Use Of Large Eddy Simulation Research Articles

Future Use of Large Eddy Simulation in Aero‐engines

Computational fluid dynamics (CFD) has become a critical tool in the design of aero-engines. Increasing demand for higher efficiency, performance, and reduced emissions of noise and pollutants has focused attention on secondary flows, small scale internal flows, and flow interactions. In conjunction with low order correlations and experimental data, RANS (Reynolds-averaged Navier–Stokes) modeling has been used effectively for some time, particularly at high Reynolds numbers and at design conditions. However, the range of flows throughout an engine is vast, with most, in reality being inherently unsteady. There are many cases where RANS can perform poorly, particularly in zones characterized by strong streamline curvature, separation, transition, relaminarization, and heat transfer. The reliable use of RANS has also been limited by its strong dependence on turbulence model choice and related ad-hoc corrections. For complex flows, large-eddy simulation (LES) methods provide reliable solutions, largely independent of turbulence model choice, and at a relatively low cost for particular flows. LES can now be used to provide in depth knowledge of flow physics, for example, in areas such as transition and real wall roughness effects. This can be used to inform RANS and lower order modeling (LOM). For some flows, LES can now even be used for design. Existing literature is used to show the potential of LES for a range of flows in different zones of the engine. Based on flow taxonomy, best practices including RANS/LES zonalization, meshing requirements, and turbulent inflow conditions are introduced, leading to the proposal of a tentative expert system for industrial use. In this way, LES becomes a well controlled tool, suitable for design use and reduces the burden on the end user. The problem sizes tackled however have lagged behind potential computing power, hence future LES use at scale requires substantial progress in several key areas. Current and future solver technologies are thus examined and the potential current and future use of LES is considered.

Development of an Engine Representative Combustor Simulator Dedicated to Hot Streak Generation

Nowadays, the lack of confidence in the prediction of combustor-turbine interactions and more specifically our ability to predict the migration of hot spots through this interface leads to the application of extra safety margins, which are detrimental to an optimized turbine design and efficiency. To understand the physics and flow at this interface, a full 360 deg nonreactive combustor simulator (CS) representative of a recent lean burn chamber together with a 1.5 turbine stage is instrumented at DLR in Gottingen (Germany) within the European project FACTOR. The chamber operates with axial swirlers especially designed to reproduce engine-realistic velocity and temperature distortion profiles, allowing the investigation of the hot streaks transport through the high pressure (HP) stage. First, a true scale three injector annular sector of the CS without turbine is assembled and tested at the University of Florence. To generate the hot steaks, the swirlers are fed by an air flow at 531 K, while the liners are cooled by an effusion system fed with air at ambient temperature. In addition to static pressure taps and thermocouples, the test rig will be equipped with an automatic traverse system which allows detailed measurements at the combustor exit by means of a 5-hole probe, a thermocouple, and hot wire anemometers. This paper presents the design process and instrumentation of the trisector CS, with a special focus on large Eddy simulations (LES) which were widely used to validate the design choices. It was indeed decided to take advantage of the ability and maturity of LES to properly capture turbulence and mixing within combustion chambers, despite an increased computational cost as compared to usual Reynolds averaged Navier Stokes (RANS) approaches. For preliminary design, simulations of a single periodic sector (representative of the DLR full annular rig) are compared to simulations of the trisector test rig, showing no difference on the central swirler predictions, comforting the choice for the trisector. In parallel, to allow hot wire anemometry (HWA) measurements, the selection of an isothermal operating point, representative of the nominal point, is assessed and validated by use of LES.

Large eddy simulation of flows in industrial compressors: a path from 2015 to 2035.

