Development of a Generalizable Data-driven Turbulence Model: Conditioned Field Inversion and Symbolic Regression

Chenyu Wu,Shaoguang Zhang,Yufei Zhang

2024-05-21

Abstract:This paper addresses the issue of predicting separated flows with Reynolds-averaged Navier-Stokes (RANS) turbulence models, which are essential for many engineering tasks. Traditional RANS models usually struggle with this task, so recent efforts have focused on data-driven methods such as field inversion and machine learning (FIML) to correct this issue by adjusting the baseline equations. However, these FIML methods often reduce accuracy in attached boundary layers. To address this issue, we developed a "conditioned field inversion" technique. This method adjusts the corrective factor \b{eta} (used by FIML) in the shear-stress transport (SST) model. It multiplies \b{eta} with a shield function f_d that is off in the boundary layer and on elsewhere. This maintains the accuracy of the baseline model for the attached flows. We applied both conditioned and classic field inversion to the NASA hump and a curved backward-facing step (CBFS), creating two datasets. These datasets were used to train two models: SR-CND (from our new method) and SR-CLS (from the traditional method). The SR-CND model matches the SR-CLS model in predicting separated flows in various scenarios, such as periodic hills, the NLR7301 airfoil, the 3D SAE car model, and the 3D Ahmed body, and significantly outperforms the baseline SST model. Importantly, the SR-CND model maintains accuracy in the attached boundary layers, whereas the SR-CLS model does not. Therefore, the proposed method improves separated flow predictions while maintaining the accuracy of the original model for attached flows, offering a better way to create data-driven turbulence models.

Fluid Dynamics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to improve the accuracy of the Reynolds - averaged Navier - Stokes (RANS) turbulence model in predicting separated flows while maintaining its precision in wall - bounded layer flows. Traditional RANS models perform poorly in predicting separated flows, and existing data - driven methods such as Field Inversion and Machine Learning (FIML) can improve this problem, but often reduce the precision in wall - bounded layer flows. Therefore, this paper proposes a new method - Conditioned Field Inversion (FI - CND), aiming to improve the prediction of separated flows without compromising the original precision of the RANS model in wall - bounded layer flows.

Specifically, the main problems mentioned in the paper include:
1. **Limitations of the RANS model**: Traditional RANS models have significant errors when predicting separated flows, which limits their application in many engineering tasks.
2. **Deficiencies of the FIML method**: Although the FIML method improves the prediction of separated flows by introducing a spatially varying correction factor \(\beta(\mathbf{x})\), it reduces the precision of the model in wall - bounded layer flows.
3. **L2 generalization ability**: The precision of existing FIML models decreases in simple wall - bounded layer flows, which is a key problem in practical engineering applications.

To address these problems, the paper proposes the FI - CND method. By introducing a shielding function \(f_{\text{wall}}\) to adjust the correction factor \(\beta(\mathbf{x})\), it makes \(\beta(\mathbf{x})\) only effective in the separation zone and remain unchanged in the wall - bounded layer. This can ensure that the precision of the model in wall - bounded layer flows is not affected while improving the prediction ability of separated flows.