Reynolds-averaged Navier-Stokes Large Eddy Simulation Research Articles

Active flow control of the vortex rope and pressure pulsations in a swirl generator

ABSTRACTThe vortex rope and pressure pulsations caused by a radial pressure gradient in the conical diffuser of a swirl generator is controlled using continuous slot jets with different momentum fluxes and angles injected from the runner crown. The swirl apparatus is designed to generate flows similar to those in the different operating conditions of a Francis turbine. The study is done with numerical modelling using the hybrid URANS-LES (Unsteady Reynolds-Averaged Navier–Stokes–Large Eddy Simulation) method with the rotor–stator interaction. The comprehensive studies of Javadi and Nilsson [Time-accurate numerical simulations of swirling flow with rotor–stator interaction. Flow, Turbulence and Combustion, Vol. 95, pp. 755–774], and Javadi, Bosioc, Nilsson, Muntean and Susan-Resiga [Experimental and numerical investigation of the precessing helical vortex in a conical diffuser, with rotor–stator interaction. ASME Journal of Fluids Engineering, doi:10.1115/1.4033416] are considered as the bench mark, and the capabilities of the technique is studied in the present work with the validated numerical results presented in those studies. The pressure pulsations caused by the pressure gradient generated by the swirl, present at off-design conditions, are cumbersome for hydropower structures. The investigation shows that the pressure pulsation, velocity fluctuations and the size of the vortex rope decrease when the jet is injected from the runner crown. The flow rate of the jet is less than 3% of the flow rate of the swirl generator. The momentum flux, angle of injection of the jet and the position of the slot are important factors for the effectiveness of the flow control technique.

An hybrid RANS/LES model for simulation of complex turbulent flow

A non-linear eddy viscosity model (NLEVM) and a scalable hybrid Reynolds averaged Navier-Stokes/ large eddy simulation (RANS/LES) strategy are developed to improve the capability of the eddy viscosity model (EVM) to simulate complex flows featuring separations and unsteady motions. To study the performance of the NLEVM, numerical simulations around S809 airfoil are carried out and the results show that the NLEVM performs much better when a large separation occurs. Calculated results of the flow around a triangular cylinder show that the NLEVM can improve the precision of the flow fields to some extents, but the error is still considerable, and the small turbulence structures can not be clearly captured as the EVM. Whereas the scalable hybrid RANS/LES model based on the NLEVM is fairly effective on resolving the turbulent structures and can give more satisfactory predictions of the flow fields.

Numerical Studies of Supersonic Planar Mixing and Turbulent Combustion using a Detached Eddy Simulation (DES) Model

We present a simulation of a hybrid Reynolds-averaged Navier Stokes / Large Eddy Simulation (RANS/LES) based on detached eddy simulation (DES) for a Burrows and Kurkov supersonic planar mixing experiment. The preliminary simulation results are checked in order to validate the numerical computing capability of the current code. Mesh refinement studies are performed to identify the minimum grid size required to accurately capture the flow physics. A detailed investigation of the turbulence/chemistry interaction is carried out for a nine species 19-step hydrogen-air reaction mechanism. In contrast to the instantaneous value, the simulated time-averaged result inside the reactive shear layer underpredicts the maximum rise in H₂O concentration and total temperature relative to the experimental data. The reason for the discrepancy is described in detail. Combustion parameters such as OH mass fraction, flame index, scalar dissipation rate, and mixture fraction are analyzed in order to study the flame structure.

