Subgrid-scale Model Research Articles

Analysis of Channel Vortex and Cavitation Performance of the Francis Turbine under Partial Flow Conditions

To realize a multienergy complementary system involving hydropower and other energy sources, hydraulic turbines frequently run under partial flow conditions in which a unique flow phenomenon, the channel vortex, occurs in the runner, causing fatigue failure and even cavitation to the turbine blade. Cavitation severely shortens the service life of the unit and terribly limits the output of the turbine under partial flow conditions. In this paper, a numerical model of a Francis turbine was created with tetrahedral grids; the large eddy simulation (LES) method based on the WALE subgrid scale model and the Schnerr–Sauer cavitation model was adopted to carry out numerical simulation of the Francis turbine; and a vortex identification method based on the Q criterion was used to capture and analyze the channel vortex. The calculation results showed that a negative impact angle at the inlet of the runner occurred when the turbine ran under partial flow conditions, leading to three different types of channel vortexes in the blade channel. Also, different channel vortexes caused cavitation on different positions on the runner, and the volume change of cavitation showed periodic properties.

Extracting turbulence parameters of smoke via video analysis

Turbulence can be observed in almost all the smoke videos, so turbulence parameters can be extracted as features of smoke in the video smoke detection system. A video-based turbulence parameter measuring method is proposed in this paper. Turbulence can be regarded as a series of irregular flow states with instantaneous velocity fluctuations. First, the instantaneous turbulent velocity field is obtained from the real-time video using an optical flow method, and based on the velocity field information, the turbulence kinetic energy (TKE) and dissipation rate are calculated using large eddy simulation and sub-grid scale (SGS) models. Three different methods, Smagorinsky model, gradient model, and dimensional analysis, are used to obtain the dissipation rate. The results obtained using different methods are compared. The dissipation rate estimated using the SGS models shows high consistency and that estimated by the dimensional analysis has a similar distribution, which shows that the choice of the SGS model is not important for the dissipation rate estimation. Then the TKE is calculated, and the processed images show that each plume structure has its own life cycle and the rising velocity of each plume structure is steady. This method is efficient for the plume turbulence measurement, and it provides benefits to the experimental study of fire plumes. Furthermore, the measured parameters can be analyzed for video-based fire detection. It is a useful method in both scientific research and industrial development.

Large eddy simulations of high-magnetic Reynolds number magnetohydrodynamic turbulence for non-helical and helical initial conditions: A study of two sub-grid scale models

Pseudo-spectral large eddy simulation (LES) calculations of high-magnetic Reynolds number (Rem) incompressible magnetohydrodynamic (MHD) turbulence are carried out for two initial conditions, namely, the non-helical Orszag–Tang vortex and the strongly helical Arnold–Beltrami–Childress (ABC) flows using two eddy-viscosity-based sub-grid scale (SGS) approaches: the cross-helicity (CH) and dynamic Smagorinsky (DS) models. Validation is conducted through comparisons of 1923 LES calculations with in-house 5123 direct numerical simulations (DNS) at Reynolds number, Re=Rem=800. The results show that the CH model performs better than the DS model. The performance of the SGS models at higher Re is further evaluated by carrying out 3843 LES calculations at Re=Rem=7500. Various quantities including turbulent kinetic energy, turbulent magnetic energy, cross-helicity, helicity, vorticity structures, cosine of angle between velocity and magnetic field, cosine of angle between velocity and vorticity field, kinetic and magnetic energy spectra, and energy fluxes are analyzed to understand the capability of the two LES models in predicting the evolution of MHD turbulence. The higher Reynolds number flow shows a delay in the maximum dissipation with increased transfer of energy toward small scales, resulting in a −5/3 Kolmogorov inertial sub-range scaling. In addition, the effect of Reynolds number on the alignment between velocity and magnetic field, and the energy transfer between kinetic and magnetic energy, is studied. With the ABC flow having strong helicity and zero cross-helicity at low and high Reynolds numbers, a strong dynamo effect is also observed using the LES models, which is consistent with previous DNS.

