Large Eddy Simulation Grid Research Articles

A transformer-based convolutional method to model inverse cascade in forced two-dimensional turbulence

The present work proposes a novel transformer-based convolutional neural network (TransCNN) method to effectively model the inverse energy cascade in two dimensional (2D) turbulence. The TransCNN structure combines large-scale features extracted by transformer with small-scale features from convolutional layers, thus is considered suitable for multi-scale modeling. The novel TransCNN method has been applied to model sub-grid scale (SGS) stress for large-eddy simulation (LES) of 2D turbulence, under the extremely challenging situation that the LES grid is too coarse to resolve the external forcing scale. The data-driven model trained by the novel TransCNN structure is compared to two deep CNN models with varying complexities. All models exhibit proficiency during a priori tests. Notably, TransCNN surpasses its counterparts in predictive accuracy and generalizability in a posteriori tests. An investigation into the receptive fields reveals that the TransCNN model can efficiently leverage global information with the transformer structure, which is key to its superior performance in representing the inverse energy cascade in the 2D turbulent simulations.

Impacts of Stochastic Coalescence Variability on Warm Rain Initiation Using Lagrangian Microphysics in Box and Large-Eddy Simulations

Abstract Various coalescence methods for Lagrangian microphysics schemes are tested in box and large-eddy simulation (LES) models, including the stochastic all-or-nothing (AON) superdroplet method (SDM) and a deterministic version of SDM (dSDM) that applies a fractional approach similar to the average impact method. In LES, variabilities driven by microphysics and by flow realizations are separated using the “piggybacking” technique. Rain initiation averaged over many realizations of the box model is delayed, and rain variability increases as the number of superdrops per collision volume NSD is decreased using SDM. In contrast, rain initiation time using SDM in LES is insensitive to NSD for 32 ≤ NSD ≤ 512. This is explained through the interaction between LES grid boxes, each acting as a separate collision volume. Variability across the ensemble of LES collision volumes using SDM results in rain quickly initiating in some of the LES grid cells at low NSD and leading to a similar overall timing of rain initiation from the cloud compared to simulations with high NSD. There is a ∼20% decrease in the total rain mass and mean rain flux as NSD is increased from 32 to 256, with little additional change as NSD is increased from 256 to 512. The fractional coalescence approach in dSDM leads to reduced microphysical variability and a 15–18-min delay in rain initiation compared to SDM. An additional LES ensemble with microphysical variability feeding back to the dynamics shows that flow variability dominates the impact of microphysical variability on rain properties. Thus, flow variability must be constrained to isolate impacts of microphysical variability. Significance Statement An emerging tool in cloud microphysics modeling represents drops by computational particles called “superdroplets” that evolve in the modeled flow. Recent studies have documented that different superdroplet models produce widely varying predictions of rain. Understanding these differences is an important step in the wider adoption of these models by the community. Here, we examine different methods for representing droplet collision-coalescence in a superdroplet model and sensitivity to the number of superdroplets employed. The timing of rain initiation is insensitive to the superdroplet number in 3D cloud simulations, but rain is substantially delayed using a coalescence method that limits random variability in droplet collisions. We also show that flow variability must be constrained to isolate impacts of microphysical variability on rain properties.

Hybrid aeroacoustic investigation of turbulent 90° pipe bend flow with source terms from Large-Eddy Simulation

