Sensitivity Analysis of Discrepancy Terms introduced in Turbulence Models using Field Inversion

TL;DR

This paper enhances RANS turbulence modeling for flow separation by introducing and optimizing discrepancy terms via field inversion and machine learning, demonstrating improved simulation accuracy and neural network augmentation.

Contribution

It introduces a discrepancy term in turbulence models, optimized with experimental data, and employs neural networks to augment RANS simulations for better flow separation predictions.

Findings

Optimized discrepancy terms improve RANS accuracy in flow separation cases.

Regularization and grid resolution significantly influence optimization results.

Neural network augmentation enhances simulation predictions on test cases.

Abstract

RANS simulations with the Spalart-Allmaras turbulence model are improved for cases with flow separation using the Field Inversion and Machine Learning approach. A compensatory discrepancy term is introduced into the turbulence model and optimized using high-fidelity reference data from experiments. Influences on the optimization results with respect to regularization, grid resolution and areas in which the optimization is active are investigated. Finally, a neural network is trained and used to augment simulations on a test case.