README.md

metadata

dataset_info:
  - config_name: geometry
    features:
      - name: node_coordinates_x
        list: float64
      - name: node_coordinates_y
        list: float64
    splits:
      - name: default
        num_bytes: 80664
        num_examples: 1
    download_size: 3772
    dataset_size: 80664
  - config_name: parameters
    features:
      - name: velocity
        dtype: float64
    splits:
      - name: default
        num_bytes: 2400
        num_examples: 300
    download_size: 3632
    dataset_size: 2400
  - config_name: snapshots
    features:
      - name: velocity_magnitude
        list: float64
      - name: pressure
        list: float64
    splits:
      - name: default
        num_bytes: 24199200
        num_examples: 300
    download_size: 18171167
    dataset_size: 24199200
configs:
  - config_name: geometry
    data_files:
      - split: default
        path: geometry/default-*
  - config_name: parameters
    data_files:
      - split: default
        path: parameters/default-*
  - config_name: snapshots
    data_files:
      - split: default
        path: snapshots/default-*

Lid-Driven Cavity Flow Dataset

Dataset Description

This dataset contains computational fluid dynamics simulations of the classic lid-driven cavity problem with varying lid velocity parameters.

Dataset Summary

The Lid-Driven Cavity dataset provides numerical simulations of viscous fluid flow in a square cavity with a moving lid. The dataset includes velocity magnitude and pressure fields, representing a fundamental benchmark problem in computational fluid dynamics suitable for validation of numerical methods and machine learning applications.

Dataset Structure

Data Instances

The dataset consists of three configurations:

• geometry: Mesh node coordinates
• snapshots: Flow field solutions (velocity magnitude and pressure)
• parameters: Lid velocity values for each simulation

Data Fields

Geometry Configuration

• node_coordinates_x: Sequence of x-coordinates of mesh nodes (float64)
• node_coordinates_y: Sequence of y-coordinates of mesh nodes (float64)

Snapshots Configuration

• velocity_magnitude: Velocity magnitude at each node (float64)
• pressure: Pressure field at each node (float64)

Parameters Configuration

• velocity: Lid velocity parameter for each simulation (float64)

Data Splits

• default: Contains all simulations

Dataset Creation

Source Data

The dataset was generated using numerical simulations of the incompressible Navier-Stokes equations in a square cavity with a moving top lid. This is a classic benchmark problem used for validating CFD codes.

Preprocessing

Solutions are stored as 1D arrays corresponding to the nodal values. The geometry uses a structured or unstructured mesh representation of the square cavity domain.

Usage

from datasets import load_dataset
import numpy as np
import matplotlib.pyplot as plt

# Load geometry
ds_geom = load_dataset("SISSAmathLab/lid-cavity", name="geometry")

# Load snapshots
ds_data = load_dataset("SISSAmathLab/lid-cavity", name="snapshots")

# Load parameters
ds_params = load_dataset("SISSAmathLab/lid-cavity", name="parameters")

# Visualize velocity magnitude for first simulation
pts_x = np.asarray(ds_geom['default']['node_coordinates_x']).flatten()
pts_y = np.asarray(ds_geom['default']['node_coordinates_y']).flatten()
velocity_mag = ds_data['default']['velocity_magnitude'][0]

plt.scatter(pts_x, pts_y, c=velocity_mag, cmap='jet', s=1)
plt.colorbar(label='Velocity Magnitude')
plt.title('Lid-Driven Cavity Flow')
plt.xlabel('x')
plt.ylabel('y')
plt.axis('equal')
plt.show()

Contact

For questions or issues, please contact SISSA mathLab.