The internal, low Mach number, turbulent flow-induced acoustic field of a cylindrical 90° pipe bend with a curvature radius Rm/D=1.02 is investigated using hybrid Computational Aero-Acoustics (CAA). Incompressible Large-Eddy Simulation (LES) is used to computationally describe the turbulent flow field at bulk Reynolds numbers ReB=5000,12000,26000 and 66000. The incompressible LES predicts all salient flow features, like the separation from the inner wall of the bend followed by a turbulent recirculation zone, in good agreement with literature. The internal flow-induced noise is predicted by Acoustic Propagation Simulations (APS) applying Lighthill’s Equation (LE), Ribner’s Dilatation Equation (RDE) and the Perturbed Convective Wave Equation (PCWE), provided with acoustic source terms obtained from the incompressible LES. The turbulent recirculation zone is the most active region for generating acoustic sources. Among the two constituent components of the PCWE source term, the local pressure time derivative term turned out as clearly dominant in this region, so that PCWE and RDE predicted similar acoustic fields. The sub-grid scale (SGS) model applied in LES notably contributes only at high frequencies to the total LE source term, starting from f=2000 Hz for the lower Reynolds numbers and beyond for the higher. The mapping of source terms from the boundary layer resolving LES grid onto the typically coarser preferably uniform APS grid leads to lower high-frequency signal amplitudes for coarser APS mesh size in regions with small turbulent structures. The impact of this amplitude reduction on the predicted acoustics sound power levels was still very small, exhibiting relative differences less than 0.7% for the different APS meshes. Despite the LES-specific limitations expected from the decrease of directly resolved content for higher Reynolds numbers, the use of increasingly underresolved source terms did not significantly lower the accuracy of finally predicted acoustics. The comparison of the acoustic results against data from dedicated measurements generally showed very good agreement for all Reynolds numbers.

Large Eddy Simulation of Separated Flows on Unconventionally Coarse Grids

Abstract We examine and benchmark the emerging idea of applying the large eddy simulation (LES) formalism to unconventionally coarse grids where Reynolds-averaged Navier–Stokes (RANS) would be considered more appropriate at first glance. We distinguish this idea from very large eddy simulation and detached eddy simulation, which require switching between RANS and LES formalism. LES on RANS grid is appealing because first, it requires minimal changes to a production code; second, it is more cost-effective than LES; third, it converges to LES; and most importantly, it accurately predicts flows with separation. This work quantifies the benefit of LES on RANS-like grids as compared to RANS on the same grids. Three canonical cases are considered: periodic hill, backward-facing step, and jet in cross flow. We conduct direct numerical simulation (DNS), proper LES on LES grids, LES on RANS-quality grids, and RANS. We show that while the LES solutions on the RANS-quality grids are not grid converged, they are twice as accurate as the RANS on the same grids.

Grid Convergence Properties of Wall-Modeled Large Eddy Simulations in the Asymptotic Regime

Abstract This study explores the grid convergence properties of wall-modeled large eddy simulation (WMLES) solutions as the large eddy simulation (LES) grid approaches the direct numerical simulation (DNS) grid. This aspect of WMLES is fundamental but has not been previously investigated or documented. We investigate two types of grid refinements: one where the LES/wall-model matching location is fixed at an off-wall grid point, and another where the matching location is fixed at a specific distance from the wall. In both cases, we refine the LES grid simultaneously in all three Cartesian directions, with grid resolution ranging from typical LES resolution to typical DNS resolution. Our focus is on examining the mean flow and turbulent kinetic energy (TKE) as the grid refines. While the turbulence statistics consistently converge toward the DNS solution, we observe nonmonotonic convergence in the mean flow statistics. We show that improving the grid resolution does not necessarily enhance the accuracy of the mean flow predictions. Specifically, we identify a negative log layer mismatch when the LES/wall-model matching location lies below the logarithmic layer, regardless of grid resolution and matching location. Finally, we demonstrate that the nonmonotonic convergence of the mean flow can lead to misleading conclusions from grid convergence studies of WMLES.

Large-Eddy-Simulation of Turbulent Non-Premixed Hydrogen Combustion Using the Filtered Tabulated Chemistry Approach

Abstract With air traffic expected to grow 5% annually until the year 2030, alternative fuels such as hydrogen are being investigated in order to tackle the current environmental crisis. Due to safety concerns, future hydrogen combustion chambers will require new designs of injection systems and are expected to operate under multimode combustion regimes. From a large-eddy-simulation (LES) perspective, a prerequisite for the shift toward new hydrogen combustion chamber technologies is a robust turbulent combustion model capable of functioning in non-premixed conditions. Turbulent combustion modeling using flame front filtering is a well-developed strategy in premixed combustion (filtered-tabulated chemistry for large-Eddy-simulation (F-TACLES)). This approach has been extended to non-premixed flames however, it suffers from high flame filter size sensitivity. Moreover, thin hydrogen flame fronts will result in lower resolution on the LES grid, potentially amplifying this issue. In order to address the feasibility of the non-premixed F-TACLES model applied to hydrogen fuel, simple one-dimensional and two-dimensional laminar counterflow diffusion flames are computed. The model is then tested on the three-dimensional Sandia hydrogen jet flame with a Reynolds number of 10,000. Simulations and a priori tests show that tabulated subgrid-scale correction terms are stiff and can result in nonphysical results, however the model is capable of correctly reproducing non-premixed flame structures for small filter sizes.

