Learned Turbulence Modelling with Differentiable Fluid Solvers: Physics-based Loss-functions and Optimisation Horizons

Abstract

In this paper, we train turbulence models based on convolutional neural networks. These learned turbulence models improve under-resolved low resolution solutions to the incompressible Navier-Stokes equations at simulation time. Our study involves the development of a differentiable numerical solver that supports the propagation of optimisation gradients through multiple solver steps. The significance of this property is demonstrated by the superior stability and accuracy of those models that unroll more solver steps during training. Furthermore, we introduce loss terms based on turbulence physics that further improve the model accuracy. This approach is applied to three two-dimensional turbulence flow scenarios, a homogeneous decaying turbulence case, a temporally evolving mixing layer, and a spatially evolving mixing layer. Our models achieve significant improvements of long-term a-posteriori statistics when compared to no-model simulations, without requiring these statistics to be directly included in the learning targets. At inference time, our proposed method also gains substantial performance improvements over similarly accurate, purely numerical methods.

Published in the Journal of Fluid Mechanics (JFM), code & data available at: https://github.com/tum-pbs/differentiable-piso

References in corpus (4)

• Machine learning accelerated computational fluid dynamics
• Prediction of Aerodynamic Flow Fields Using Convolutional Neural Networks
• Embedded training of neural-network sub-grid-scale turbulence models
• High-order methods for decaying two-dimensional homogeneous isotropic turbulence

Cited by in corpus (9)

• Weak baselines and reporting biases lead to overoptimism in machine learning for fluid-related partial differential equations
• Optimizing a DIscrete Loss (ODIL) to solve forward and inverse problems for partial differential equations using machine learning tools
• XLB: A differentiable massively parallel lattice Boltzmann library in Python
• A Multifidelity deep operator network approach to closure for multiscale systems
• Discretize first, filter next: learning divergence-consistent closure models for large-eddy simulation
• Optimized Flow Control based on Automatic Differentiation in Compressible Turbulent Channel Flows
• Reduced Data-Driven Turbulence Closure for Capturing Long-Term Statistics
• Online learning in idealized ocean gyres
• PICT -- A Differentiable, GPU-Accelerated Multi-Block PISO Solver for Simulation-Coupled Learning Tasks in Fluid Dynamics