Wall-modeled Large Eddy Simulation Research Articles

Near-wall numerical coherent structures and turbulence generation in wall-modelled large-eddy simulation

Near-wall turbulence structures and generation in the wall-modelled large-eddy simulation (WMLES) are revealed. To elucidate the turbulence structures driving a near-wall turbulence generation in the WMLES, flat-plate turbulent boundary-layer flows calculated by the WMLES and direct numerical simulation (DNS) are closely investigated. A conditional-averaging technique is applied to the instantaneous flow fields and the near-wall statistical structures of the ejection and sweep pairs, which produce the turbulence, are revealed to exist even in the WMLES although the structures are non-physically elongated compared with those obtained by the DNS. Since the near-wall turbulence structures in the WMLES are revealed not to be disordered, but to be coherent structures with low- and high-speed fluids alternating in the spanwise direction, it is suggested that the near-wall turbulence generation in the WMLES is explained by the numerically elongated coherent structures. Furthermore, the Reynolds number effects of wall-bounded turbulent flows, i.e. the appearance of the outer peak in the energy spectrum of the streamwise velocity fluctuations at increasing Reynolds numbers, is found not to be reproduced by the WMLES, and the origin of the outer peak is discussed in association with the inner–outer-layer interactions. The near-wall turbulence structures in the WMLES could depend heavily on the computational grids and the numerical methods. Therefore, additional cases varying the grid resolutions and the numerical methods (numerical schemes and sub-grid-scale models) are also conducted to confirm the consistency of the present conclusions.

Wall-resolved versus wall-modeled LES of the flow field and surface forced convective heat transfer for a low-rise building

Large eddy simulation (LES) is widely used to investigate the aerodynamics and convective heat transfer (CHT) at the surfaces of sharp-edged bluff bodies for a wide range of Reynolds (Re) numbers. Due to the heavy computational costs associated with implicit filtering in LES at high Reynolds number flows (Re ≥ 105), wall-modeled (WM) rather than wall-resolved (WR) LES is often adopted. However, the performance of LES-WM for such applications has not yet been systematically investigated. Therefore, this study evaluates the performance of LES-WM and LES-WR for the flow and thermal field at the facades of a low-rise building immersed in an atmospheric boundary layer. Four grids are constructed for LES-WM, each employing different resolution at the building surfaces reaching maximum non-dimensional wall distance y+ = 43, 57, 70, and 95. In addition, the performance of two wall functions, namely the Werner and Wengle and the enhanced wall function is investigated. The results show that the use of LES-WM can result in significant deviations in the predicted near-facade flow pattern and the surface convective heat transfer coefficient (CHTC). Grid resolution significantly impacts the CHTC results and deviations go up to 88% (at the base of the windward facade). Considerable deviations among the employed wall functions are apparent only on the finest grid. In this case, the implementation of the enhanced wall function indicates better performance compared to the non-blended law of the wall (combined with the Werner and Wengle) for CHTC in the regions of the leeward facade where the flow remains attached to the wall. The deviation of the enhanced wall function for surface-averaged CHTC is found to be 10.8% against the wall-resolved LES results, while for the non-blended law of the wall this is 19.2%.

Improved Delayed Detached Eddy Simulation with Reynolds-stress background modelling

A novel variant of Improved Delayed Detached-Eddy Simulation based on a differential Reynolds-stress background model is presented. The approach aims to combine the advantages of anisotropy-resolving Reynolds-stress closures in the modelled RANS regions with consistent LES and wall-modelled LES behaviour in the resolved flow regions. In computations of decaying isotropic turbulence with a low-dissipative flow solver it is shown that a straightforward hybridised Reynolds-stress model provides insufficient turbulent dissipation as sub-grid closure in the LES regions and is therefore locally replaced by scalar viscosity modelling. Simulations of periodic channel flows at different Reynolds numbers and grid resolutions are used to calibrate and validate the wall-modelled LES branch of the new model. A final application in embedded wall-modelled LES of a flat-plate boundary layer is widely consistent with results using the SST-RANS background model, but shows some deviations from the Coles–Fernholz skin-friction correlation. In this regard, initial sensitivity studies indicate possible adverse effects due to the synthetic-turbulence approach used in these simulations.

