Koopman-Based Surrogate Modelling of Turbulent Rayleigh-B\'enard Convection

TL;DR

This paper demonstrates that Koopman-inspired neural networks, specifically LRAN, can effectively model complex turbulent Rayleigh-Bénard convection flows, outperforming traditional methods like KDMD in highly turbulent regimes.

Contribution

The study introduces the use of LRAN for reduced-order modeling of turbulent convection flows, showing its advantages over KDMD in capturing complex dynamics.

Findings

LRAN outperforms KDMD in high turbulence scenarios.

LRAN effectively learns complex observables from turbulent data.

Koopman-based models enable potential control applications in turbulent flows.

Abstract

Several related works have introduced Koopman-based Machine Learning architectures as a surrogate model for dynamical systems. These architectures aim to learn non-linear measurements (also known as observables) of the system's state that evolve by a linear operator and are, therefore, amenable to model-based linear control techniques. So far, mainly simple systems have been targeted, and Koopman architectures as reduced-order models for more complex dynamics have not been fully explored. Hence, we use a Koopman-inspired architecture called the Linear Recurrent Autoencoder Network (LRAN) for learning reduced-order dynamics in convection flows of a Rayleigh B\'enard Convection (RBC) system at different amounts of turbulence. The data is obtained from direct numerical simulations of the RBC system. A traditional fluid dynamics method, the Kernel Dynamic Mode Decomposition (KDMD), is used to compare the LRAN. For both methods, we performed hyperparameter sweeps to identify optimal settings. We used a Normalized Sum of Square Error measure for the quantitative evaluation of the models, and we also studied the model predictions qualitatively. We obtained more accurate predictions with the LRAN than with KDMD in the most turbulent setting. We conjecture that this is due to the LRAN's flexibility in learning complicated observables from data, thereby serving as a viable surrogate model for the main structure of fluid dynamics in turbulent convection settings. In contrast, KDMD was more effective in lower turbulence settings due to the repetitiveness of the convection flow. The feasibility of Koopman-based surrogate models for turbulent fluid flows opens possibilities for efficient model-based control techniques useful in a variety of industrial settings.