Muhammad I. Zafar,Xuhui Zhou,Christopher J. Roy,David Stelter,Heng Xiao
Abstract:Wall-cooling effect in hypersonic boundary layers can significantly alter the near-wall turbulence behavior, which is not accurately modeled by traditional RANS turbulence models. To address this shortcoming, this paper presents a turbulence modeling approach for hypersonic flows with cold-wall conditions using an iterative ensemble Kalman method. Specifically, a neural-network-based turbulence model is used to provide closure mapping from mean flow quantities to Reynolds stress as well as a variable turbulent Prandtl number. Sparse observation data of velocity and temperature are used to train the turbulence model. This approach is analyzed using direct numerical simulation database for boundary layer flows over a flat plate with a Mach number between 6 and 14 and wall-to-recovery temperature ratios ranging from 0.18 to 0.76. Two training cases are conducted: 1) a single training case with observation data from one flow case, 2) a joint training case where data from two flow cases are simultaneously used for training. Trained models are also tested for generalizability on the remaining flow cases in each of the training cases. The results are also analyzed for insights to inform the future work towards enhancing the generalizability of the learned turbulence model.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in hypersonic boundary - layer flows, cold - wall conditions have a significant impact on near - wall turbulent behavior, while traditional RANS (Reynolds - Averaged Navier - Stokes) turbulence models are unable to accurately simulate this impact. Specifically, the cold - wall effect will lead to a significant increase in the thermal gradient in the near - wall region, thus changing the turbulence structure, which has not been fully considered in traditional turbulence models.
To solve this problem, this paper proposes a data - driven turbulence modeling method based on the iterative ensemble Kalman method, especially for hypersonic cold - wall boundary - layer flows. This method uses neural networks to provide a closure mapping from mean - flow variables to Reynolds stresses and variable turbulence Prandtl numbers, and is trained with sparse velocity and temperature observation data. Through this method, researchers hope to improve the prediction accuracy of hypersonic boundary - layer turbulent behavior under cold - wall conditions, especially the prediction of wall heat flux and friction.
### Specific problem description
1. **Impact of the cold - wall effect**:
- In hypersonic boundary - layer flows under cold - wall conditions, the turbulent behavior in the near - wall region is different from what is assumed by traditional RANS models.
- Traditional RANS models will overestimate wall heat flux and friction under cold - wall conditions, especially in high - Mach - number cases, and the error may be as high as 30%.
2. **Limitations of existing models**:
- Existing turbulence models are usually calibrated for incompressible flow conditions and perform poorly for hypersonic flows, especially those under cold - wall conditions.
- Assuming a constant turbulence Prandtl number is a source of uncertainty in existing models, especially in hypersonic flows, and this assumption may lead to large prediction errors.
### Proposed solutions
1. **Data - driven turbulence model**:
- Use neural networks (NN) to learn the mapping relationship from mean - flow variables to Reynolds stresses and variable turbulence Prandtl numbers.
- Use the iterative ensemble Kalman method to optimize model parameters, ensuring that the model can adapt to data under different working conditions.
2. **Utilization of sparse observation data**:
- Use sparse velocity and temperature observation data for model training, which are from the direct numerical simulation (DNS) database.
- Train and verify the generalization ability of the model by considering flow situations with different Mach numbers and wall - to - recovery temperature ratios.
3. **Consistent training framework of the model**:
- During the training process, combine with the RANS solver to ensure that the model output is consistent with actual physical phenomena.
- Through non - invasive methods (such as the ensemble Kalman method), reduce the implementation complexity and improve the training efficiency.
### Summary
This paper aims to improve the prediction of turbulent behavior in hypersonic cold - wall boundary - layer flows, especially to improve the prediction accuracy of wall heat flux and friction, by introducing a new data - driven turbulence modeling method. This method not only considers the impact of the cold - wall effect on the turbulence structure, but also improves the robustness and generalization ability of the model through machine - learning techniques.