Reliable coarse-grained turbulent simulations through combined offline learning and neural emulation

TL;DR

This paper presents a method combining offline learning and neural emulation to enhance the stability and accuracy of coarse-grained turbulent fluid simulations by better modeling unresolved dynamics.

Contribution

It introduces an additional loss network that emulates future states, improving stability and capturing subgrid processes in offline-trained ML models for turbulence simulation.

Findings

Enhanced stability of ML-integrated turbulence simulations

Improved modeling of subgrid-scale processes

Successful application to 2D quasi-geostrophic turbulence

Abstract

Integration of machine learning (ML) models of unresolved dynamics into numerical simulations of fluid dynamics has been demonstrated to improve the accuracy of coarse resolution simulations. However, when trained in a purely offline mode, integrating ML models into the numerical scheme can lead to instabilities. In the context of a 2D, quasi-geostrophic turbulent system, we demonstrate that including an additional network in the loss function, which emulates the state of the system into the future, produces offline-trained ML models that capture important subgrid processes, with improved stability properties.