On hybrid RANS-LES of transition in a separated boundary layer

The numerical simulation of laminar separation followed by transition and re-attachment of a turbulent boundary layer has been investigated with particular focus on the prediction of the transition process. The turbulence has been modeled using hybrid RANS-LES models, namely the variants DDES and IDDES. For reference, a computation with nearly DNS-like resolution is featured. Using differently fine meshes, the prediction of the shear layer vortex breakup process into turbulence has been observed in terms of turbulence intensities and turbulence length scales. It is shown how these properties develop during the transition process going along with flow re-attachment. Slight differences arise between DDES and IDDES, the latter one operating in a mode of wall-modeled LES. To describe the impact on the aerodynamic properties, the probability of flow separation along the transition process is presented and the correlation between length scales, turbulence intensity and flow separation is discussed. Conclusions are drawn on the applicability and qualification of the turbulence models as well as on requirements concerning mesh resolution and discretization schemes.

Large Eddy Simulation of the ACAT1 fan stage for broadband noise prediction

A numerical study is carried out to assess the aerodynamic and noise radiation specific features of the ACAT1 fan stage, which was tested in the framework of the European project TurbonoiseBB. This study deals with the prediction of the broadband noise radiated by the fan stage at approach conditions, and particularly focuses on the rotor–stator interaction (RSI) noise, which results from the impact of the fan wakes onto the outlet guide vanes (OGV). To this end, two wall-modeled Large Eddy Simulations (LES) with different levels of mesh refinement have been performed on the full fan-OGV stage. The finer mesh significantly improved the aerodynamic predictions in terms of mean flow profiles, but also in terms of RMS profiles and velocity spectra, for which significant disparities are highlighted for the coarser LES. Only slight disparities regarding the turbulent content of the flow remain for the finer LES. The typical flow features at approach conditions are examined and the noise sources on both the stator and the rotor are analyzed, revealing the presence of additional broadband noise sources. Noise predictions are performed by using the free-field Ffowcs Williams and Hawkings (FW-H) analogy, and Goldstein’s in-duct acoustic analogy informed with the pressure fluctuations recorded on the stator and rotor surfaces. FW-H predictions of the stator noise, though well recovering the shape of the experimental spectra, overestimate the radiated noise especially upstream of the fan. Using the finer simulation data substantially reduces the gap with the measurements. Accounting for the duct effect on the sound propagation, further improves the predictions in terms of absolute levels, leading to a good agreement between the finer simulation and the noise measurements. The breakdown of the different stator noise source contributions, in conjunction with rotor broadband noise predictions made it possible to confirm the presence of additional significant broadband noise sources, questioning the common assumption considering the RSI sources as the only dominant fan noise mechanism at approach conditions.

Survey of machine-learning wall models for large-eddy simulation

Survey of machine-learning wall models for large-eddy simulation

Wall-modeled large-eddy simulation integrated with synthetic turbulence generator for multiple-relaxation-time lattice Boltzmann method

The synthetic turbulence generator (STG) lies at the interface of the Reynolds averaged Navier–Stokes (RANS) simulation and large-eddy simulation (LES). This paper presents an STG for the multiple-relaxation-time lattice Boltzmann method (LBM) framework at high friction Reynolds numbers, with consideration of near-wall modeling. The Reichardt wall law, in combination with a force-based method, is used to model the near-wall field. The STG wall-modeled LES results are compared with turbulent channel flow simulations at Reτ=1000,2000,5200 at different resolutions. The results demonstrate good agreement with direct numerical simulation, with the adaptation length of 6–8 boundary layer thickness. This method has a wide range of potentials for hybrid RANS/LES-LBM related applications at high friction Reynolds numbers.

Assessment of subgrid-scale models in wall-modeled large-eddy simulations of turbulent channel flows

Assessment of subgrid-scale models in wall-modeled large-eddy simulations of turbulent channel flows

Grey area in embedded wall-modelled LES on a transonic nacelle-aircraft configuration

A scale-resolving hybrid RANS–LES technique is applied to an aircraft-nacelle configuration under transonic flow conditions using the unstructured, compressible TAU solver. In this regard, a wall-modelled LES methodology is locally applied to the nacelle lower surface to examine shock-induced separation. To circumvent the grey-area issue of delayed turbulence onset, a Synthetic Turbulence Generator (STG) is used at the RANS–LES interface. Prior to the actual examinations, fundamental features of the simulation technique are validated by simulations of decaying isotropic turbulence as well as flat plate flows. For the aircraft-nacelle configuration at a Reynolds number of 3.3 million, a sophisticated mesh with 420 million points was designed which refines 32 % of the outer casing surface of the nacelle. The results show a development of a well-resolved turbulent boundary layer with a broad spectrum of turbulent scales which demonstrates the applicability of the mesh and method for aircraft configurations. Furthermore, the necessity of a low-dissipation low-dispersion scheme is demonstrated. However, a noticeable drop of the surface skin friction downstream of the STG motivates further research on the impact of the interface modelling on the shock–boundary layer interaction.