A consistent MMC-LES approach for turbulent premixed flames

A multiple mapping conditioning (MMC) approach is coupled with large eddy simulations (LES) for the prediction of the flame propagation, the flame structure and finite rate chemistry effects including NO formation in turbulent premixed flames. The mixing term in MMC relies on a reference field that characterizes the chemical composition within the flame and it is typically taken to be the reaction progress variable. In LES, this reference field needs to be artificially thickened to be resolved on the LES grid. A novel MMC mixing time scale model that had been developed with the aid of direct numerical simulation (DNS), has been adapted to account for the thickening of the reference field. Equally, the particle selection mechanism for particle mixing has been modified to ensure localness of mixing in composition space and to preserve the prediction of a realistic (thin) flame thickness by the stochastic particles. A set of freely propagating turbulent premixed flames with varying equivalence ratios is used to test the performance of the MMC model. Results of MMC-LES are compared with reference DNS that include full chemistry. Additional simulations were carried out where MMC is coupled with quasi-DNS where the flow field is fully resolved but the reaction progress variable field is thickened and its source term is closed with a flamelet approach. This allows to isolate the different modelling assumptions invoked in MMC-LES. Deviations from the exact solution can then be associated with the MMC approach and with thickening of the reference field separately and they are small. The MMC-LES predictions of the turbulent premixed flame propagation speed and flame structure are very good and finite rate chemistry effects can be captured well which is demonstrated by good agreement of predicted NO with DNS data.

Numerical investigation on turbulence-radiation interaction in the UMD turbulent line fires

This work aims to study the turbulence-radiation interaction (TRI) in the UMD turbulent line fires with increasing levels of underventilation until strong local extinction occurs. Large eddy simulations (LES) are performed with a radiative flamelet/progress variable model to account for local extinction events. The effects of the resolved-scale and subgrid-scale (SGS) TRI on global radiant fractions and local radiative power are quantified. Results show that the total TRI makes a considerable contribution to radiation for all oxygen concentrations (XO2) tested in this work. Specifically, in the most under-ventilated case of XO2=13% featuring strong local extinction, more than 95% of the global radiant fraction is induced by the TRI. More interestingly, it is found that on a regular LES grid resolving more than 80% of the turbulent kinetic energy, the resolved-scale TRI plays a leading role in the XO2=21% case, while the SGS TRI becomes dominant in the XO2=13% case. Further investigations reveal that the temperature self-correlation and the absorption coefficient-temperature correlation are the key components for TRI, which is confirmed by the corresponding conditional statistics. This study highlights the importance of including an SGS TRI model to accurately predict radiation fields in LES of fires with local extinction.

A new coupling of a GPU‐resident large‐eddy simulation code with a multiphysics wind turbine simulation tool

AbstractThe development of new wind farm control strategies can benefit from combined analysis of flow dynamics in the farm and the behavior of individual turbines within one simulation environment. In this work, we present such an environment by developing a new coupling between the large‐eddy simulation (LES) code GRASP and the multiphysics wind turbine simulation tool OpenFAST via an actuator line model (ALM). In addition, the implementation of the recently proposed filtered actuator line model (FALM) within the coupling is described. The new ALM implementation is cross‐verified with results from four other commonly used research LES codes. The results for the blade loads and the near wake obtained with the new coupling are consistent with the other codes. Deviations are observed in the far wake. The results further indicate that the FALM is able to reduce the lift and power overprediction from which the traditional ALM suffers on coarse LES grids. This new simulation environment paves the way for future wind farm simulations under realistic weather conditions by leveraging GRASP's ability to impose data from large‐scale meteorological models as boundary conditions.

