174 research outputs found
Direct numerical simulation of turbulence at lower costs
No full textDirect Numerical Simulation (DNS) is the most accurate, but also the most expensive, way of computing turbulent flow. To cut the costs of DNS we consider a family of second-order, explicit one-leg time-integration methods and look for the method with the best linear stability properties. It turns out that this method requires about two times less computational effort than Adams–Bashforth. Next, we discuss a fourth-order finite-volume method that is constructed as the Richardson extrapolate of a classical second-order method. We compare the results of this fourth-order method and the underlying second-order method for a DNS of the flow in a cubical driven cavity at Re = 10^4. Experimental results are available for comparison. For this example, the fourth-order results are clearly superior to the second-order results, whereas their computational effort is about twenty times less. With the improved simulation method, a DNS of a turbulent flow in a cubical lid-driven flow at Re = 50,000 and a DNS of a turbulent flow past a square cylinder at Re = 22,000 are performed.
Parameter-Free Symmetry-Preserving Regularization Modelling of Turbulent Natural Convection Flows
No full textNumerical Simulation of a Turbulent Flow in a Channel with Surface Mounted Cubes
No full textIn this paper we report on a fourth-order, spectro-consistent simulation of a complex turbulent flow. A spatial discretization of a convection-diffusion equation is termed spectro-consistent if the spectral properties of the convective and diffusive operators are preserved, i.e. convection ↔ skew-symmetric; diffusion ↔ symmetric positive definite. We consider a fully developed flow in a channel, where a matrix of cubes is placed at a wall of the channel. The Reynolds number (based on the channel width and the mean bulk velocity) is equal to Re = 13,000. The three-dimensional flow around the surface mounted cubes has served at a test case at the 6th ERCOFTAC/IAHR/COST workshop on refined flow modeling (Delft, June 1997). Here, mean velocity profiles as well as Reynolds stresses at various locations in the channel have been computed without using any turbulence models. The results agree well with the available experimental data.
Truncation of scales by relaxation
No full textThis paper is about a relaxation model for large-eddy simulation of turbulent flow that truncates the too small scales of motion by making sure that they do not get energy from the larger eddies. To verify that a box filter is introduced and the relaxation parameter is determined in such a way that the production of small, box-fitting scales is counteracted by the modeled dissipation. This dissipation-production balance is worked out with the help of Poincar\'{e}'s inequality, which results in a relaxation model that depends on the invariants of the velocity gradient. This model is discretized and equipped with a Schumann filter. It is successfully tested for isotropic turbulence as well as for turbulent channel flow
Uncertainty quantification analysis for simulation of wakes in wind-farms using a stochastic RANS solver, compared with a deep learning approach
No full textQuantification of uncertainties in Reynolds–Averaged Navier–Stokes (RANS) simulations has gained a considerable interest in turbulence modeling. We present two different approaches for the quantification and propagation of model-form and operational uncertainties in context of wind turbine RANS simulations. The first approach is based on a stochastic RANS solver in OpenFOAM using intrusive polynomial chaos method (Parekh and Verstappen, 2023). Here the uncertainties are propagated through a single (large) simulation for the coupled coefficients of the polynomial expansion. The second approach is a surrogate based uncertainty quantification (SBUQ) method. The surrogate model comprises of a 3D U-Net neural network (trained over a single wind turbine) combined with a wake superposition model in order to the prediction of flow field in an array of wind turbines. The above-mentioned approaches are applied for uncertainty quantification analysis in RANS simulations of two turbulent engineering flow problems — (i) a wake past a single wind turbine, and (ii) wake interactions and power losses in an array of wind turbines. The results show that the uncertain RANS solutions from the two approaches are able to reasonably capture the reference high-fidelity solution. We also discuss comparisons between the two approaches including computational cost, applicability, generality etc. The two methods can be further explored and applied to engineering applications where it is critical to compute the turbulent RANS solution in presence of various sources of uncertainties
Uncertainty quantification analysis for simulation of wakes in wind-farms using a stochastic RANS solver, compared with a deep learning approach
No full textQuantification of uncertainties in Reynolds–Averaged Navier–Stokes (RANS) simulations has gained a considerable interest in turbulence modeling. We present two different approaches for the quantification and propagation of model-form and operational uncertainties in context of wind turbine RANS simulations. The first approach is based on a stochastic RANS solver in OpenFOAM using intrusive polynomial chaos method (Parekh and Verstappen, 2023). Here the uncertainties are propagated through a single (large) simulation for the coupled coefficients of the polynomial expansion. The second approach is a surrogate based uncertainty quantification (SBUQ) method. The surrogate model comprises of a 3D U-Net neural network (trained over a single wind turbine) combined with a wake superposition model in order to the prediction of flow field in an array of wind turbines. The above-mentioned approaches are applied for uncertainty quantification analysis in RANS simulations of two turbulent engineering flow problems — (i) a wake past a single wind turbine, and (ii) wake interactions and power losses in an array of wind turbines. The results show that the uncertain RANS solutions from the two approaches are able to reasonably capture the reference high-fidelity solution. We also discuss comparisons between the two approaches including computational cost, applicability, generality etc. The two methods can be further explored and applied to engineering applications where it is critical to compute the turbulent RANS solution in presence of various sources of uncertainties
Uncertainty quantification analysis for simulation of wakes in wind-farms using a stochastic RANS solver, compared with a deep learning approach
No full textQuantification of uncertainties in Reynolds–Averaged Navier–Stokes (RANS) simulations has gained a considerable interest in turbulence modeling. We present two different approaches for the quantification and propagation of model-form and operational uncertainties in context of wind turbine RANS simulations. The first approach is based on a stochastic RANS solver in OpenFOAM using intrusive polynomial chaos method (Parekh and Verstappen, 2023). Here the uncertainties are propagated through a single (large) simulation for the coupled coefficients of the polynomial expansion. The second approach is a surrogate based uncertainty quantification (SBUQ) method. The surrogate model comprises of a 3D U-Net neural network (trained over a single wind turbine) combined with a wake superposition model in order to the prediction of flow field in an array of wind turbines. The above-mentioned approaches are applied for uncertainty quantification analysis in RANS simulations of two turbulent engineering flow problems — (i) a wake past a single wind turbine, and (ii) wake interactions and power losses in an array of wind turbines. The results show that the uncertain RANS solutions from the two approaches are able to reasonably capture the reference high-fidelity solution. We also discuss comparisons between the two approaches including computational cost, applicability, generality etc. The two methods can be further explored and applied to engineering applications where it is critical to compute the turbulent RANS solution in presence of various sources of uncertainties
When Does Eddy Viscosity Damp Subfilter Scales Sufficiently?