Dynamic iterative approximate deconvolution models for large-eddy simulation of turbulence

Dynamic iterative approximate deconvolution (DIAD) models with Galilean invariance are developed for subgrid-scale (SGS) stress in the large-eddy simulation (LES) of turbulence. The DIAD models recover the unfiltered variables using the filtered variables at neighboring points and iteratively update model coefficients without any a priori knowledge of direct numerical simulation (DNS) data. The a priori analysis indicates that the DIAD models reconstruct the unclosed SGS stress much better than the classical velocity gradient model and approximate deconvolution model with different filter scales ranging from viscous to inertial regions. We also propose a small-scale eddy viscosity (SSEV) model as an artificial dissipation to suppress the numerical instability based on a scale-similarity-based dynamic method without affecting large-scale flow structures. The SSEV model can predict a velocity spectrum very close to that of DNS data, similar to the traditional implicit large-eddy simulation. In the a posteriori testing, the SSEV-enhanced DIAD model is superior to the SSEV model, dynamic Smagorinsky model, and dynamic mixed model, which predicts a variety of statistics and instantaneous spatial structures of turbulence much closer to those of filtered DNS data without significantly increasing the computational cost. The types of explicit filters, local spatial averaging methods, and initial conditions do not significantly affect the accuracy of DIAD models. We further successfully apply DIAD models to the homogeneous shear turbulence. These results illustrate that the current SSEV-enhanced DIAD approach is promising in the development of advanced SGS models in the LES of turbulence.

Numerical assessment of LES subgrid-scale turbulence models for expandable particles in fire suppression

Numerical assessment of LES subgrid-scale turbulence models for expandable particles in fire suppression

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.

COMPARISON OF SUB-GRID-SCALE MODELS BASED ON PREDICTION OF COHERENT STRUCTURES IN PLANE TURBULENT JETS

A compressible plane jet at Mach 0.9 is numerically studied by large-eddy simulation. The governing equations are discretized by fourth-order spatial and third-order temporal schemes. Five sub-grid-scale (SGS) models, namely, the standard Smagorinsky model (SM), the coherent structure kinetic model (CKM) and the selective mixed-scale model (SMSM), the localized dynamic Smagorinsky model (LDSM), and the coherent-structure Smagorinsky model (CSM), are employed for closure of the sub-grid scale (SGS) terms, respectively, and compared. Proper orthogonal decomposition (POD) method is applied to extract the leading modes of the fluctuating velocity components, i.e., \begin{document}$u'$\end{document} in the streamwise, \begin{document}$v'$\end{document} in the lateral and \begin{document}$w'$\end{document} in the spanwise. The averaged flow fields, dissipation, the instantaneous vortical structures and the coherent structures represented by the leading POD modes, are compared. The leading POD modes of \begin{document}$u'$\end{document} are two longitudinal stripes with fracted contours in the turbulent region, representing the multi-scaled flow and the decay of fluctuation strength. The leading modes of \begin{document}$v'$\end{document} are a row of ribs with the lateral size growing along the streamwise, while the modes of \begin{document}$\left( {u', v'} \right)$\end{document} are in a circular flow pattern around the ends of the ribs. The circular pattern penetrates both the jet flow and the peripheral flow, representing the flow entrainment. The leading modes of \begin{document}$w'$\end{document} are a train of ridges along streamwise. Their positive and negative values represent the spanwise stretching pattern of the coherent structures. In comparison of these five SGS models, the instantaneous multi-scale vortical flow is not effectively predicted by the SM and the CKM, as the small vortical scale is smeared. The POD modes are found to be sensitive to the sub-grid dissipation, since the valley regions of the mode of \begin{document}$u'$\end{document} coincide with the peak dissipation regions predicted by the CKM. The circular flow pattern is also not clearly predicted by the CKM. Neither the ridge pattern of the mode of \begin{document}$w'$\end{document} is predicted by the SM. On the other hand, the multi-scales turbulent flow and the flow patterns of POD modes are well predicted by the CSM, the SMSM, and the LDSM, in which the CSM is computationally more efficient.