A better understanding of turbulent unsteady flows is a necessary step towards a breakthrough in the design of modern compressors. Owing to high Reynolds numbers and very complex geometry, the flow that develops in such industrial machines is extremely hard to predict. At this time, the most popular method to simulate these flows is still based on a Reynolds-averaged Navier-Stokes approach. However, there is some evidence that this formalism is not accurate for these components, especially when a description of time-dependent turbulent flows is desired. With the increase in computing power, large eddy simulation (LES) emerges as a promising technique to improve both knowledge of complex physics and reliability of flow solver predictions. The objective of the paper is thus to give an overview of the current status of LES for industrial compressor flows as well as to propose future research axes regarding the use of LES for compressor design. While the use of wall-resolved LES for industrial multistage compressors at realistic Reynolds number should not be ready before 2035, some possibilities exist to reduce the cost of LES, such as wall modelling and the adaptation of the phase-lag condition. This paper also points out the necessity to combine LES to techniques able to tackle complex geometries. Indeed LES alone, i.e. without prior knowledge of such flows for grid construction or the prohibitive yet ideal use of fully homogeneous meshes to predict compressor flows, is quite limited today.

Large Eddy Simulation of wind farm aerodynamics: A review

To study wind farm aerodynamics (WFA), the Wind Power industry currently relies on simple Engineering Models (EM) that simulate wind farms using basic principles of physics and empirically established approximations. EMs are fast and accurate for an overview of WFA and gauging mean power production, but cannot resolve phenomena like wake meandering, effect of atmospheric stratification on wake development, a turbine׳s response to partial wake interaction and yawed inflows etc., and their relation with turbine loading, which require a first principle physics-based model, namely Computational Fluid Dynamics (CFD). Although advances in computer technology have promoted the application of CFD, the study of WFA is yet unworkable with Direct Numerical Simulation, which is the most comprehensive CFD technique. Thus, as a trade-off between cost and detail, researchers must resort to Large Eddy Simulation (LES) to garner thorough knowledge of WFA, which could in succession help the industry improve engineering models. This paper summarises the contributions of various LES investigations into WFA and how they have helped broaden our understanding of the subject. Additionally, the article touches upon the optimal use of LES and of the resultant data, and also the challenges faced by LES.

Wall-Modelled Large-Eddy Simulation of a hot Jet-In-Cross-Flow with turbulent inflow generation

Hot jets-in-cross-flow are frequently encountered in aeronautics and the accurate estimation of the wall temperature in the jet wake is crucial during the early design of a new aircraft. However, common two-equation RANS models fail at estimating the wall temperature in the jet wake. The use of Large-Eddy Simulation, which seems to be a promising solution at first sight, is not applicable due to its prohibitive computational cost on such large Reynolds number wall-bounded flows. For an affordable cost, we propose a strategy which consists in: reducing the computational domain to a small region around the phenomenon of interest (RANS-LES embedded approach), perform a Wall-Modelled Large-Eddy Simulation (WMLES) in the reduced domain and generate a turbulent inflow at the reduced domain inlet. The test case selected is a hot Jet-In-Cross-Flow experimentally studied by Albugues (2005) [1]. We simulate the real geometry of the wind-tunnel model, which imposes strong constraints on the inflow generation and numerical method. It is shown that an advanced inflow generation, combining a stochastic velocity fluctuation injection and a dynamic forcing term (Laraufie et al., 2011) [17], is mandatory to obtain a realistic turbulent flow upstream of the jet. In the jet wake, the wall temperature estimated by the WMLES agrees well with the experimental measurements.

A velocity divergence constraint for large-eddy simulation of low-Mach flows

The velocity divergence (rate of fluid volumetric expansion) is a flow field quantity of fundamental importance in low-Mach flows. It directly affects the local mass density and therefore the local temperature through the equation of state. In this paper, starting from the conservative form of the sensible enthalpy transport equation, we derive a discrete divergence constraint for use in large-eddy simulation (LES) of low-Mach flows. The result accounts for numerical transport of mass and energy, which is difficult to eliminate in relatively coarse, engineering LES calculations when total variation diminishing (TVD) scalar transport schemes are employed. Without the correction terms derived here, unresolved (numerical) mixing of gas species with different heat capacities or molecular weights may lead to erroneous mixture temperatures and ultimately to an imbalance in the energy budget. The new formulation is implemented in a publicly available LES code called the Fire Dynamics Simulator (FDS). Accuracy of the flow solver for transport is demonstrated using the method of manufactured solutions. The conservation properties of the present scheme are demonstrated on two simple energy budget test cases, one involving a small fire in a compartment with natural ventilation and another involving mixing of two gases with different thermal properties.

