The present paper tests the capability of a standard Reynolds-Averaged Navier-Stokes (RANS) turbulence model for predicting the turbulent heat transfer in a generic trailing-edge situation with a cutback on the pressure side of the blade. The model investigated uses a gradient-diffusion assumption with a scalar turbulent-diffusivity and constant turbulent Prandtl number. High-fidelity Large-Eddy Simulations (LES) were performed for three blowing ratios to provide reliable target data and the mean velocity and eddy viscosity as input for the heat transfer model testing. Reasonably good agreement between the LES and recent experiments was achieved for mean flow and turbulence statistics. The LES yielded coherent structures which were analysed, in particular with respect to their effect on the turbulent heat transfer. For increasing blowing ratio, the LES replicated an also experimentally observed counter-intuitive decrease of the cooling effectiveness caused by the coherent structures becoming stronger. In contrast, the RANS turbulent heat transfer model failed in predicting this behaviour and yielded significantly too high cooling effectiveness. It is shown that the model cannot predict the strong upstream and wall-directed turbulent heat fluxes caused by large coherent structures, which were found to be responsible for the counter-intuitive decrease of the cooling effectiveness.

@en

In the present study the flow and heat transfer in a tandem cylinder setup are simulated by means of embedded direct numerical simulation (DNS). The influence of wake turbulence on the heat transfer in the stagnation region of the rear cylinder is investigated. The oncoming flow is varied by increasing the distance between the two cylinders, causing a change of the turbulent wake characteristics and the heat transfer. The data of both simulations show good agreement with an existing experimental correlation in the literature. For the small wake generator distance, a clear shift of the maximum heat transfer away from the stagnation line is observed. This shift is less pronounced for the larger distance.

@en

From experimental studies it is well established that free-stream and wake-induced turbulence can increase the heat transfer on curved surfaces such as turbine blades or cylinders even if the mean flow matches corresponding laminar velocity profiles. Here, preliminary results of an investigation using (embedded) Direct Numerical Simulation (DNS) are reported that aim at identifying the responsible physical mechanisms for this phenomenon. The setup follows an existing experiment where a heated cylinder is subjected to the turbulent wake of a smaller cylinder placed upstream. The Reynolds number based on the diameter of the heated cylinder is Re = 48, 000. A key parameter identified in an experimental correlation is the characteristic turbulence level Tu_a that includes a dependency on Re and the mean flow acceleration. For the present study, Tu_a is an order of magnitude higher than those reported in any previous DNS. The obtained Nusselt number is in good agreement with the experimental correlation. More and better resolved data is currently produced for further analysis.

@en

Heat transfer in the stagnation point area of a heated cylinder is investigated using Direct Numerical Simulation (DNS). The heated cylinder is subjected to the turbulent wake of a smaller cylinder placed upstream. Two Reynolds numbers based on the diameter of the heated cylinder of 13,200 and 48,000 are chosen. In accordance with correlations in the literature, an increase in heat transfer compared to fully laminar flow is found for all angles along the front circumferential area of the heated cylinder. However, due to the presence of the wake, the maximum increase is shifted away from the centerline. The characteristic turbulence level and Nusselt number in the present study are an order of magnitude higher than those reported in previous simulations. The DNS results obtained, are in good agreement with an existing experimental correlation. Finally, relevant flow structures and instantaneous temperature fields are visualized.

@en

The turbulent separated flow through an asymmetric diffuser with and without manipulation of incoming turbulence-driven mean secondary vortices (MSVs) from a rectangular duct is investigated by large-eddy simulations. The simulations carried out for two diffuser geometries reveal that by introducing a small amount of mean-flow kinetic energy via the MSVs into the flow, the complex three-dimensional separation behaviour and pressure recovery can be effectively controlled. Manipulated MSVs were found to enhance cross-sectional transport of high-momentum fluid, which determined the location, shape, and size of the separation bubble. The integral effect was a delay or expedition in the onset of separation. This change strongly affected the conversion of mean-flow kinetic energy to pressure, in particular for the front part of the diffuser. In addition, a substantial reduction in total pressure loss could be achieved. The manipulation of the MSVs is an efficient mechanism for performance enhancement in the cases investigated. The results have important implications for both control and statistical modelling of turbulent separated flow in rectangular diffusers.@en

In the present study a tandem cylinder setup is simulated by means of embedded DNS. The influence of wake turbulence on the heat transfer in the stagnation region of the rear cylinder is investigated. The oncoming flow is varied by increasing the distance between the two cylinders, causing a change of the turbulent wake characteristics and the heat transfer. The data of both simulations show good agreement with an existing experimental correlation in the literature. For the small wake generator distance, a clear shift of the maximum heat transfer is observed away from the stagnation line. This shift is less pronounced for the larger distance.@en

Large-eddy simulations (LES) and Reynolds-averaged Navier-Stokes (RANS) calculations of the flow in two asymmetric three-dimensional diffusers were performed. The setup was chosen to match an existing experiment with separation. Both diffusers possess the same expansion ratio but differ in performance. The aim of the present study is to find the least expensive method to reliably and with reasonable accuracy account for the impact of the change in geometry. RANS calculations failed to predict both the extent and location of the separation. In contrast, LES with wall-functions delivered results within the accuracy of the experimental data.@en

The paper presents developments for a segregated approach to the coupling of Reynolds-Averaged Navier-Stokes (RANS) calculations with a zone computed as Large-Eddy Simulation (LES). The mean velocity fields are matched at predefined interfaces and velocity fluctuations of the LES zone are treated according to the type of interface. If the RANS zone is downstream of the LES, fluctuations are allowed to leave the domain by employing a convective boundary condition. If the RANS zone is placed between the LES and a wall, the fluctuations at the interface are scaled to match the statistics predicted by the RANS computation. The proposed method was applied to turbulent channel flow and the flow over periodic hills. It was found that the method delivered an improvement over alternative techniques in the literature while removing the need for calibration constants. For incompressible flows, it is also necessary to prescribe conditions for the pressure or an equivalent variable. Several alternatives were tested. Decoupling of the pressure combined with explicitly enforced mass conservation at the interface yielded the best and, for the hill flow with aggressively placed interface, the only acceptable results.@en

A RANS calculation is connected to an upstream LES via explicit coupling conditions at a pre-defined interface. The role of the interface is to allow for mean flow information to propagate upstream and for fluctuations to leave the LES domain without reflections. To this end, the mean velocity is directly coupled across the domain boundaries whereas the fluctuations at the downstream end of the LES zone are treated using either the so-called enrichment strategy (Quéméré & Sagaut, 2002) or a parameter-free generalization of this method based on a convective condition (von Terzi, Fröhlich & Mary, 2006). For incompressible flows, both techniques require a complementary coupling condition for the pressure or an equivalent variable enforcing continuity. Two distinct techniques are investigated: (i) The instantaneous pressure is computed in a coupled fashion for the union of the LES and RANS domains and (ii) the pressure is completely decoupled and mass conservation across the interface is ensured by an adjustment of the velocities on both sides. The performance of the different methods is scrutinized for turbulent channel flow and the flow over periodic hills. It was found that the convective condition with decoupled pressure fields and an explicit mass flux correction was the most robust technique delivering results of equal or increased quality in comparison to other combinations considered.@en