Impact of Swirl Flow on Combustor Liner Heat Transfer and Cooling: A Numerical Investigation With Hybrid Reynolds-Averaged Navier–Stokes Large Eddy Simulation Models

This paper reports the main findings of a numerical investigation aimed at characterizing the flow field and the wall heat transfer resulting from the interaction of a swirling flow provided by lean-burn injectors and a slot cooling system, which generates film cooling in the first part of the combustor liner. In order to overcome some well-known limitations of Reynolds-averaged Navier–Stokes (RANS) approach, e.g., the underestimation of mixing, the simulations were performed with hybrid RANS–large eddy simulation (LES) models, namely, scale-adaptive simulation (SAS)–shear stress transport (SST) and detached eddy simulation (DES)–SST, which are proving to be a viable approach to resolve the main structures of the flow field. The numerical results were compared to experimental data obtained on a nonreactive three-sector planar rig developed in the context of the EU project LEMCOTEC. The analysis of the flow field has highlighted a generally good agreement against particle image velocimetry (PIV) measurements, especially for the SAS–SST model, whereas DES–SST returns some discrepancies in the opening angle of the swirling flow, altering the location of the corner vortex. Also the assessment in terms of Nu/Nu0 distribution confirms the overall accuracy of SAS–SST, where a constant overprediction in the magnitude of the heat transfer is shown by DES–SST, even though potential improvements with mesh refinement are pointed out.

Analysis of pressure perturbation sources on a generic space launcher after-body in supersonic flow using zonal turbulence modeling and dynamic mode decomposition

A sparsity promoting dynamic mode decomposition (DMD) combined with a classical data-based statistical analysis is applied to the turbulent wake of a generic axisymmetric configuration of an Ariane 5-like launcher at Ma∞ = 6.0 computed via a zonal Reynolds-averaged Navier-Stokes/large-eddy simulation (RANS/LES) method. The objective of this work is to gain a better understanding of the wake flow dynamics of the generic launcher by clarification and visualization of initially unknown pressure perturbation sources on its after-body in coherent flow patterns. The investigated wake topology is characterized by a subsonic cavity region around the cylindrical nozzle extension which is formed due to the displacement effect of the afterexpanding jet plume emanating from the rocket nozzle (Mae = 2.52, pe/p∞ = 100) and the shear layer shedding from the main body. The cavity region contains two toroidal counter-rotating large-scale vortices which extensively interact with the turbulent shear layer, jet plume, and rocket walls, leading to the shear layer instability process to be amplified. The induced velocity fluctuations in the wake and the ultimately resulting pressure perturbations on the after-body feature three global characteristic frequency ranges, depending on the streamwise position inside the cavity. The most dominant peaks are detected at SrD r3 = 0.85 ± 0.075 near the nozzle exit, while the lower frequency peaks, in the range of SrD r2 = 0.55 ± 0.05 and SrD r1 = 0.25 ± 0.05, are found to be dominant closer to the rocket’s base. A sparse promoting DMD algorithm is applied to the time-resolved velocity field to clarify the origin of the detected peaks. This analysis extracts three low-frequency spatial modes at SrD = 0.27, 0.56, and 0.85. From the three-dimensional shape of the DMD modes and the reconstructed modulation of the mean flow in time, it is deduced that the detected most dominant peaks of SrD r3 ≈ 0.85 are caused by the radial flapping motion of the shear layer, while the middle-frequency range of SrD r2 ≈ 0.55 is found to be associated with its swinging motion. The less intensive peaks of SrD r1 ≈ 0.25 pronounced on the base wall are caused by the low-frequency longitudinal pumping of the two toroidal large-scale vortices inside the cavity.