Study of a stall cell using stereo particle image velocimetry

The structure of Stall Cells (SCs) on wings is analyzed on the basis of stereo particle image velocimetry measurements. All experiments regard a Reynolds number 0.87 × 106 flow around a rectangular wing with endplates and an aspect ratio of 2.0. The inherently unstable stall cell is stabilized by means of a localized spanwise disturbance. Velocity, vorticity, and Reynolds stress data above the wing and in the wake are presented and discussed, also in combination with Computational Fluid Dynamics data. The present study completes and clarifies the previously suggested models regarding the SC structure. The SC emerges in between the separation and trailing edge shear layer where three different types of vortices are identified: (a) the stall cell vortices that start normal to the wing surface and continue downstream aligned with the free stream, (b) the separation line vortex, and (c) the trailing edge line vortex that both run parallel to the wing trailing edge and grow significantly at the center of the stall cell. Analysis of the Reynolds stress data reveals high anisotropy. Concentration of high streamwise shear stress values is connected to the two shear layers and high cross shear Reynolds stresses are connected to vortex stretching. High normal Reynolds stress values are observed (a) in the separation but not in the trailing edge shear layer indicating the flapping of the former and (b) along the stall cell vortices indicating their wandering motion. The eddy viscosity based Reynolds averaged Navier-Stokes simulations are found in good qualitative agreement with the experiments in terms of the type and position of the identified vortex structures, an agreement which is linked to the correct trend in the predicted shear Reynolds stresses distributions. Quantitative deviations of the numerical results from the measurements are attributed to the isotropic definition of the turbulence model. Therefore, use of large eddy simulation is suggested for better prediction of the flow.

Image-based large-eddy simulation in a realistic left heart

A numerical framework allowing insight in fluid dynamics inside patient-specific human hearts is presented. The heart cavities and their wall dynamics are extracted from medical images, with the help of a non-linear image registration algorithm, in order to obtain a patient-specific moving numerical domain. Flow equations are written on a conformal moving computational domain, using an Arbitrary Lagrangian–Eulerian framework. Resulting equations are solved numerically with a fourth-order finite-volume technique. Application of this framework to compute a patient-specific left heart flow is presented as well. The blood flow is characterized by its transitional nature, resulting in a complex cyclic flow. Flow dynamics is analysed in order to reveal the main fluid phenomena and to obtain insights into the physiological patterns commonly detected. It is demonstrated that the flow is neither laminar nor fully turbulent, thus justifying a posteriori the use of Large Eddy Simulation.

Large-eddy simulation of decaying isotropic turbulence across a grid refinement interface using explicit filtering and reconstruction

This paper explores numerical errors that arise when large-eddy simulation (LES) is used with adaptive mesh refinement (AMR). LES and AMR combined can reduce the computational cost of turbulence simulations compared to direct numerical simulations, but are rarely used together due to complications that arise with the application of the turbulence closure model at different grid resolutions. Errors appear at grid refinement interfaces due to dependence of computed quantities on the LES filter width and insufficient smoothness of the solution at the grid scale. Here, explicit filtering and approximate reconstruction of the unfiltered velocity field are used to mitigate the effects of these errors in a simulation of decaying isotropic turbulence advected past a grid refinement interface. In particular, different explicit filter types and levels of reconstruction are tested. Explicit filtering with zero-level reconstruction is found to produce the best long-term convergence to a uniform grid solution with minimum perturbation at the interface. Higher levels of reconstruction yield better near-interface convergence. When explicit filtering is used, the explicit filter width transition is more important than the grid spacing transition in terms of solution convergence and interface perturbation. These results inform the use of LES on more complicated AMR grids.

