Towards scalable surrogate models based on Neural Fields for large scale aerodynamic simulations

2025-05-14

AI Summary

Traditional CFD simulations for large-scale aerodynamic analysis suffer from prohibitive computational cost, while existing surrogate models rely on mesh-based discretizations and parametric geometry representations, limiting generalizability. To address these challenges, this paper proposes MARIO, a scalable neural-field-based aerodynamic surrogate model. Its core contributions are: (1) a novel modulation-based aerodynamic resolution-invariant operator that decouples shape encoding from multi-scale modeling, enabling non-parametric geometric representation and mesh-agnostic inference; and (2) support for coarse-mesh training and full-resolution inference, drastically reducing memory footprint and computational overhead. On the AirfRANS dataset, MARIO improves pressure and aerodynamic coefficient prediction accuracy by an order of magnitude and accurately captures boundary-layer structures. In NASA CRM 3D wing tests at million-node scale, it achieves hundreds-of-times speedup over CFD while preserving physical fidelity—demonstrating strong scalability and accuracy.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

This paper introduces a novel surrogate modeling framework for aerodynamic applications based on Neural Fields. The proposed approach, MARIO (Modulated Aerodynamic Resolution Invariant Operator), addresses non parametric geometric variability through an efficient shape encoding mechanism and exploits the discretization-invariant nature of Neural Fields. It enables training on significantly downsampled meshes, while maintaining consistent accuracy during full-resolution inference. These properties allow for efficient modeling of diverse flow conditions, while reducing computational cost and memory requirements compared to traditional CFD solvers and existing surrogate methods. The framework is validated on two complementary datasets that reflect industrial constraints. First, the AirfRANS dataset consists in a two-dimensional airfoil benchmark with non-parametric shape variations. Performance evaluation of MARIO on this case demonstrates an order of magnitude improvement in prediction accuracy over existing methods across velocity, pressure, and turbulent viscosity fields, while accurately capturing boundary layer phenomena and aerodynamic coefficients. Second, the NASA Common Research Model features three-dimensional pressure distributions on a full aircraft surface mesh, with parametric control surface deflections. This configuration confirms MARIO's accuracy and scalability. Benchmarking against state-of-the-art methods demonstrates that Neural Field surrogates can provide rapid and accurate aerodynamic predictions under the computational and data limitations characteristic of industrial applications.

Problem

• Develop scalable Neural Field surrogates for large aerodynamic simulations
• Address non-parametric geometric variability with efficient shape encoding
• Reduce computational cost while maintaining full-resolution accuracy

Innovation

• Neural Fields for scalable surrogate models
• Efficient shape encoding for geometric variability
• Training on downsampled meshes, full-resolution inference