Delayed Detached Eddy Simulation of Airfoil Stall Flows Using High-Order Schemes

An advanced hybrid Reynolds-Averaged Navier–Stokes/large eddy simulation (RANS/LES) turbulence model delayed detached eddy simulation (DDES) is conducted in thispaper to investigate the dynamic stall flows over 3D NACA0012 airfoil at 17 deg, 26 deg, 45 deg, and 60 deg angle of attack (AOA). The spatially filtered unsteady 3D Navier–Stokes equations are solved using a fifth-order weighted essentially nonoscillatory (WENO) reconstruction with a low diffusion E-CUSP (LDE) scheme for the inviscid fluxes and a conservative fourth-order central differencing for the viscous terms. An implicit second-order time marching scheme with dual time stepping is employed to achieve high stability and convergency rate. A 3D flat plate is validated for the DDES model. For quantitative prediction of lift and drag of the stalled NACA0012 airfoil flows, the detached eddy simulation (DES) and DDES achieve much more accurate results than the Unsteady Reynolds-Averaged Navier–Stokes (URANS) simulation. In addition to the quantitative difference, the DES/DDES and URANS also obtain qualitatively very different unsteady stalled flows of NACA0012 airfoil with different vortical structures and frequencies. This may bring a significantly different prediction if those methods are used for fluid–structural interaction. For comparison purpose, a third-order WENO scheme with a second-order central differencing is also employed for the DDES stalled NACA0012 airfoil flows. Both the third- and fifth-order WENO schemes predict the stalled flow similarly for lift and drag at AOA less than 45 deg, while at AOA of 60 deg, the fifth-order WENO scheme shows better agreement with the experiment than the third-order WENO scheme. The high-order scheme of WENO 5 also resolves more small scales of flow structures than the second-order scheme. The prediction of the stalled airfoil flow using DDES with both the high-order scheme and second-order scheme is overall significantly more accurate than the URANS simulation.

A new formulation of scale-adaptive simulation approach to predict complex wall-bounded shear flows

A new formulation of scale-adaptive simulation (SAS) approach for complex wall-bounded shear flows is presented. This approach makes use of a unique modelling representation and requires moderate computational costs. Based on the Rotta original transport equation for turbulence integral length scale, the suggested model is able to resolve unsteady turbulent structures with sufficient spatial resolution. Similar to the classical SAS-turbulence model (SST–SAS) proposed by Menter et al. [5] that represents an improvement of the proven turbulence model for unsteady calculations, the new model appears advantageous where classical LES (large eddy simulation) or hybrid LES–RANS models are too expensive. However, in its zonal formulation, the new model provides accurate predictions in (1) so-called stable flows (e.g. channel flow) in which the classical SAS () model will not be able to switch from Reynolds-averaged Navier–Stokes (RANS) to scale-resolving simulations without an explicit introduction of synthetic turbulence, (2) flows with relatively weak instabilities also challenging for the classical SAS-type models (e.g. swirling flow with sudden expansion). Such improvements result from the appropriately treating additional source terms in modified transport equation of dissipation rate. For demonstration, comparisons with experimental data and direct numerical simulation are provided.

Numerical study on supersonic mixing and combustion with hydrogen injection upstream of a cavity flameholder

Characteristics of supersonic mixing and combustion with hydrogen injection upstream of a cavity flameholder are investigated numerically using hybrid RANS/LES (Reynolds-Averaged Navier–Stokes/Large-Eddy Simulation) method. Two types of inflow boundary layer are considered. One is a laminar-like boundary layer with inflow thickness of $$\delta_{\inf } = 0.0$$ and the other is a turbulent boundary layer with inflow thickness of $$\delta_{\inf } = 2.5\,{\text{mm}}$$ . The hybrid RANS/LES method acts as a DES (Detached Eddy Simulation) model for the laminar-like inflow condition and a wall-modeled LES for the turbulent inflow condition where the recycling/rescaling method is adopted. Although the turbulent inflow seems to have just minor influences on the supersonic cavity flow without fuel injection, its effects on the mixing and combustion processes are great. It is found that the unsteady turbulent structures in upstream incoming boundary layer interact with the injection jet, resulting in fluctuations of the upstream recirculation region and bow shock, and induce quick dispersion of the hydrogen fuel jet, which enhances the mixing as well as subsequent combustion.

Large eddy simulation based studies of jet–cavity interactions in a supersonic flow

Interactions of a cavity flameholder with an upstream injected jet in a Ma 2.52 supersonic flow are investigated numerically. A hybrid RANS/LES (Reynolds-Averaged Navier–Stokes/Large Eddy Simulation) method acting as wall-modeled LES is adopted, for which the recycling/rescaling method is introduced to treat the unsteady turbulent inflow. Patterns of the fluid entrainment into the cavity and escape from the cavity are identified using a scalar-tracing method. It is found that the jet–cavity interactions remarkably enhanced the mass exchange between the fluids in and out of the cavity, resulting in reduced residence time of the cavity fluids. Increasing the distance between the fuel injection and the cavity leading edge tends to attenuate the jet–cavity interactions, leading to weaker mass exchange. Raising the injection pressure appears to enhance the jet–cavity interactions, resulting in a shorter residence time of the cavity fluids. Moreover, the mass decay processes for the fuel and air within the cavity are basically the same while the entrainment processes for the fuel and air into the cavity seem quite different.