Machine learning building-block-flow wall model for large-eddy simulation

A wall model for large-eddy simulation (LES) is proposed by devising the flow as a combination of building blocks. The core assumption of the model is that a finite set of simple canonical flows contains the essential physics to predict the wall shear stress in more complex scenarios. The model is constructed to predict zero/favourable/adverse mean pressure gradient wall turbulence, separation, statistically unsteady turbulence with mean flow three-dimensionality, and laminar flow. The approach is implemented using two types of artificial neural networks: a classifier, which identifies the contribution of each building block in the flow, and a predictor, which estimates the wall shear stress via a combination of the building-block flows. The training data are obtained directly from wall-modelled LES (WMLES) optimised to reproduce the correct mean quantities. This approach guarantees the consistency of the training data with the numerical discretisation and the gridding strategy of the flow solver. The output of the model is accompanied by a confidence score in the prediction that aids the detection of regions where the model underperforms. The model is validated in canonical flows (e.g. laminar/turbulent boundary layers, turbulent channels, turbulent Poiseuille–Couette flow, turbulent pipe) and two realistic aircraft configurations: the NASA Common Research Model High-lift and NASA Juncture Flow experiment. It is shown that the building-block-flow wall model outperforms (or matches) the predictions by an equilibrium wall model. It is also concluded that further improvements in WMLES should incorporate advances in subgrid-scale modelling to minimise error propagation to the wall model.

Effects of discharge-to-submergence ratio on evolution of air-core vortex

An air-core vortex is generated in a tidal power station when the discharge increases or the submergence decreases at the intake. To better understand the complex evolution of the air-core vortex, we conduct high-fidelity coupled level-set and volume-of-fluid and wall-modeled large-eddy simulations for various ratios of discharge and submergence. We characterize the air-core vortex evolution process into two stages: the inception stage and the propagation stage. Our results show that the streamwise and spanwise velocity fluctuations exhibit similar distributions and temporal evolutions and are responsible for generating asymmetrical flow transitioned from a completely symmetrical pattern in the inception stage. The turbulence quadrant decomposition analysis in the confluence region quantifies the velocity fluctuation signal and Reynolds shear stress distribution. Under a specific Reynolds shear stress distribution, eddies converge and evolve into a main vortex. We introduce a vortex burst index to quantitatively characterize the contribution of the Reynolds shear stress to the main vortex evolution. The vortex burst index is positively correlated with the discharge-to-submergence ratio. In the propagation stage, the enstrophy stretching mechanism dominates the air-vortex evolution. The tilting terms nearly cancel each other because of the similar tangential velocity distribution in the axial direction. The results show that the increase in the discharge-to-submergence ratio accelerates the formation of the air-core vortex, owing to the increase in the vertical velocity gradient and the enhancement of the stretching effect.

Hydrodynamics of a swimming batoid fish at Reynolds numbers up to 148 000

Flow around a tethered model of a swimming batoid fish is studied by using the wall-modelled large-eddy simulation in conjunction with the immersed boundary method. A Reynolds number ( $Re$ ) up to 148 000 is chosen, and it is comparable to that of a medium-sized aquatic animal in cruising swimming state. At such a high $Re$ , we provide, to the best of our knowledge, the first evidence of hairpin vortical (HV) structures near the body surface using three-dimensional high-fidelity flow field data. It is observed that such small-scale vortical structures are mainly formed through two mechanisms: the leading-edge vortex (LEV)–secondary filament–HV and LEV–HV transformations in different regions. The HVs create strong fluctuations in the pressure distribution and frequency spectrum. Simulations are also conducted at $Re=1480$ and 14 800 to reveal the effect of Reynolds number. Variations of the flow separation behaviour and local pressure with $Re$ are presented. Our results indicate that low- $Re$ simulations are meaningful when the focus is on the force variation tendency, whereas high- $Re$ simulations are needed when concerning flow fluctuations and turbulence mechanisms.

A wall model learned from the periodic hill data and the law of the wall

Toward data-driven wall-modeled large-eddy simulations of different wall-bounded turbulent flows, a wall model is learned in this work using the wall-resolved large-eddy simulation (WRLES) data of the flow over periodic hills (PH) and the law of the wall (LoW). The feedforward neural network (FNN) is employed to construct the model. The obtained FNN_PH-LoW model is successfully tested using the direct numerical simulation data of turbulent channel flows and the WRLES data of PH cases, and applied to turbulent channel flows for a wide range of Reynolds numbers.