Structural subgrid scale model based on wavelet filter for large eddy simulation of particle-laden turbulence

Coherent vortex structures in subgrid scale (SGS) motions are important for preferential concentration and collision of particles with small and intermediate Stokes number (Xiong et al., 2019). In this study, a new large eddy simulation (LES) strategy that combines wavelet filtered large eddy simulation with a differential filter SGS model is proposed. First, a wavelet filtered direct numerical simulation (WFDNS) enables good preservation of SGS structures due to the high compressibility and local fidelity of the wavelet filter. Second, a differential filter (DF) model, containing only a parameter related to the nominal filter width, is used to dynamically reconstruct unresolved eddies. The SGS model presented here is verified using direct numerical simulation (DNS) data for particle-laden homogenous isotropic turbulence. Compared to the classical spectral-filtered DNS (FDNS) model, the new SGS model enables to achieve better agreement with DNS results related to the dispersed-phase statistics, such as particle acceleration, particle-seen fluid kinetic energy, particle-seen Lagrangian integral time etc. Furthermore, obtained results also exhibit the advantages of a proposed model over the stochastic Langevin model in the prediction of particle dynamics, especially collision-related statistics such as radial distribution function, radial relative velocities, and particle collision rates. Moreover, the model also demonstrates good performance in application to particle-pair related statistics at different Reynolds numbers and different filter depths. This is attributed to the ability of this model to multispectrally enhance coherent vortex structures in SGS, which is promising to apply for recovering SGS effects in the LES of turbulent particle-laden flows.

Topological investigation of junction flow between cylinder and flat plate

The paper presents a numerical investigation on the junction flow occurring at the intersection between a wall and a protruding circular cylinder. Simulations of flow for Reynolds number ( Re) in the range [125 - 20000] have been carried out using OpenFOAM, an open source CFD tool. Plots of stream tracers have been used to qualitatively characterize the flow topology as either attachment or separation based on the type of singular points, classified as nodes and saddle points. Quantitative variations of momentum flux have been applied to identify a key mechanism for the transition of topology using the relative momentum strength of the incoming and reverse flow. Effect of a thinner boundary layer has been assessed by (i) imposing a reduced wall shear stress, and (ii) increasing the Reynolds number. Features of a typical unsteady flow, in the transition regime, at Re = 20000 have been predicted using Large Eddy Simulation (LES) with a one-equation eddy viscosity sub-grid scale model. Description of the time evolution of the topology at Re = 20000 has been able to validate the one-equation model in OpenFoam as well as to further validate the key mechanism identified for topology transition.

A novel buoyancy-modified subgrid-scale model for large-eddy simulation of turbulent convection

PurposeThe purpose of this paper is to develop a subgrid-scale (SGS) model for large eddy simulation (LES) of buoyancy- and thermally driven transitional and turbulent flows and further examine its performance.Design/methodology/approachFavre-filtered, non-dimensional LES equations are solved using non-dissipative, fully implicit, kinetic energy conserving, finite-volume algorithm which uses an iterative predictor-corrector approach based on pressure correction. Also, to develop a new SGS model which accounts for buoyancy, turbulent generation term in SGS viscosity is properly modified and enhanced by buoyancy production.FindingsThe proposed model has been successfully applied to turbulent Rayleigh–Bénard convection. The results show that the model is able to reproduce the complex physics of turbulent thermal convection. In comparison with the original wall-adapting local eddy-viscosity (WALE) and buoyancy-modified (BM) Smagorinsky models, turbulent diagnostics predicted by the new model are in better agreement with direct numerical simulation.Originality/valueA BM variant of the WALE SGS model is newly developed and analyzed.