Numerical investigation of air curtain flows in the doorway of a building using RANS and LES

Air curtains are commonly adopted in building flows to facilitate aerodynamic sealing against the exchange flow that occurs across an open doorway due to the density differences owing to buoyancy. This paper reports numerical simulations of air curtain flows using Reynolds-averaged Navier–Stokes (RANS) and large-eddy simulation (LES) methodologies. Computations are performed for various operating conditions that are described using a deflection modulus (Dm) which ranges from 0 to 1.25. Physically, Dm represents the relative strength of the jet momentum flux compared to the transverse stack effect. The value of this deflection modulus, along with qualitative distribution of buoyancy, demarcates the operating regime of an air curtain into unstable (Dm⪅ 0.1), stable (Dm⪆ 0.2), and strong installation (Dm⪆ 0.6). The quantification of aerodynamic sealing is established using effectiveness, E, which suggests that the optimal operation lies in the stable regime with the poorest performance in the unstable regime. The comparison between various simulation methodologies suggests that 2D RANS computations generally over-predict the effectiveness, whereas both 3D RANS and LES are able to estimate E reasonably well, especially at practical values of Dm. If the employed LES grid is sufficiently fine, the predicted effectiveness falls within the experimental uncertainty limits. Furthermore, we quantify the statistics of turbulence in the flow field using the transport equation of turbulent kinetic energy. This analysis reveals that shear production and turbulent dissipation are dominant processes in the flow. In comparison, the effects of transport terms and buoyancy production are rather small. These quantities are shown to scale well with w3/l where w and l are the characteristic velocity and length scales associated with an air curtain. Visualization of large-scale coherent structures using the Q-criterion depicted the presence of long spanwise vortices in the shear layer with a wide range of scales of motion at higher Dm. These vortical structures break down significantly upon impinging on the wall (floor), resulting in a complex spatiotemporal distribution of vortices.

Modeling Near-Surface Turbulence in Large-Eddy Simulations of a Tornado: An Application of Thin Boundary Layer Equations

Abstract Unsteadiness and horizontal heterogeneities frequently characterize atmospheric motions, especially within convective storms, which are frequently studied using large-eddy simulations (LES). The models of near-surface turbulence employed by atmospheric LES, however, predominantly assume statistically steady and horizontally homogeneous conditions (known as the equilibrium approach). The primary objective of this work is to investigate the potential consequences of such unrealistic assumptions in simulations of tornadoes. Cloud Model 1 (CM1) LES runs are performed using three approaches to model near-surface turbulence: the “semi-slip” boundary condition (which is the most commonly used equilibrium approach), a recently proposed nonequilibrium approach that accounts for some of the effects of turbulence memory, and a nonequilibrium approach based on thin boundary layer equations (TBLE) originally proposed by the engineering community for smooth-wall boundary layer applications. To be adopted for atmospheric applications, the TBLE approach is modified to account for the surface roughness. The implementation of TBLE into CM1 is evaluated using LES results of an idealized, neutral atmospheric boundary layer. LES runs are then performed for an idealized tornado characterized by rapid evolution, strongly curved air parcel trajectories, and substantial horizontal heterogeneities. The semi-slip boundary condition, by design, always yields a surface shear stress opposite the horizontal wind at the lowest LES grid level. The nonequilibrium approaches of modeling near-surface turbulence allow for a range of surface-shear-stress directions and enhance the resolved turbulence and wind gusts. The TBLE approach even occasionally permits kinetic energy backscatter from unresolved to resolved scales. Significance Statement The traditional approach of modeling the near-surface turbulence is not suitable for a tornado characterized by rapid evolution, strongly curved air parcel trajectories, and substantial horizontal heterogeneities. To understand the influence of statistically unsteady and horizontally heterogeneous near-surface conditions on tornadoes, this work adopts a fairly sophisticated approach from the engineering community and implements it into a widely used atmospheric model with necessary modifications. Compared to the traditional approach, the newly implemented approach produces more turbulent near-surface winds, more flexible surface-drag directions, and stronger wind gusts. These findings suggest a simulated tornado is very sensitive to the modeling approach of near-surface turbulence.

The Time Scale of Shallow Convective Self-Aggregation in Large-Eddy Simulations Is Sensitive to Numerics.