Performance assessment of wall-modeled large-eddy simulation for modeling aeolian two-phase flow

Wall-modeled large-eddy simulation (WMLES) is an effective way to improve the computational efficiency of two-phase flow. Based on a wall-resolved large-eddy simulation (WRLES) of aeolian two-phase flow at sand-free friction Reynolds number of Reτ=4200, the performance of different subgrid stress (SGS) and wall models in modeling aeolian two-phase flow are checked. Three classical eddy-viscosity-type SGS models are employed, i.e., Smagorinsky (SMAG) model, dynamic Smagorinsky (DYNA) model and scale-dependent dynamic (SDDY) model. The involved wall models are two equilibrium wall models and a non-equilibrium wall model. By comparing the WMLES and WRLES of sand-laden flow, it is found that wall models have less influence on particle velocities for the same SGS model. On the contrary, the particle volume fraction and mass flux strongly depend on the SGS and wall models. Although, the performance of WMLES with different models varies from case to case, the combination of dynamic Smagorinsky model and integral model performs the best in this case. By analyzing the particle and fluid stress, we suggest that the influence of particles on LES models is in urgent need.

Data-driven wall modeling for turbulent separated flows

The large-eddy simulation of wall-bounded turbulent flows at high Reynolds numbers is made more efficient by the use of wall models that predict the wall shear stress, allowing coarser cell sizes at the wall. In this paper, a data-driven approach for the modeling of the wall shear stress is examined using filtered high-fidelity numerical data from two fully developed turbulent channel flows and two turbulent flows with separated regions: a three-dimensional diffuser and a backward-facing step. The model is a multilayer perceptron based on the flow information in the vicinity, given by the distance to the wall and the velocity components at a given number of grid points above the wall. The model is Mach number equivariant at the quasi-incompressible limit, Galilean invariant, statistically rotational invariant and can extrapolate to flow conditions unseen in the training dataset. The relevance of the machine-learning procedure is verified a priori using the filtered numerical data and a posteriori by performing wall-modeled large-eddy simulations implementing the model. The results show that the model is able to leverage the local spatial information to discriminate developed wall turbulence and separated regions in flow configurations not included in the training dataset.

A foundational step towards understanding and improving the transition between URANS and LES in hybrid URANS-LES methodology

Appropriate transition between URANS and LES plays an integral role in the performance of hybrid URANS-LES techniques. However, dynamics of the transition region (also known as the Gray Area) have not yet been well understood, despite several efforts to mitigate negative impacts of inappropriate modeling of this region such as deviation from law-of-the-wall (also referred to as log-layer mismatch (LLM)) in turbulent attached boundary layers. The current paper centers on Scale Adaptive Simulation (SAS) hybrid URANS-LES methods and aims to address some shortcomings of the SAS methodology while providing new insights into the dynamics of the Gray Area. More specifically, necessary requirements, including the length scale, grid design and underlying RANS model to appropriately model essential dynamics of the Gray Area and, thus, mitigating the log-layer mismatch, are presented, which could be used for broader framework of hybrid URANS/LES and probably wall-modeled LES approaches.SAS methods traditionally have difficulty transitioning from URANS to LES without a sufficiently unstable base flow. Here we introduce an improved triggering mechanism based on departure-from-equilibrium dynamics formulated for the SAS method, primarily to enable the model to transition from URANS to scale resolving mode in attached/mildly separated flows, a well-known deficiency of the classical SAS formulations. The improved triggering mechanism considers deviations from equilibrium through a vortex stretching term in the dynamical equation for turbulent dissipation rate. The present study demonstrates that the length scale obtained for the gray area using straightforward blending of URANS and LES regions results in inaccurate transition dynamics. Therefore, we introduce a transport equation for the length scale of energy-containing eddies that allows an appropriate transition of length scale from URANS to LES mode. The modified framework uses the k−ɛ−ζ−f model as the underlying RANS model and is shown to trigger a transition from URANS to LES in stable attached stationary turbulent channel flow and mitigate the log-layer mismatch problem to a great extent. When applied to flows that feature boundary layer separation (flow over a periodic hill and flow over the wall-mounted hump), the current hybrid framework delivers results in agreement with existing experimental data and comparable to high fidelity LES calculations.

