Minimum-dissipation model for large-eddy simulation using symmetry-preserving discretization in OpenFOAM

Abstract

The minimum-dissipation model is applied to channel flow up to $Re_\tau = 2000$, flow past a circular cylinder at $Re=3900$, and flow over periodic hills at $Re=10595$. Numerical simulations were performed in OpenFOAM which is based on the finite volume methods. We used both symmetry-preserving and standard second-order accurate discretization methods in OpenFOAM on structured meshes. The results are compared to DNS and experimental data. The results of channel flow demonstrate a static QR model performs equally well as the dynamic models while reducing the computational cost. The model constant of $C=0.024$ gives the most accurate prediction, and the contribution of the sub-grid model decreases with the increase of the mesh resolution and becomes very small (less than 0.2 molecular viscosity) if a fine mesh is used. Furthermore, the QR model is able to predict the mean and rms velocity accurately up to $Re_\tau = 2000$ without a wall damping function. The symmetry-preserving discretization outperforms the standard OpenFOAM discretization at $Re_\tau=1000$. The results for the flow over a cylinder show that the mean velocity, drag coefficient, and lift coefficient are in good agreement with the experimental data and the central difference schemes conjugated with the QR model predict better results. The various comparisons carried out for flows over periodic hills demonstrate the need to use central difference schemes in OpenFOAM in combination with the minimum dissipation model. The best model constant is again $C=0.024$. The single wind turbine simulation shows that the QR model is capable of predicting accurate results in complex rotating scenarios.