Numerical simulations of the tropical mesoscales often exhibit a self-reinforcing feedback between cumulus convection and shallow circulations, which leads to the self-aggregation of clouds into large clusters. We investigate whether this basic feedback can be adequately captured by large-eddy simulations (LESs). To do so, we simulate the non-precipitating, cumulus-topped boundary layer of the canonical "BOMEX" case over a range of numerical settings in two models. Since the energetic convective scales underpinning the self-aggregation are only slightly larger than typical LES grid spacings, aggregation timescales do not converge even at rather high resolutions (<100m). Therefore, high resolutions or improved sub-filter scale models may be required to faithfully represent certain forms of trade-wind mesoscale cloud patterns and self-aggregating deep convection in large-eddy and cloud-resolving models, and to understand their significance relative to other processes that organize the tropical mesoscales.

Influence of grid resolution of large-eddy simulations on foehn-cold pool interaction.

Numerical simulations are performed to assess the influence of horizontal and vertical model resolution on the turbulent erosion of a cold-air pool (CAP) by foehn winds in an Alpine valley near Innsbruck, Austria. Strong wind shear in the transition zone from the CAP to the overlying foehn generates turbulence by shear-flow instability and contributes to the CAP erosion. The sensitivity of this process to grid resolution in the "grey zone" of turbulence is studied with the Weather Research and Forecasting model in large-eddy simulation (LES) mode with a horizontal grid spacing of 200, 40, and 13.33 m and in mesoscale mode with a grid spacing of 1 km. Moreover, two different vertical resolutions are tested. The mesoscale simulation exhibits deficiencies in the CAP development and is neither able to resolve nor parametrize the effect of Kelvin-Helmholtz (K-H) instability. In contrast, the LES with the coarsest horizontal grid spacing begins to explicitly permit K-H instability, albeit individual K-H waves are not completely resolved, and thereby greatly improves the stability and wind profile of the foehn. Refining the LES grid spacing leads to a more explicit and realistic representation of turbulence, but surprisingly has little impact on mean quantities. An increase in the vertical resolution shows the greatest benefit in the turbulent upper part of the foehn jet, whereas an increase in the horizontal resolution improves the turbulence characteristics, especially at the foehn-CAP interface. However, spectral analysis indicates that even a horizontal grid spacing of 40 m does not fully capture the energy cascade in the inertial subrange. Eddies remain too large and foehn-CAP interaction is too vigorous compared with the simulation with 13.33 m grid spacing. Nevertheless, results illustrate the potential benefit of an (100 m) model resolution for improving numerical weather predictions in complex terrain.

Supersaturation Variability from Scalar Mixing: Evaluation of a New Subgrid-Scale Model Using Direct Numerical Simulations of Turbulent Rayleigh–Bénard Convection

Abstract Supersaturation fluctuations in the atmosphere are critical for cloud processes. A nonlinear dependence on two scalars—water vapor and temperature—leads to different behavior than single scalars in turbulent convection. For modeling such multiscalar processes at subgrid scales (SGS) in large-eddy simulations (LES) or convection-permitting models, a new SGS scheme is implemented in CM1 that solves equations for SGS water vapor and temperature fluctuations and their covariance. The SGS model is evaluated using benchmark direct-numerical simulations (DNS) of turbulent Rayleigh–Bénard convection with water vapor as in the Michigan Tech Pi Cloud Chamber. This idealized setup allows thorough evaluation of the SGS model without complications from other atmospheric processes. DNS results compare favorably with measurements from the chamber. Results from LES using the new SGS model compare well with DNS, including profiles of water vapor and temperature variances, their covariance, and supersaturation variance. SGS supersaturation fluctuations scale appropriately with changes to the LES grid spacing, with the magnitude of SGS fluctuations decreasing relative to those at the resolved scale as the grid spacing is decreased. Sensitivities of covariance and supersaturation statistics to changes in water vapor flux relative to thermal flux are also investigated by modifying the sidewall conditions. Relative changes in water vapor flux substantially decrease the covariance and increase supersaturation fluctuations even away from boundaries.

