New pipeline generates ML-ready datasets for 3D fluid dynamics simulations

Researchers have developed ChannelFlow-Tools, an open-source pipeline designed to generate machine-learning-ready datasets for 3D obstructed channel flows. This configuration-driven system integrates procedural obstacle generation, signed-distance-field voxelization, and Lattice-Boltzmann simulations to create datasets for computational fluid dynamics. The pipeline ensures byte-identical reproducibility for geometry generation and has undergone extensive validation, demonstrating its capability to produce physically consistent data for training surrogate models. AI

IMPACT Enables more robust and auditable training data generation for CFD surrogate models, potentially accelerating research in the field.

RANK_REASON Research paper detailing a new tool for generating ML datasets. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• 3D U-Net
• alphaXiv
• arXiv
• CatalyzeX
• ChannelFlow-Tools
• DagsHub
• Fourier Neural Operators
• Gotit.pub
• Hugging Face
• ScienceCast
• Shubham Kavane

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →