external shock wave in water. The interaction with the shock wave results in a collapse of the air bubble. Here we investigate a scenario with open boundary conditions at every wall.

About the data

Metadata	Description
Solver	ALPACA
PDE	2D compressible Euler equations
Dimension	2D
Number of Trajectories	300
Train-Test-Split	0.8 - 0.2
Number of Timesteps	101
Simulation End Time	15 $\mu s\backslash mu\; s$
Save Interval	0.15 $\mu s\backslash mu\; s$
Fields	- density_mixed - density_water - density_air - pressure - velocityX - velocityY - volume_fraction_water
Simulation Resolution	2048x2560 (Clipped and Downsampled)
Dataset Resolution	512x512 & 256x256
Grid Type	Cartesian Uniform Grid
Initial Condition	The bubbles are in equilibrium with the surrounding pre-shock environment.
Boundary Conditions	North: Open South: Open East: Open West: Open
Conditioning Parameters	Shock Mach Number [1.3 - 2.1] Bubble Count [1 - 5] Bubble Radius [2mm - 3.5mm] Bubble Center Position [Random in Domain]

Further metadata information and channel-wise statistics can be found in the corresponding metadata.json file.

OOD Data

In addition to the in-distribution train/test data, we provide multiple out-of-distribution test datasets:

• OOD Interpolation
  The shock mach number is randomly sampled within the in-distribution data mach number range (1.3 - 2.1).
• OOD Extrapolation SB1-5
  The shock mach number is chosen outside of the in-distribution data mach number range (1.3 - 2.1). The trajectories are evenly split into the two shock mach number values 1.1 and 2.3.
• OOD Extrapolation SB6-7
  The shock mach number is chosen outside of the in-distribution data mach number range (1.3 - 2.1). The trajectoreis are evenly split into the two shock mach number values 1.1 and 2.3. In addition, there are either 6 or 7 bubbles placed inside the domain.

Apart from the mentioned parameters, the trajectory setup is kept the same as the metadata described in section About the data.

Key Challenges for Surrogate Modelling

• Accurate interface capturing
• Peak pressure in the domain
• Understanding wave dynamics
• Tracking bubble fragments

How to Download the Dataset

Download the whole dataset:

hf download FluidVerse/2D_SABW_OOOO --repo-type dataset

Download only part of the dataset with resolution 256x256:

from huggingface_hub import snapshot_download

snapshot_download(
    repo_id="FluidVerse/2D_SABW_OOOO",
    repo_type="dataset",
    allow_patterns="256x256/*",
    local_dir="."
)