The eruption of a high-pressure cylindrical heavy gas cloud

The physical explosion of a high-pressure and cylindrical dense gas (SF6) cloud has been simulated with the use of large eddy simulation (LES), and hybrid high-order schemes have been employed to solve the LES equations. Our results show that, while the shockwave is exploding from the SF6 to the air, it bifurcates into the reflected rarefaction wave and the transmitted shock, and a reverse shock also appears. The rarefaction wave moves inward first, the Richtmyer–Meshkov (RM) instabilities occur as the transmitted shock accelerates the interface between the SF6 and the air. Later, the rarefaction wave merges with the reverse shock, and finally converges at the origin of the cloud, which generates a strong circular reflected shock, and makes the flow field complex.

A Computational Study of Turbulent Airflow and Tracer Gas Diffusion in a Generic Aircraft Cabin Model

In order to study the capability of computational methods in investigating the mechanisms associated with disease and contaminants transmission in aircraft cabins, the computational fluid dynamics (CFD) models are used for the simulation of turbulent airflow and tracer gas diffusion in a generic aircraft cabin mockup. The CFD models are validated through the comparisons of the CFD predictions with corresponding experimental measurements. It is found that using large eddy simulation (LES) with the Werner-Wengle wall function, one can predict unsteady airflow velocity field with relatively high accuracy. However in the middle region of the cabin mockup, where the recirculation of airflow takes place, the accuracy is not as good as that in other locations. By examining different k-ε models, the current study recommends the use of the RNG k-ε model with the nonequilibrium wall function as an Reynolds averaged Navier-Stokes model for predicting the steady-state airflow velocity. It is also found that changing the nozzle height has a significant effect on the flow behavior in the middle and upper part of the cabin, while the flow pattern in the lower part is not affected as much. Through the use of LES and species transport model in simulating tracer gas diffusion, a very good agreement between predicted and measured tracer gas concentration is achieved for some monitoring locations, but the agreement level is not uniform for all the locations. The reasons for the deviations between prediction and measurement for those locations are discussed.

Effective eddy viscosity in stratified turbulence

This paper investigates the effective eddy viscosity inferred from direct numerical simulations of decaying stratified and non-stratified turbulence. It is shown that stratification affects the horizontal eddy viscosity dramatically, by increasing non-local energy transfer between large and small horizontal scales. This non-local horizontal energy transfer is around 20% of the local horizontal energy transfer at the cutoff wavenumber kc = 40. The non-local horizontal energy transfer occurs at large vertical wavenumbers, which may be larger than the buoyancy wavenumber kb = N/urms, where N is the buoyancy frequency and urms is the root-mean-square velocity. By increasing the value of the test cutoff wavenumber kc from large scales to the dissipation range, the non-local horizontal eddy viscosity decreases and the local eddy viscosity is dominant. Overall, the presence of stratification can significantly change the features of subgrid-scale (SGS) motions. Current SGS models should, therefore, be modified for use in large-eddy simulation of stratified turbulence.

Large eddy simulation of unsteady turbulent flow in a semi-industrial size spray dryer

This paper is concerned with the numerical simulation of gas flow inside spray dryers, which is known to be highly transient. Experimental evidence presented in the literature to date is acquired either at a scale much smaller than industrial or at a limited number of locations in the system (e.g. the central axis). Alternatively, computational fluid dynamics (CFD) simulation, in particular methods based on the Reynolds-averaged Navier–Stokes (RANS) equations, may provide the information of the complete flow field at industrial scales. However, RANS methods to some extent are hampered by the limitations of turbulence modelling. In this study, it has been investigated whether the current limitations of experiments and RANS CFD simulations can be omitted by the use of large eddy simulation (LES). The spray dryer studied was of semi-industrial size with a volume of 60 m3. Via grid size and time step refinement studies, the minimum numerical requirements for obtaining converged time averaged velocity profiles could be identified. Spectral analysis and estimates of the Kolmogorov scales demonstrated that LES resolves scales well into the inertial subrange of turbulence scales. The LES calculation shows a complex precession of the jet. The jet moves around in the drying chamber seemingly with no reoccurring modes. Even with very long sampling times, no typical frequencies could be found. This work demonstrates the feasibility of LES for a spray dryer of semi-industrial scale and the ability of LES to predict large-scale motions. Future work will include the further validation of the LES via comparison with laser Doppler anemometry measurements and, subsequently, the comparison of RANS, unsteady RANS and LES simulations.

