Benchmarking Autoregressive Conditional Diffusion Models for Turbulent Flow Simulation

TL;DR

This paper benchmarks autoregressive conditional diffusion models for turbulent flow simulation, demonstrating their potential for improved accuracy and stability over traditional methods, with added probabilistic benefits.

Contribution

It introduces a comprehensive benchmark for data-driven flow prediction methods, highlighting the effectiveness of diffusion models in turbulent flow simulation.

Findings

Diffusion-based models outperform traditional methods in accuracy and stability.

Probabilistic diffusion approaches can generate multiple physics-consistent predictions.

Traditional architectures are faster but lack probabilistic sampling capabilities.

Abstract

Simulating turbulent flows is crucial for a wide range of applications, and machine learning-based solvers are gaining increasing relevance. However, achieving temporal stability when generalizing to longer rollout horizons remains a persistent challenge for learned PDE solvers. In this work, we analyze if fully data-driven fluid solvers that utilize an autoregressive rollout based on conditional diffusion models are a viable option to address this challenge. We investigate accuracy, posterior sampling, spectral behavior, and temporal stability, while requiring that methods generalize to flow parameters beyond the training regime. To quantitatively and qualitatively benchmark the performance of various flow prediction approaches, three challenging 2D scenarios including incompressible and transonic flows, as well as isotropic turbulence are employed. We find that even simple diffusion-based approaches can outperform multiple established flow prediction methods in terms of accuracy and temporal stability, while being on par with state-of-the-art stabilization techniques like unrolling at training time. Such traditional architectures are superior in terms of inference speed, however, the probabilistic nature of diffusion approaches allows for inferring multiple predictions that align with the statistics of the underlying physics. Overall, our benchmark contains three carefully chosen data sets that are suitable for probabilistic evaluation alongside various established flow prediction architectures.