Subgrid scale modeling considerations for large eddy simulation of supercritical turbulent mixing and combustion

This paper presents a systematic investigation of large eddy simulation (LES) and subgrid scale (SGS) modeling with application to transcritical and supercritical turbulent mixing and combustion. There remains uncertainty about the validity of extending the LES formalism developed for low-pressure, ideal-gas flows to simulations of high-pressure real-fluid flows. To address this concern, we reexamine the LES theoretical framework and the underlying assumptions in the context of real-fluid mixing and combustion. Two-dimensional direct numerical simulations of nonreacting and reacting mixing layers of gaseous methane and liquid oxygen in the thermodynamically transcritical and supercritical fluid regimes are performed. The computed results are used to evaluate the exact terms in the LES governing equations and associated SGS models. Order of magnitude analysis of the exact filtered and subgrid terms in the LES equations and a priori analysis of the simplifications are performed at different filter widths. It is shown that several of these approximations do not hold for supercritical turbulent mixing. Subgrid scale terms, which are neglected in the LES framework for ideal-gas flows, become significant in magnitude compared to the other leading terms in the governing equations. In particular, the subgrid term arising from the filtering of the real-fluid equation of state is shown to be important.

Intercomparison of Subgrid Scale Models in Large-Eddy Simulation of Sunset Atmospheric Boundary Layer Turbulence: Computational Aspects

Intercomparison of Subgrid Scale Models in Large-Eddy Simulation of Sunset Atmospheric Boundary Layer Turbulence: Computational Aspects

Mixing matters

ABSTRACT All hydrodynamical simulations of turbulent astrophysical phenomena require sub-grid scale models to properly treat energy dissipation and metal mixing. We present the first implementation and application of an anisotropic eddy viscosity and metal mixing model in Lagrangian astrophysical simulations, including a dynamic procedure for the model parameter. We compare these two models directly to the common Smagorinsky and dynamic variant. Using the mesh-free finite mass method as an example, we show that the anisotropic model is best able to reproduce the proper Kolmogorov inertial range scaling in homogeneous, isotropic turbulence. Additionally, we provide a method to calibrate the metal mixing rate that ensures numerical convergence. In our first application to cosmological simulations, we find that all models strongly impact the early evolution of galaxies, leading to differences in enrichment and thermodynamic histories. The anisotropic model has the strongest impact, with little difference between the dynamic variant and the constant-coefficient variant. We also find that the metal distribution functions in the circumgalactic gas are significantly tighter at all redshifts, with the anisotropic model providing the tightest distributions. This is contrary to a recent study that found metal mixing to be relatively unimportant on cosmological scales. In all of our experiments, the constant-coefficient Smagorinsky and anisotropic models rivalled their dynamic counterparts, suggesting that the computationally inexpensive constant-coefficient models are viable alternatives in cosmological contexts.

Large-eddy simulations of the flow over a semi-circular cylinder at Re = 50000

Large-eddy simulations (LES) of the flow over a semi-circular cylinder at a diameter-based Reynolds number 50000 and zero angle of attack are presented. In comparison with the prominent benchmark flows past circular and square cylinders, this test problem has been less investigated in the literature from both numerical and experimental perspectives. Also, this case can be even more challenging for validation and verification of the computational codes due to a negative lift force, which characterizes this flow regime. A standard numerical platform based on a second-order finite volume method and the k-equation eddy viscosity subgrid scale model and its dynamic version, implemented in the OpenFOAM CFD toolbox, is used for the present LES. A detailed spectral analysis is provided to examine dynamics of the vortex shedding and shear layer instabilities. The moderate Reynolds number is chosen in order to replicate experimental data, which is limited to integral flow parameters like the lift, drag and pressure coefficients and a Strouhal number. The numerical and experimental data available for the flows over the circular and square cylinders at the Reynolds number 20000−50000 are used for qualitative assessment of the present results. Similarity of the turbulent wakes between the circular, square and semi-circular cylinders are discussed and compared both, for the integral flow parameters and high order statistics as well. Overall, present simulations provide reasonable agreement with available measurements and numerical calculations. Consistency and satisfactory agreement with the previous numerical and experimental results obtained for the circular and square cylinders are shown as well.