Hybrid Reynolds-averaged Navier-Stokes/large-eddy simulation of jet mixing in a supersonic crossflow

A sonic under-expanded transverse jet injection into a Ma 1.6 supersonic crossflow is investigated numerically using our hybrid RANS/LES (Reynolds-averaged Navier-Stokes/large eddy simulation) method. First, a calculation is carried out to validate the code, where both the instantaneous and statistical results show good agreement with the existing experimental data. Then the jet-mixing characteristics are analyzed. It is observed that the large-scale vortex on the windward portion of the jet boundary is formed mainly by the intermittent impingement of the incoming high-speed fluid on the relatively low-speed region of the upstream jet boundary, where the interaction between the upstream separated region and the jet supplies a favorable pressure condition for the sustaining acceleration of the high-speed fluid during the vortex forming, associated with which the incoming fluid is entrained into the jet boundary and large-scale mixing occurs. Meanwhile, the secondary recirculation zone between the upstream separated region and the jet is observed to develop evidently during the vortex forming, inducing the entrainment of jet fluid into the upstream separated region. Moreover, effects of the incoming boundary layer on the jet mixing are addressed.

Hybrid RANS/LES of plane jets impinging on a flat plate at small nozzle-plate distances

A k-omega BASED HYBRID RANS/LES (Reynolds-averaged Navier-Stokes/large eddy simulation) model is tested for simulation of plane impinging jets at various nozzle-plate distances (H/B), where H is the distance and B is the slot's width) and various Reynolds numbers (based on the slot's width and the velocity in the symmetry plane). The studied combinations are H/B = 2 for Re = 10000, H/B = 4 for Re = 18 000 and H/B = 9.2 for Re = 20000. The focus is on small distance of the nozzle exit to the plate. In LES mode, the hybrid RANS/LES model uses two definitions of the local grid size, one based on the maximum distance between the cell faces in the destruction term of the turbulent kinetic energy equation and one based on the cube root of the cell volume in the eddy-viscosity formula. This allows accounting for flow inhomogeneity on anisotropic grids. In RANS mode, the hybrid model turns into the newest version of the k-omega model by Wilcox.

Large-Eddy/Reynolds-averaged Navier–Stokes simulation of combustion oscillations in a cavity-based supersonic combustor

Combustion oscillations in a supersonic combustor with hydrogen injection upstream of a cavity flameholder are investigated numerically using a hybrid RANS/LES (Reynolds-Averaged Navier–Stokes/Large-Eddy Simulation) method acting as a wall-modeled LES. A turbulent boundary layer with thickness of δinf = 2.5 mm is considered and a recycling/rescaling method is adopted to treat the unsteady inflow. The results show that combustion oscillations can mainly be attributed to two mechanisms. One is the unsteady flame spreading from the cavity shear layer to the main stream, which is greatly influenced by the interaction of the jet-with-cavity shear layer. This mechanism leads to relatively low-frequency oscillations that correspond to the cavity-shear layer oscillations. The other is the auto-ignition of the combustible fluid packets formed around the fuel jet accompanied by the generation of the hairpin-like vortices, which leads to relatively high-frequency oscillations that correspond to the jet instabilities.