Effects of Horizontal Resolution, Domain Size, Boundary Conditions, and Surface Heterogeneity on Coarse LES of a Convective Boundary Layer

Abstract Atmospheric properties in a convective boundary layer vary over a wide range of spatial scales and are commonly studied using large-eddy simulations (LES) in various configurations. We examine how the boundary layer depth and distribution of variability across scales are affected by LES grid spacing, domain size, inhomogeneity of surface properties, and external forcing. Two different setups of the Weather Research and Forecasting (WRF) model are analyzed. A semi-idealized configuration uses a periodic domain, flat surface, prescribed homogeneous surface heat fluxes, and horizontally uniform profiles of large-scale advective tendencies. A nested LES setup employs a larger domain and realistic initial and boundary conditions, including an interactive land surface model with representative topography and vegetation and soil types. Subdomains of identical size are analyzed for all simulations. Characteristic structure sizes are quantified using the variability scales L50 and L95, defined such that features smaller than that contain 50% and 95% of the total variance, respectively. Progressive increase in L50 from vertical velocity to temperature and moisture structures is systematically reproduced in all simulation configurations. This dependence of L50 on the considered variable complicates the development of scale-aware parameterizations for models with grid spacing in the “terra incognita”. In simulations using a larger domain with heterogeneous surface properties, the development of internal mesoscale patterns significantly affects variance distributions inside analyzed subdomains. Sizes of boundary layer structures also strongly depend on the LES grid spacing and, in case of heterogeneous surface and topography, on location of the subdomain inside a larger computational domain.

Large-Eddy Simulation of a wind turbine using a Filtered Actuator Line Model

When dealing with multirotor devices such as quadcopters or wind farms, the cost of blade-resolved large-eddy simulation (LES) becomes prohibitive. Combining LES with a family of lower-fidelity models, called actuator line models (ALMs), has grown in popularity in the past decade. ALM replaces full blade resolution with an array of actuator points or lines parameterized by aerodynamic lift/drag polar plots along the blades. Body forces computed based on these actuator points are then projected onto the LES flow mesh, mimicking the effect of rotating blades on the flow. However, the optimal projection radius and the associated LES grid size is often too restrictive for multirotor simulations. Recently, a new tip-correction-based filtered ALM (F-ALM) was proposed by Martínez-Tossas and Meneveau (2019), which allows coarser-than-optimal grids by avoiding the associated overprediction of thrust. In this work, F-ALM is implemented into a high-order, in-house LES code to simulate National Renewable Energy Laboratory Phase VI wind turbine. It is then followed by a comparison between the baseline ALM and the newly implemented F-ALM in terms of instantaneous and time-averaged flow fields and blade loads, revealing the advantages of F-ALM in preventing the overprediction of power on coarse grids. This encourages accurate and affordable simulations of multirotor devices in the future.

Resolved and subgrid-scale crossing trajectory effects in Eulerian large eddy simulations of size-dependent droplet transport

We study the dispersion characteristics of slightly buoyant droplets in a turbulent jet using large eddy simulations (LES). The droplet number density fields are represented using an Eulerian approach, with the dispersed phase modelled using the Fast-Eulerian method (Ferry &amp; Balachandar, Intl J. Multiphase Flow, vol. 27, issue 7, pp. 1199–1226, 2001) that includes the droplet rise velocity. Radial concentration profiles and turbulent concentration fluxes for droplets of different sizes are analysed to quantify the ‘trajectory crossing effect’, when relative motions between particles and turbulent eddies tend to reduce turbulent diffusion. For finer LES grid resolutions, the model captures the differential, size-based dispersion characteristics of the droplets with the transverse dispersion of the larger droplet sizes suppressed, since trajectory crossing effects are explicitly resolved in LES. We examine a similarity solution model for the size-dependent radial concentration profiles based on a modified Schmidt number derived from the theory of turbulent diffusion of particles in the atmosphere proposed by Csanady (J. Atmos. Sci., vol. 20, pp. 201–208, 1963). The results are validated with the high resolution LES data and show good agreement. Then the size-dependent Schmidt number model is reformulated as a model for unresolved subgrid-scale trajectory crossing effects and used to calculate the subgrid concentration flux in a coarse LES of a turbulent jet, with slightly buoyant droplets injected at the centreline in the self-similar region of the jet. The results are compared to a simulation with higher grid resolution and a coarse simulation with a constant Schmidt number subgrid-scale model. We find that the subgrid model enhances the prediction accuracy of the concentration profiles and turbulent concentration flux for the coarse LES.