Development of a large-eddy simulation subgrid model based on artificial neural networks: a case study of turbulent channel flow

Abstract. Atmospheric boundary layers and other wall-bounded flows are often simulated with the large-eddy simulation (LES) technique, which relies on subgrid-scale (SGS) models to parameterize the smallest scales. These SGS models often make strong simplifying assumptions. Also, they tend to interact with the discretization errors introduced by the popular LES approach where a staggered finite-volume grid acts as an implicit filter. We therefore developed an alternative LES SGS model based on artificial neural networks (ANNs) for the computational fluid dynamics MicroHH code (v2.0). We used a turbulent channel flow (with friction Reynolds number Reτ=590) as a test case. The developed SGS model has been designed to compensate for both the unresolved physics and instantaneous spatial discretization errors introduced by the staggered finite-volume grid. We trained the ANNs based on instantaneous flow fields from a direct numerical simulation (DNS) of the selected channel flow. In general, we found excellent agreement between the ANN-predicted SGS fluxes and the SGS fluxes derived from DNS for flow fields not used during training. In addition, we demonstrate that our ANN SGS model generalizes well towards other coarse horizontal resolutions, especially when these resolutions are located within the range of the training data. This shows that ANNs have potential to construct highly accurate SGS models that compensate for spatial discretization errors. We do highlight and discuss one important challenge still remaining before this potential can be successfully leveraged in actual LES simulations: we observed an artificial buildup of turbulence kinetic energy when we directly incorporated our ANN SGS model into a LES simulation of the selected channel flow, eventually resulting in numeric instability. We hypothesize that error accumulation and aliasing errors are both important contributors to the observed instability. We finally make several suggestions for future research that may alleviate the observed instability.

Mixing and combustion at low heat release in large eddy simulations of a reacting shear layer

In this paper, the mixing and combustion at low-heat release in a turbulent mixing layer are studied numerically using large eddy simulation. The primary aim of this paper is to successfully replicate the flow physics observed in experiments of low-heat release reacting mixing layers, where a duty cycle of hot structures and cool braid regions was observed. The nature of the imposed inflow condition shows a dramatic influence on the mechanisms governing entrainment, and mixing, in the shear layer. An inflow condition perturbed by Gaussian white noise produces a shear layer which entrains fluid through a nibbling mechanism, which has a marching scalar probability density function where the most probable scalar value varies across the layer, and where the mean-temperature rise is substantially over-predicted. A more sophisticated inflow condition produced by a recycling and rescaling method results in a shear layer which entrains fluid through an engulfment mechanism, which has a non-marching scalar probability density function where a preferred scalar concentration is present across the thickness of the layer, and where the mean-temperature rise is predicted to a good degree of accuracy. The latter simulation type replicates all of the flow physics observed in the experiment. Extensive testing of subgrid-scale models, and simple combustion models, shows that the WALE model coupled with the Steady Laminar Flamelet model produces reliable predictions of mixing layer diffusion flames undergoing with fast chemistry.

Computational efficiency of CFD modeling for building engineering: An empty domain study

In computational wind engineering, Large Eddy Simulation (LES) operates with superior accuracy. However, this comes with a higher computational cost compared to Reynolds Averaged Navier-Stokes (RANS) models. We propose multiple strategies to reduce the computational cost without compromising accuracy. The turbulence synthesis approach is adopted for efficient inflow generation. To alleviate the issues with tedious and resource-consuming computations, the concept of ‘interpolation’ is endorsed. This approach allows the usage of identical inflow information for different arrangements of meshing. The effect of employing larger time-step and coarser mesh, on the performance of LES with wall functions, is also investigated. Several LES cases are studied, initially with special near-wall treatment, and later, with three different Subgrid Scale (SGS) models. The profiles of mean velocities, turbulence intensities, and integral length scales are examined, in addition to the velocity spectra. The horizontal homogeneity is assessed for all cases. The results show that the wall-adapting eddy viscosity (WALE) SGS model with LES can reduce the computational time at a lower error range. Besides, among three investigated hybrid models, the Detached Eddy Simulation (DES) and the Delayed Detached Eddy Simulation (DDES) provide a significant reduction in the computational cost, compared to LES with the dynamic one-equation eddy-viscosity SGS model, when the objective is to achieve a similar level of accuracy.