Wall-modelled large-eddy simulation of three-dimensional turbulent boundary layer in a bent square duct

We conduct wall-modelled large-eddy simulation (WMLES) of a pressure-driven three-dimensional turbulent boundary layer developing on the floor of a bent square duct to investigate the predictive capability of three widely used wall models, namely, a simple equilibrium stress model, an integral non-equilibrium model, and a partial differential equation (PDE) non-equilibrium model. The numerical results are compared with the experiment of Schwarz & Bradshaw (J. Fluid Mech., vol. 272, 1994, pp. 183–210). While the wall-stress magnitudes predicted by the three wall models are comparable, the PDE non-equilibrium wall model produces a substantially more accurate prediction of the wall-stress direction, followed by the integral non-equilibrium wall model. The wall-stress direction from the wall models is shown to have separable contributions from the equilibrium stress part and the integrated non-equilibrium effects, where how the latter is modelled differs among the wall models. The triangular plot of the wall-model solution reveals different capabilities of the wall models in representing variation of flow direction along the wall-normal direction. In contrast, the outer LES solution is unaffected by the type of wall model used, resulting in nearly identical predictions of the mean and turbulent statistics in the outer region for all the wall models. This is explained by the vorticity dynamics and the inviscid skewing mechanism of generating the mean three-dimensionality. Finally, the LES solution in the outer layer is used to study the anisotropy of turbulence. In contrast to the canonical two-dimensional wall turbulence, the Reynolds stress anisotropy exhibits strong non-monotonic behaviour with increasing wall distance.

Aeroacoustic analysis of the tip-leakage flow of an ultrahigh bypass ratio fan stage

A detailed aeroacoustic analysis of the flow induced by the clearance between the fan tip and the shroud is performed in a scale-model fan stage of an ultrahigh bypass ratio turbofan engine, which was designed to operate at transonic regimes. A wall-modeled large eddy simulation is performed at approach condition, which corresponds to a fully subsonic operating point. The contributions of the tip-gap noise to the total fan noise are investigated using the Ffowcs Williams and Hawkings analogy. The surface is split into two parts: the tip region and the rest of the blade in order to analyze the acoustic contributions of these two regions separately. It is shown that the tip-gap region generates a significant noise component above 2 kHz, which corresponds to approximately 1.2 times the blade passing frequency. Two separate tip-leakage vortices are identified in the vicinity of the fan tip. The dominant noise sources in the tip-gap region are observed at the trailing edge of the fan blade. The wall pressure spectra in the tip-gap region and the coherence of pressure fluctuations between monitor points at different positions show an acoustic contribution of the tip-leakage flow at two different frequency ranges. The first range corresponds to medium frequencies between 2 and 9 kHz, and the second range corresponds to high frequencies between 10 and 25 kHz. The analysis of dynamic mode tracking, fluctuating pressure and velocity spectra, and instantaneous flow fields relates specific vortices in the tip-gap flow to their spectral signature and paves the way for further analytical modeling of tip-gap noise sources.

Wall-modeled large eddy simulation for the flows around an axisymmetric body of revolution

Wall-modeled large eddy simulation for the flows around an axisymmetric body of revolution

FFVHC-ACE: Fully Automated Cartesian-Grid-Based Solver for Compressible Large-Eddy Simulation

This study presents a fully automated Cartesian-grid-based compressible flow solver, named FrontFlow/Violet Hierarchical Cartesian for Aeronautics based on Compressible-flow Equations (FFVHC-ACE), for large-eddy simulation (LES) and its aeronautical applications. FFVHC-ACE enables high-fidelity LES of high-Reynolds-number flows around complex geometries by adopting three key numerical methods: hierarchical Cartesian grids, wall modeling in LES, and the kinetic-energy and entropy preserving (KEEP) scheme. The hierarchical Cartesian grids allow fully automated grid generation for complex geometries in FFVHC-ACE, and high-fidelity LES of high-Reynolds-number flows around complex geometries is realized by the wall modeling and the KEEP scheme on the non-body-fitted hierarchical Cartesian grids. We apply FFVHC-ACE to wall-modeled LES around high-lift aircraft configurations at wind-tunnel-scale Reynolds number and real-flight Reynolds number , demonstrating the capability of FFVHC-ACE for fully automated grid generation and high-fidelity simulations around complex aircraft configurations. The computed flowfield and aerodynamic forces at the wind-tunnel-scale Reynolds number agree well with the experimental data provided in the past AIAA High Lift Prediction Workshop (Rumsey et al., Journal of Aircraft, Vol. 56, No. 2, 2019, pp. 621–644). Furthermore, the wall-modeled LES at the real-flight Reynolds number shows good agreement of the lift coefficient with flight-test data.