1 April 2022 in Physics of Fluids
Turbulence modeling based on the Reynolds-averaged Navier–Stokes (RANS) method has been widely applied in industry, but its performance in some complex flows is far from satisfactory. The improvement of turbulence models based on the traditional framework has not made breakthrough progress for decades. In this study, a data-driven turbulence modeling framework based on the reconstruction of Reynolds stress discrepancies is used to aid in the improvement of turbulence models, with the Reynolds stresses of the shear-stress transport model being modified in the eigenspace. The large eddy simulation (LES) dataset of a set of bump cases is used to provide high-fidelity information on adverse-pressure-gradient flows for the modeling framework. First, the Reynolds stress tensors of RANS and LES are compared in terms of amplitude, shape, and orientation. Then, the random forest (RF) algorithm is employed to map the mean flow features to the Reynolds stress discrepancies. The well-trained RF model greatly improves the predictions of Reynolds stresses and other flow variables for the attachment and separation states and enables the numerical simulations to have predictive accuracy close to LES and computation time of the same order of magnitude as RANS.
1 February 2023 in Physics of Fluids
The classical four-equation γ-Re transition model has presented excellent accuracy in low-speed boundary layer transition prediction. However, once the incoming flow reaches hypersonic speed, the original model is no longer applicable due to the compressibility problem and the appearance of multiple instability modes. Recently, there has been widespread interest in data-driven modeling for quantifying uncertainty or improving model prediction accuracy. In this paper, a data-driven framework based on field inversion and machine learning is performed to extend the prediction capability of the original γ-Re transition model for the hypersonic boundary layer transition.First, the iterative regularized ensemble Kalman filter method is applied to obtain the spatial distribution of the perturbation correction term β for the switching function Fonset1, and the effectiveness of this method is initially verified in the hypersonic flat plate case. Then, the random forest algorithm is adopted to construct a mapping from the average flow features to β. The generalizability of the well-trained learning model is fully validated in the blunt cone cases with different unit Reynolds numbers, free-stream flow temperature, and bluntness. The simulation results indicate that the performance of the original γ-Re transition model in the hypersonic boundary layer transition prediction is significantly improved, and the boundary layer transition onset location and the length of transition zone can be correctly obtained. In addition, the machine learning model investigates the importance of the input features and confirms that the effective length scale plays a significant role in the numerical simulation of the hypersonic boundary layer transition.
13 June 2024 in Acta Astronautica
1 November 2023 in Physics of Fluids
Due to the difficulties of precisely calculating the turbulence intensity within the separated shear layer using traditional turbulent models, computing strongly separated flows is a key task of considerable interest. In this paper, the Reynolds stress functional expression is improved toward an explicit algebraic stress model for separated flows that are similar to hump flows using a data-driven framework of field inversion and machine learning that can take model consistency into account during the model correction process. The iterative Kalman filter algorithm is utilized to address the inversion problem, and the inversion results are used as training data for correction models that are trained using random forest regression. For model verification and validation, we employ both the curved backward-facing step and bump cases. The findings indicate that the inversion produces favorable outcomes, and the enhanced model developed utilizing the inversion data exhibits good generalizability.
4 January 2024 in Acta Astronautica
25 September 2025 in Journal of Fluid Mechanics
Turbulence amplification is crucial in shock-wave/turbulent boundary layer interaction (SWTBLI). To examine the impact of interaction intensity on turbulence amplification and inter-component energy transfer, direct numerical simulations of impinging oblique shock reflections at strong ( $37^\circ$ ) and weak ( $33.2^\circ$ ) incident angles are conducted. The results indicate that strong interaction generates a larger permanent separation zone, featuring the unique ‘oblique platform’ in Reynolds stress peaks and ‘secondary turbulence amplification’ downstream. Reynolds stress budget and spanwise spectral analyses reveal that $\widetilde {u^{\prime \prime}u^{\prime \prime}}$ and $-\!\widetilde{\ u^{\prime\prime}v^{\prime\prime}}$ amplify primarily by production terms. $u''$ , $v''$ and $w''$ represent the streamwise, wall-normal and spanwise velocity fluctuations. At the investigated Reynolds number, deceleration effect dominates the initial amplification of $\widetilde {u^{\prime \prime}u^{\prime \prime}}$ , influencing multi-scale wall-bounded turbulence structures, while shear effect remains active along the shear layer and may primarily affects streaky structures. The initial amplification of $-\!\widetilde{\ u^{\prime\prime}v^{\prime\prime}}$ is driven by the adverse pressure gradient, which reshapes the velocity profile and affects the wall-normal velocity. The primary energy for $\!\widetilde{\ v^{\prime\prime}v^{\prime\prime}}$ and $\widetilde {w^{\prime \prime}w^{\prime \prime}}$ amplification originates from $\widetilde{ u^{\prime \prime}u^{\prime \prime}}$ via the pressure-strain term. The delayed amplification of $\!\widetilde{\ v^{\prime\prime}v^{\prime\prime}}$ is influenced by its production term and energy redistribution, with $\widetilde {w^{\prime \prime}w^{\prime \prime}}$ exhibiting higher spectral consistency with $\widetilde {u^{\prime \prime}u^{\prime \prime}}$ and receiving more energy. In strong interaction, the ‘oblique platform’ serves as a stable dissipation region, formed by increased separation–incident shock distance, characterised by progressively concentrated stress spectra and the transition to large-scale streaks. The downstream ‘secondary amplification’ process resembles the initial amplification near the separation shock foot, driven by intermittent compression waves that strengthen shear instabilities and the deceleration effect. These findings detail the streamwise stress evolution, providing a more comprehensive turbulence amplification mechanism in SWTBLI.