Large-eddy simulation of flow around an isolated building: A step-by-step analysis of influencing factors on turbulent statistics

We designed and exclusively performed a wind tunnel experiment to obtain high-quality and high-resolution measurement results for three components of mean and fluctuating velocities around a 1: 1: 2 isolated building to establish a benchmark case for large-eddy simulation (LES) as a part of the activities of the Architectural Institute of Japan. Subsequently, we conducted LES to investigate the effect of computational grid arrangement on turbulent statistics by changing three grid arrangements. Finally, we performed a comparison by changing spatial discretization schemes, sub-grid scale (SGS) turbulence models, and convergence criteria. A grid in which the width of a building was discretized into 20 grids yielded sufficient resolution to reproduce the mean flow and second-order turbulent statistics. Under this appropriate grid arrangement, in the condition that the blending ratio of the first-order upwind interpolation scheme with the second-order linear interpolation scheme from 5 to 10%, the appropriate second-order statistics of wind velocity around the isolated building was obtained and numerical oscillation was effectively removed. Then, with this grid, any SGS model yielded appropriate results. In addition, the convergence criteria had little effects on the LES results if stable solution was obtained. Because the convergence criteria greatly affect the total calculation cost, weaker appropriate criteria should be set.

Investigation of aerodynamic noise reduction of exterior side view mirror based on bionic shark fin structure

The LES (Large Eddy Simulation) model based on different subgrid-scale models was used to calculate the aerodynamic noise of the side view mirror. The simulation result of Algebraic Wall-Modeled LES Model (WMLES) in LES was proved to be more accurate through the wind tunnel test. In order to explore the noise reduction mechanism of bionic shark fin, the shape of shark fin was studied and applied to the shell of side view mirror. The simulation results indicated that the bionic shark fin structure could effectively comb the streamline of the side view mirror surface and reduce the negative pressure area. The turbulence intensity at the rear of the side view mirror was reduced, which was beneficial to reduce the aerodynamic noise of the side view mirror on the front side window area. The maximum noise reduction could reach 7.3 dB.

A data-driven subgrid scale model in Large Eddy Simulation of turbulent premixed combustion

We derive a data-driven model of a subgrid scale (SGS) closure for turbulent premixed combustion in the context of Large Eddy Simulation (LES) using deep learning. We validate the model through LES of the direct numerical simulation (DNS) flame configuration and compare it to other subgrid models from the literature. The filtered DNS used as training data was provided by Lapeyre et al. (2019). The neural network developed in this study was designed to estimate the SGS flame surface density, using only local progress variable values as a basis. A priori tests show that the results inferred from the frozen neural network were comparable to results obtained from the convolutional neural networks (CNNs) using the full nonlocal set of variables, and were in good agreement with the filtered DNS. A model-agnostic method for interpreting machine learning was employed to investigate the behavior of the trained neural network. A posteriori evaluation using the network as an LES subgrid model demonstrates that the proposed data-driven modeling is more accurate than classical algebraic models in terms of the integrated flame area in the axial direction. This illustrates that the proposed data-driven subgrid model to represent the non-linear unresolved terms is a successful approximation from both an a priori and an a posteriori perspective and that only fully local values in the filtered domain suffice to yield good agreement with DNS results. This is in contrast to earlier attempts, which use the full LES domain dataset as input to the network.