README.md

metadata

license: cc-by-nc-4.0

SURF: A Generalisation Benchmark for GNNs Predicting Fluid Dynamics

SURF, is a benchmark designed to test the generalization of learned graph-based fluid simulators. The benchmark consists of seven independent datasets:

• Base
• Turned
• Topo
• Range
• Dynamic
• Full
• FullFiner

Each dataset is available as separate *.zip file and consists of at least 1200 2D incompressible fluid flow simulations with 300 timesteps. The data structure is as follows:

• folder: dataset_name
  • folders: dpx
    • files: sim.npz, triangles.py, constrained_kmeans_20.npy, Simulation_dp1_Timestep_50.png
  • folder: Splits
    • files: train.txt, test.txt, valid.txt

The file sim.npz (numpy archive) contains the result of the simulation for each timestep at each node:

• 'pointcloud': x, y coordinates
• 'VX': velocity in x-direction
• 'VY': velocity in y-direction
• 'PS': static pressure
• 'PG': dynamic pressure
• 'T': temperature
• 'TC': thermal conductivity of fluid
• 'HC': heat capacity of fluid

The results have the following shape: VX.shape=(#timesteps, #nodes, 1).

The file triangles.py contains the mesh connectivity. triangles.shape=(#timesteps, #elements, 3). Each triangle is defined by the node numbers in counter clockwise direction.