The thickened flame approach for non-premixed combustion: Principles and implications for turbulent combustion modeling

Modeling turbulent non-premixed combustion remains a challenge in the context of Large Eddy Simulation (LES) in complex geometries and for realistic conditions, taking into account all physical phenomena impacting the flame such as heat loss, dilution, or liquid fuel atomization and evaporation. In this work, the Thickened Flame concept, which allows to resolve the flame front on the LES grid while preserving the consumption speed, and initially derived for premixed combustion, is adapted to diffusion flames. It is demonstrated that the concept holds for these flames, with however, a different formulation of the model due to but their specific nature and properties. In particular, in the high-Damköhler regime, the thickening factor is applied only to the diffusion coefficients. The behavior of thickened diffusion flames is illustrated on laminar steady strained flames for both simple and complex chemistry, showing how the Thickened Flame concept applies. Based on these results, an expression for the thickening factor related to mesh coarsening is derived. For a complete turbulent combustion model, the thickening factor should also describe the sub-grid scale flame-turbulence interaction, which is left for future work.

CFD simulations of two opposing plane wall jets in a generic empty airplane cabin: Comparison of RANS and LES

Accurate computational fluid dynamics (CFD) simulations of the ventilation flow in airplane cabins are important for ventilation design optimisation regarding passenger health, thermal comfort and energy efficiency. The complex unsteady flow phenomena inside the cabin are challenging in numerical modelling and therefore dedicated validation of the simulation outcomes is required. The majority of airplane ventilation studies performed steady Reynolds-averaged Navier-Stokes (RANS) simulations and focused mainly on global flow field characteristics. This paper presents both steady RANS and transient simulations, including unsteady RANS and large eddy simulations (LES), of the flow driven by opposing plane wall jets in a reduced-scale isothermal water-filled generic airplane cabin. A detailed analysis of the mean velocity and turbulence characteristics (including Reynolds stresses) of the cabin flow is outlined, encompassing a thorough investigation of the fundamental flow components such as the opposing jets and the merged jet. Specific attention is devoted to LES grid sensitivity and the performance of different RANS turbulence and LES subgrid-scale (SGS) models is assessed by comparison with particle image velocimetry (PIV) measurements. It is shown that LES in general performs much better than RANS, the latter being incapable of providing accurate mean flow characteristics within the interaction zone and merged jet due to the underlying dynamic opposing jet interaction. Unsteady RANS partly covers the unsteadiness, however, turbulence levels remain systematically underpredicted and LES is required for a correct representation. Differences between the LES SGS model predictions are limited to the SGS kinetic energy budget of the total turbulent kinetic energy.

A dynamic spatial gradient model for the subgrid closure in large-eddy simulation of turbulence

A dynamic spatial gradient model (DSGM) is proposed for the subgrid-scale (SGS) closure of large-eddy simulation (LES). The velocity gradients at neighboring LES grids are incorporated to improve the accuracy of the SGS stress. Compared to the previous machine-learning-based multi-point gradient models, the current model is free from the need of a priori knowledge. The model coefficients are dynamically determined by the least-square method using the Leonard stress. The a priori tests show that the correlation coefficients of the SGS stress for the DSGM framework are much larger than the traditional velocity gradient model over different tested filter widths from viscous to inertial scales. The analysis of the model coefficients in the a priori test suggests that the number of the model coefficients can be significantly reduced, leading to a simpler version of the model. A small-scale eddy viscosity (SSEV) model is introduced as an artificial viscosity to mimic the flux of kinetic energy to smaller scales which cannot be resolved at an LES grid. The velocity spectrum predicted by SSEV-based implicit LES is very close to that of direct numerical simulation (DNS) data. In the a posteriori tests, both the flow statistics and the instantaneous field are accurately recovered with the SSEV-enhanced DSGM model. Compared with the SSEV-based implicit LES, the dynamic Smagorinsky model, and the dynamic mixed model, the results predicted by the current model have overall closer agreements with the filtered DNS result, suggesting that the DSGM framework is well-suited for highly accurate LES of turbulence.