Generative Modeling of Turbulence

TL;DR

This paper introduces a mathematically grounded approach using GANs to generate high-resolution turbulent flow fields efficiently, demonstrating their ability to learn invariant measures and generalize across geometries in complex flow scenarios.

Contribution

The paper provides a theoretical proof that GANs can learn the invariant measure of chaotic systems and applies this to turbulent flow modeling, showcasing efficient, high-resolution flow synthesis with generalization capabilities.

Findings

GANs can learn the invariant measure of chaotic systems.

GAN-based turbulence modeling is faster and requires less data than classical methods.

Conditional GANs can generalize to unseen geometries in turbulent flow simulations.

Abstract

We present a mathematically well founded approach for the synthetic modeling of turbulent flows using generative adversarial networks (GAN). Based on the analysis of chaotic, deterministic systems in terms of ergodicity, we outline a mathematical proof that GAN can actually learn to sample state snapshots form the invariant measure of the chaotic system. Based on this analysis, we study a hierarchy of chaotic systems starting with the Lorenz attractor and then carry on to the modeling of turbulent flows with GAN. As training data, we use fields of velocity fluctuations obtained from large eddy simulations (LES). Two architectures are investigated in detail: we use a deep, convolutional GAN (DCGAN) to synthesise the turbulent flow around a cylinder. We furthermore simulate the flow around a low pressure turbine stator using the pix2pixHD architecture for a conditional DCGAN being conditioned on the position of a rotating wake in front of the stator. The settings of adversarial training and the effects of using specific GAN architectures are explained. We thereby show that GAN are efficient in simulating turbulence in technically challenging flow problems on the basis of a moderate amount of training data. GAN training and inference times significantly fall short when compared with classical numerical methods, in particular LES, while still providing turbulent flows in high resolution. We furthermore analyse the statistical properties of the synthesized and LES flow fields, which agree excellently. We also show the ability of the conditional GAN to generalize over changes of geometry by generating turbulent flow fields for positions of the wake that are not included in the training data.