No full textLarge eddy simulation (LES) seeks to predict the dynamics of spatially filtered turbulent flows. The very essence is that the LES-solution contains only scales of size ≥Δ, where Δ denotes some user-chosen length scale. This property enables us to perform a LES when it is not feasible to compute the full, turbulent solution of the Navier-Stokes equations. Therefore, in case the large eddy simulation is based on an eddy viscosity model we determine the eddy viscosity such that any scales of size <Δ are dynamically insignificant. In this paper, we address the following two questions: how much eddy diffusion is needed to (a) balance the production of scales of size smaller than Δ; and (b) damp any disturbances having a scale of size smaller than Δ initially. From this we deduce that the eddy viscosity νe has to depend on the invariants q = ½tr(S^2) and r =−⅓tr(S^3) of the (filtered) strain rate tensor S. The simplest model is then given by νe = 3/2(Δ/π)^2|r|/q. This model is successfully tested for a turbulent channel flow (Reτ = 590).
Uncertainty quantification analysis for simulation of wakes in wind-farms using a stochastic RANS solver, compared with a deep learning approach
No full textQuantification of uncertainties in Reynolds–Averaged Navier–Stokes (RANS) simulations has gained a considerable interest in turbulence modeling. We present two different approaches for the quantification and propagation of model-form and operational uncertainties in context of wind turbine RANS simulations. The first approach is based on a stochastic RANS solver in OpenFOAM using intrusive polynomial chaos method (Parekh and Verstappen, 2023). Here the uncertainties are propagated through a single (large) simulation for the coupled coefficients of the polynomial expansion. The second approach is a surrogate based uncertainty quantification (SBUQ) method. The surrogate model comprises of a 3D U-Net neural network (trained over a single wind turbine) combined with a wake superposition model in order to the prediction of flow field in an array of wind turbines. The above-mentioned approaches are applied for uncertainty quantification analysis in RANS simulations of two turbulent engineering flow problems — (i) a wake past a single wind turbine, and (ii) wake interactions and power losses in an array of wind turbines. The results show that the uncertain RANS solutions from the two approaches are able to reasonably capture the reference high-fidelity solution. We also discuss comparisons between the two approaches including computational cost, applicability, generality etc. The two methods can be further explored and applied to engineering applications where it is critical to compute the turbulent RANS solution in presence of various sources of uncertainties
Uncertainty quantification analysis for simulation of wakes in wind-farms using a stochastic RANS solver, compared with a deep learning approach
No full textQuantification of uncertainties in Reynolds–Averaged Navier–Stokes (RANS) simulations has gained a considerable interest in turbulence modeling. We present two different approaches for the quantification and propagation of model-form and operational uncertainties in context of wind turbine RANS simulations. The first approach is based on a stochastic RANS solver in OpenFOAM using intrusive polynomial chaos method (Parekh and Verstappen, 2023). Here the uncertainties are propagated through a single (large) simulation for the coupled coefficients of the polynomial expansion. The second approach is a surrogate based uncertainty quantification (SBUQ) method. The surrogate model comprises of a 3D U-Net neural network (trained over a single wind turbine) combined with a wake superposition model in order to the prediction of flow field in an array of wind turbines. The above-mentioned approaches are applied for uncertainty quantification analysis in RANS simulations of two turbulent engineering flow problems — (i) a wake past a single wind turbine, and (ii) wake interactions and power losses in an array of wind turbines. The results show that the uncertain RANS solutions from the two approaches are able to reasonably capture the reference high-fidelity solution. We also discuss comparisons between the two approaches including computational cost, applicability, generality etc. The two methods can be further explored and applied to engineering applications where it is critical to compute the turbulent RANS solution in presence of various sources of uncertainties
- …
