FLUID-LLM: Learning Computational Fluid Dynamics with Spatiotemporal-aware Large Language Models

TL;DR

FLUID-LLM introduces a novel framework that combines pre-trained large language models with spatiotemporal-aware encoding to improve the prediction of unsteady fluid dynamics, addressing complex geometries in CFD.

Contribution

The paper presents a new method integrating LLMs with spatiotemporal encoding specifically designed for CFD prediction, which is a novel application of LLMs in fluid dynamics.

Findings

Significant performance improvements on standard CFD benchmarks.

Effective integration of spatiotemporal information into LLMs.

Enhanced prediction accuracy for unsteady fluid flows.

Abstract

Learning computational fluid dynamics (CFD) traditionally relies on computationally intensive simulations of the Navier-Stokes equations. Recently, large language models (LLMs) have shown remarkable pattern recognition and reasoning abilities in natural language processing (NLP) and computer vision (CV). However, these models struggle with the complex geometries inherent in fluid dynamics. We introduce FLUID-LLM, a novel framework combining pre-trained LLMs with spatiotemporal-aware encoding to predict unsteady fluid dynamics. Our approach leverages the temporal autoregressive abilities of LLMs alongside spatial-aware layers, bridging the gap between previous CFD prediction methods. Evaluations on standard benchmarks reveal significant performance improvements across various fluid datasets. Our results demonstrate that FLUID-LLM effectively integrates spatiotemporal information into pre-trained LLMs, enhancing CFD task performance.