Large Eddy Simulation for Dispersed Bubbly Flows: A Review

Large eddy simulations (LES) of dispersed gas-liquid flows for the prediction of flow patterns and its applications have been reviewed. The published literature in the last ten years has been analysed on a coherent basis, and the present status has been brought out for the LES Euler-Euler and Euler-Lagrange approaches. Finally, recommendations for the use of LES in dispersed gas liquid flows have been made.

Entrainment and detrainment in cumulus convection: an overview

AbstractEntrainment and detrainment processes have been recognised for a long time as key processes for cumulus convection and have recently witnessed a regrowth of interest mainly due to the capability of large‐eddy simulations (LES) to diagnose these processes in more detail. This article has a twofold purpose. Firstly, it provides a historical overview of the past research on these mixing processes, and secondly, it highlights more recent important developments. These include both fundamental process studies using LES aiming to improve our understanding of the mixing process, but also more practical studies targeted toward an improved parametrised representation of entrainment and detrainment in large‐scale models. A highlight of the fundamental studies resolves a long‐lasting controversy by showing that lateral entrainment is the dominant mixing mechanism in comparison with the cloud‐top entrainment in shallow cumulus convection. The more practical studies provide a wide variety of new parametrisations with sometimes conflicting approaches to the way in which the effect of the free tropospheric humidity on the lateral mixing is taken into account. An important new insight that will be highlighted is that, despite the focus in the literature on entrainment, it appears that it is rather the detrainment process that determines the vertical structure of the convection in general and the mass flux especially. Finally, in order to speed up progress and stimulate convergence in future parametrisations, stronger and more systematic use of LES is advocated. Copyright © 2012 Royal Meteorological Society

Prediction of the spread of smoke in a huge transit terminal subway station under six different fire scenarios

A fire at a subway station can cause tremendous loss of property and heavy casualties, and smoke has been reported to be the main cause of these casualties. This article reports the use of Large Eddy Simulation (LES) simulation to study the dispersion of fire-induced smoke that is buoyancy driven in a subway station. This station was located within a huge transit terminal at Xi’an, China. The influence of natural and mechanical ventilation was studied in this station, which has atrium height ranging from 0 to 15m, roof window dimensions of 1×1m2 and 5×1m2, and rates of air change ranging between 5 and 11h−1. In addition, six different fire-source locations with heat-release rates (HRRs) of 4 and 7.5MW are discussed. Four grid systems were compared, and the sensitivity of these grid systems was checked to ensure grid independence. Preliminary analysis of results indicates that the atrium of a subway station poses significant limitations on the horizontal dispersion of smoke. A large soot-density gradient has also been observed in stations with a high atrium. An exponential relationship has been proposed to define the correlation between atrium height and duration of smoke filling, with a relative coefficient (R) of 0.9995. The results indicate that mechanical ventilation can markedly control the spread of smoke in the horizontal direction but it has little effect on the dispersion of smoke in the vertical direction. If a fire starts directly below the atrium, most of the smoke generated will converge toward the upper regions of the atrium. In addition, the simulation helps predict the maximum soot density of smoke generated by different heat releases (4 and 7.5MW).