Characteristics of Oscillations in Supersonic Open Cavity Flows

Characteristics of oscillations in supersonic open cavity flows are investigated numerically using hybrid RANS/LES (Reynolds-Averaged Navier-Stokes/Large Eddy Simulation) method. The oscillation regimes and feedback mechanisms for the supersonic cavity flows are identified and analyzed. The calculation captures a mixed shear-layer/wake oscillation mode in the flow of Ma = 1.75, where these two modes occur alternately. The shear-layer mode and wake mode are driven by vortex convection-acoustic feedback and absolute instability, respectively. In particular, the results indicate that the feedback-acoustic-wave in the shear-layer mode is probably generated by the reflection of the downstream-traveling pressure wave, associated with the shed vortex in the shear layer, on the aft wall. The cavity flow of Ma = 2.52 is then simulated to see the influence of Mach number. It is found that the increase of Mach number may decrease the amplitude of the fluctuations in the shear layer, inhibiting the transition to wake mode. Furthermore, the influence of upstream injection is also studied, where the results show that the injection only weakens the oscillations and faintly shifts the resonant frequencies.

Unstructured overset incompressible computational fluid dynamics for unsteady wind turbine simulations

ABSTRACTOverset computational fluid dynamics (CFD) methods are the most sophisticated methods currently available to predict the unsteady motion of wind turbine blades without the need for additional simplifications or restrictions on the turbine operational conditions. An unstructured implementation of the governing equations of motion permits rapid modeling of the salient components, such as nacelles, towers and other localized obstructions of interest. A time‐accurate incompressible formulation accelerates the convergence of the solution, in addition to eliminating the need for low‐Mach number preconditioning, which can be problematic and computationally expensive for time‐accurate simulations. The use of a hybrid Reynolds‐averaged Navier–Stokes/large eddy simulation (RANS/LES) turbulence method is observed to improve the prediction and extent of separation, as well as integrated performance variables for stalled rotors under fully turbulent conditions. Copyright © 2012 John Wiley & Sons, Ltd.

Hybrid RANS/LES of round impinging jets

Fluid flow and heat transfer characteristics are presented for simulations of round impinging jets at two nozzle-plate distances H/D = 2 and 10 (D is the nozzle exit diameter) and two Reynolds numbers Re = 5000 and 70,000 with hybrid RANS/LES (Reynolds-averaged Navier-Stokes/Large Eddy Simulation), dynamic Smagorinsky LES and RANS k−ω models. Three k−ω based hybrid RANS/LES models are analyzed. With the hybrid RANS/LES models, improved heat transfer results are obtained, when compared to RANS, in the impact region and in the developing wall-jet region. For accurate predictions at low nozzle-plate distance, it is necessary to sufficiently resolve the formation and development of the near-wall vortices in the jet impingement region. At high nozzle-plate distance, it is important to capture the evolution and breakup of the unsteady vortices in the shear layer of the jet, so that realistic mean and fluctuating velocity profiles are obtained in the impact jet region.

A dynamic pressure-sink method for improving large eddy simulation and hybrid Reynolds-averaged Navier–Stokes/large eddy simulation of wall-bounded flows

A dynamic method based on introducing pressure sinks at the wall boundary is proposed to improve hybrid Reynolds-averaged Navier–Stokes/large eddy simulation (RANS/LES) and LES on under-resolved meshes of wall-bounded flows. The basic spirit is to enhance the momentum transfer in the boundary layer from the outer layer to the inner layer by introducing a small pressure gradient in the wall-normal direction, so that the skin friction, usually underestimated in hybrid RANS/LES and LES on under-resolved meshes, can recover to the correct level. The dynamic pressure-sink method, applied to LES and hybrid RANS/LES, is first numerically tested in simulating a supersonic flat-plate boundary flow, for which the turbulent inflow condition is treated by the recycling/rescaling approach. The results show obvious improvement as compared to the baseline LES and hybrid RANS/LES results on the same mesh. The method is then applied to LES of a supersonic flow over a compression–expansion ramp, where improved results are also achieved.

Hybrid RANS/LES computation of plane impinging jet flow

Flow characteristics are presented of simulation results of plane impinging jets at high nozzle-plate distances, with two k - ω based hybrid RANS/LES (Reynolds Averaged Navier–Stokes/Large–Eddy Simulation) models and a k - ω RANS model. The first hybrid RANS/LES model is obtained by substitution of the turbulent length scale by the local grid size in the destruction term of the turbulent kinetic energy equation and in the definition of the eddy-viscosity. The second hybrid model is obtained by a latency factor in the definition of the eddy-viscosity. The RANS model overpredicts the length of the jet core region, caused by too weak turbulent mixing in the shear layers of the jet. This results in erroneous near-wall shear stress along the impingement plate. The hybrid RANS/LES models overcome the deficiency of the RANS model. Further, the hybrid models represent the flow with much more detail. For instance, the Gortler vortices are well reproduced in the stagnation flow region by the hybrid RANS/LES models.

