NeurIPS 2024 ML4CFD Competition: Harnessing Machine Learning for Computational Fluid Dynamics in Airfoil Design

TL;DR

This paper presents the NeurIPS 2024 ML4CFD competition, which targets ML-based surrogate modeling for CFD in airfoil design using the AirfRANS dataset within the Learning Industrial Physical Simulation (LIPS) framework. The setup emphasizes industrially relevant evaluation criteria—ML accuracy, computational speed, OOD generalization, and physics compliance—to benchmark surrogate models against OpenFOAM-based ground truth. It details the data, tasks, scoring rules, baselines, and available resources, including GPU-powered evaluation on Codabench and a structured progression of phases, tutorials, and open-source incentives. The work demonstrates a concerted effort to drive practical, physics-aware ML methods for fast, reliable CFD in engineering contexts, with a roadmap for broader industrial applicability and community engagement.

Abstract

The integration of machine learning (ML) techniques for addressing intricate physics problems is increasingly recognized as a promising avenue for expediting simulations. However, assessing ML-derived physical models poses a significant challenge for their adoption within industrial contexts. This competition is designed to promote the development of innovative ML approaches for tackling physical challenges, leveraging our recently introduced unified evaluation framework known as Learning Industrial Physical Simulations (LIPS). Building upon the preliminary edition held from November 2023 to March 2024, this iteration centers on a task fundamental to a well-established physical application: airfoil design simulation, utilizing our proposed AirfRANS dataset. The competition evaluates solutions based on various criteria encompassing ML accuracy, computational efficiency, Out-Of-Distribution performance, and adherence to physical principles. Notably, this competition represents a pioneering effort in exploring ML-driven surrogate methods aimed at optimizing the trade-off between computational efficiency and accuracy in physical simulations. Hosted on the Codabench platform, the competition offers online training and evaluation for all participating solutions.

Figures (2)

• Figure 1: Learning Industrial Physical Simulation (LIPS) framework
• Figure 2: (a) Airfoil mesh. (b) $x$-velocity field.