The Effect of Scale on the Applicability of Taylor’s Frozen Turbulence Hypothesis in the Atmospheric Boundary Layer

Taylor’s frozen turbulence hypothesis is the central assumption invoked in most experiments designed to investigate turbulence physics with time resolving sensors. It is also frequently used in theoretical discussions when linking Lagrangian to Eulerian flow formalisms. In this work we seek to quantify the effectiveness of Taylor’s hypothesis on the field scale using water vapour as a passive tracer. A horizontally orientated Raman lidar is used to capture the humidity field in space and time above an agricultural region in Switzerland. High resolution wind speed and direction measurements are conducted simultaneously allowing for a direct test of Taylor’s hypothesis at the field scale. Through a wavelet decomposition of the lidar humidity measurements we show that the scale of turbulent motions has a strong influence on the applicability of Taylor’s hypothesis. This dependency on scale is explained through the use of dimensional analysis. We identify a ‘persistency scale’ that can be used to quantify the effectiveness of Taylor’s hypothesis, and present the accuracy of the hypothesis as a function of this non-dimensional length scale. These results are further investigated and verified through the use of large-eddy simulations.

A filter-independent model identification technique for turbulent combustion modeling

In this paper, we address a method to reduce the number of species equations that must be solved via application of Principal Component Analysis (PCA). This technique provides a robust methodology to reduce the number of species equations by identifying correlations in state-space and defining new variables that are linear combinations of the original variables. We show that applying this technique in the context of Large Eddy Simulation allows for a mapping between the reduced variables and the full set of variables that is insensitive to the size of filter used. This is notable since it provides a model to map state variables to progress variables that is a closed model.As a linear transformation, PCA allows us to derive transport equations for the principal components, which have source terms. These source terms must be parameterized by the reduced set of principal components themselves. We present results from a priori studies to show the strengths and weaknesses of such a modeling approach. Results suggest that the PCA-based model can identify manifolds that exist in state space which are insensitive to filtering, suggesting that the model is directly applicable for use in Large Eddy Simulation. However, the resulting source terms are not parameterized with an accuracy as high as the state variables.

Statistical properties of an ideal subgrid-scale correction for Lagrangian particle tracking in turbulent channel flow

One issue associated with the use of Large-Eddy Simulation (LES) to investigate the dispersion of small inertial particles in turbulent flows is the accuracy with which particle statistics and concentration can be reproduced. The motion of particles in LES fields may differ significantly from that observed in experiments or direct numerical simulation (DNS) because the force acting on the particles is not accurately estimated, due to the availability of the only filtered fluid velocity, and because errors accumulate in time leading to a progressive divergence of the trajectories. This may lead to different degrees of inaccuracy in the prediction of statistics and concentration. We identify herein an ideal subgrid correction of the a-priori LES fluid velocity seen by the particles in turbulent channel flow. This correction is computed by imposing that the trajectories of individual particles moving in filtered DNS fields exactly coincide with the particle trajectories in a DNS. In this way the errors introduced by filtering into the particle motion equations can be singled out and analyzed separately from those due to the progressive divergence of the trajectories. The subgrid correction term, and therefore the filtering error, is characterized in the present paper in terms of statistical moments. The effects of the particle inertia and of the filter type and width on the properties of the correction term are investigated.

Large Eddy Simulation of Transitional Separated Flow over a Flat Plate and a Compressor Blade

The ability of Large-Eddy Simulation (LES) to predict transitional separation bubbles is investigated, with particular emphasis being placed on the response to free-stream-turbulence. The principal objective is to quantify the penalties, relative to Direct Numerical Simulations (DNS), that arise from the coarser resolution and the use of subgrid-scale models. Two flow configurations are considered: a flat-plate boundary layer, subjected to different free-stream-turbulence levels, ranging from 0 to 2% (at the point of separation), and the flow over a compressor blade at 0 and 3.25% turbulence levels. For both cases, results are compared with DNS data. A number of challenges associated with the use of LES in transitional flows are addressed, including the representation of the decay of free-stream turbulence and the mesh resolution needed for a correct description of the growth of instability waves in the early stage of transition.