Analysis of filtered Navier–Stokes equation for hybrid RANS/LES simulation

Hybrid Reynolds-averaged Navier–Stokes/large eddy simulation (RANS/LES) is expected to accurately predict wall-bounded turbulent flows at high Reynolds numbers. It is known that extra terms due to the noncommutivity between the hybrid filter and the spatial derivative appear in the hybrid-filtered equations. In this paper the filtered Navier–Stokes equation is investigated using direct numerical simulation data of turbulent channel flow. In particular, the transport equations for the resolved and modeled energies are evaluated to examine the contribution of the extra terms. The RANS and LES regions are located by setting the width of the hybrid filter varying from the grid spacing to the channel half width. In the first case the RANS region is located near the wall for wall modeling in LES. In the second case the RANS and LES regions are located upstream and downstream, respectively. The extra terms in the Navier–Stokes and continuity equations show fairly large values in the RANS/LES interface region in both cases. The extra terms for the convection and the turbulent diffusion represent the energy transfer from the modeled part to the resolved part. This energy transfer suggests that conventional hybrid simulations neglecting the extra terms may underpredict velocity fluctuations in the interface region, causing an insufficient momentum transport and velocity mismatch in channel flow. These results show that the effect of the extra terms should be taken into account in the filtered Navier–Stokes equation as well as in the turbulent-energy equation in hybrid simulations.

Computations of Flow Past a Circular Cylinder using a Continuous-Turbulence Model

modeling is calculated with Reynolds-averaged Navier–Stokes equations using a function that is uniquely determined by the ratio of the resolved small-scale length to the integral length scale. The unified treatment of Reynolds-averaged Navier–Stokes large eddy simulation could be achieved with this continuous model by controlling the coarseness of the mesh. The continuous Reynolds-averaged Navier–Stokes large eddy-simulation model is initially tested with flow past a circular cylinder at the Reynolds number Re 3900. Both constant integral length scales and varied integral length scales based on the mixing length are used in the simulations. The mesh dependence is also studied with both coarse and fine meshes. The comparisons are made and analyzed between the results with different integral length scales and different mesh sizes. These results are also compared with the experimental results and Reynolds-averaged Navier–Stokes results. The results demonstrate that with the continuous modeling, the large eddy-simulation-like simulation can be achieved by solving Reynolds-averaged Navier–Stokes equations alone. The spatially varied integral length scale is necessary to capture the more accurate turbulence quantities for anisotropic wall-bounded turbulent flow.

A modified blending function for zonal hybrid Reynolds-averaged Navier—Stokes/large-eddy simulation methodology

A modified blending function for zonal hybrid Reynolds averaged Navier—Stokes/large eddy simulation (RANS/LES) methodology was developed using an empirical analogy from Menter k—ω shear stress transport (SST) turbulent model (Menter, 1994) to predict complex turbulent flows. Tests of slot jet in supersonic flow and supersonic flow over compression—expansion ramp was conducted and prediction of separations was well improved when certain model constant was forced on the traditional blending function (Baurle et al., 2003). Analysis based on calculations of flat plate boundary layer demonstrated that an efficient empirical constant could be used in blending function and boundary layer could be well calculated without heavy contamination of RANS on wake region. Validation of the modified zonal hybrid RANS/LES approach for slot jet in supersonic flow, supersonic flow over compression—expansion ramp, supersonic flow over backward facing step, and supersonic cavity flow was conducted. The simulated results showed that the modified blending function performs well on complex turbulent flows. Deficiencies of traditional hybrid zonal RANS/LES method in over-prediction of separations associated with adverse pressure gradient flows were favourably improved.