Turbulence closure modeling with data-driven techniques: Investigation of generalizable deep neural networks

TL;DR

This paper investigates the ability of standard neural networks to generalize in turbulence closure modeling, revealing that larger networks are needed for accuracy and that extrapolation significantly reduces performance, highlighting challenges in practical applications.

Contribution

The study systematically examines the approximation capabilities of moderate-sized fully-connected neural networks for turbulence modeling using proxy-physics systems, emphasizing the effects of complexity, sampling, and optimization.

Findings

Neural networks require more degrees of freedom than proxy models for accurate approximation.

Approximation capability drops significantly when trained on limited parameter space data.

Finding optimal architectures for extrapolation remains challenging.

Abstract

Generalizability of machine-learning (ML) based turbulence closures to accurately predict unseen practical flows remains an important challenge. At the Reynolds-averaged Navier-Stokes (RANS) level, NN-based turbulence closure modeling is rendered difficult due to two important reasons: inherent complexity of the constitutive relation arising from flow-dependent non-linearity and bifurcations; and, inordinate difficulty in obtaining high-fidelity data covering the entire parameter space of interest. In this context, the objective of the work is to investigate the approximation capabilities of standard moderate-sized fully-connected NNs. We seek to systematically investigate the effects of: (i) intrinsic complexity of the solution manifold; (ii) sampling procedure (interpolation vs. extrapolation) and (iii) optimization procedure. To overcome the data acquisition challenges, three proxy-physics turbulence surrogates of different degrees of complexity (yet significantly simpler than turbulence physics) are employed to generate the parameter-to-solution maps. Even for this simple proxy-physics system, it is demonstrated that feed-forward NNs require more degrees of freedom than the original proxy-physics model to accurately approximate the true model even when trained with data over the entire parameter space (interpolation). Additionally, if deep fully-connected NNs are trained with data only from part of the parameter space (extrapolation), their approximation capability reduces considerably and it is not straightforward to find an optimal architecture. Overall, the findings provide a realistic perspective on the utility of ML turbulence closures for practical applications and identify areas for improvement.