DNS Database Research Articles

RANS representation of transition and separation over a low-Re number blade section at high angle of attack

Systems based on wind energy harvesting can successfully meet part of the increasing green energy demand worldwide. However, wind turbines operation might be undermined by varying atmospheric conditions, which could result in an increase of angle of attack and consequent onset of flow separation phenomena, especially at low Reynolds numbers. Such conditions are strongly influenced by blades geometry, and they negatively affect structural integrity and power output of wind turbines. For this reason, it is crucial to define a tool capable of swiftly allowing numerical investigations on different geometrical configurations to delay and mitigate flow separation occurrence. The present work aims at modelling laminar-turbulent transition and turbulent flow separation over a wind turbine blade section operating at angle of attack = 15°, Re = 66000 and Pr = 0.71 by means of a steady RANS approach. Turbulence is treated by means of the Transition SST k-ω and the Transition k-kL-ω models. The main aerodynamic and thermal coefficients are evaluated and compared against a high-order accurate DNS database for validation. The results highlight, for the present test case, a better capability of the Transition SST k-ω of perceiving the main thermo-fluid dynamic features of the separated flow over the blade section.

Novel drag and Nusselt number models based on direct numerical simulations of a bidisperse gas–solid system

AbstractFlow and heat transfer in a bidisperse gas–solid system with freely moving spheres are simulated by particle‐resolved direct numerical simulation (PR‐DNS). Gas–solid coupling is enforced by the direct‐forcing immersed boundary method. Compard with the DNS database, it is found that the existing polydisperse drag correction model developed from static systems combined with various monodisperse drag models underestimates the drag force on dynamic arrays of particles. The existing Nusselt number correction model developed from static systems combined with various monodisperse models overestimates the Nusselt number of dynamic arrays of particles. Sensitivity analysis indicates that the effects of the granular temperature on the drag force and Nusselt number are negligible. Novel polydisperse drag and Nusselt number models are derived based on the database. The advantages of the derived polydisperse drag and Nusselt number models are demonstrated and confirmed by comparing the results of the computational fluid dynamics–discrete element method using various drag and Nusselt number models with experimental or additional PR‐DNS data.

Data-based instantaneous conditional progress variable dissipation rate modeling for turbulent premixed combustion

Manifold-based turbulent combustion modeling reduces computational cost considerably by projecting the thermochemical state onto a lower-dimensional space of a few manifold variables and reconstructing the state from a set of manifold equations. In turbulent premixed combustion, a one-dimensional space is defined by the progress variable, and the progress variable dissipation rate appears as a parameter in the resulting one-dimensional manifold equations. Closure of the manifold equations requires a model for the dependence of this dissipation rate on the progress variable. Existing models for these profiles are often ad hoc, derived analytically with crude assumptions, and assumed to be spatially and temporally ‘universal’ — yet the resulting thermochemical state can be sensitive to these profiles. To evaluate the assumptions in traditional models, instantaneous conditional progress variable dissipation rate profiles are extracted from DNS databases of turbulent premixed jet flames. At low Karlovitz number, the instantaneous conditional progress variable dissipation rate profiles exhibit some variance but are relatively well-characterized by a ‘universal’ model. However, a large variance in the instantaneous conditional progress variable dissipation rate profiles is observed at high Karlovitz number, indicating the form of the conditional dissipation rate is certainly not ‘universal’. To capture this variation at high Karlovitz number, a data-based model for the instantaneous conditional progress variable dissipation rate profiles is generated, and the resulting model predictions demonstrate excellent agreement with profiles extracted from DNS and far less error than ‘universal’ models. The sensitivity of the solutions of the one-dimensional premixed manifold equations to the dissipation rate profile is demonstrated a priori, and the error associated with the observed variance in the instantaneous conditional progress variable dissipation rate profiles (i.e., relative to the DNS mean) indicates that an instantaneous modeling approach is necessary. A variable importance study is conducted, and the conditional velocity gradient is identified as the critical quantity that characterizes the observed variance in the instantaneous conditional progress variable dissipation rate profiles, providing further motivation for a data-based modeling approach.

Machine learning for the prediction of the local skin friction factors and Nusselt numbers in turbulent flows past rough surfaces

Turbulent flows past rough surfaces can create substantial energy losses in engineering equipment. During the last decades, developing accurate correlations to predict the thermal and hydrodynamic behavior of rough surfaces has proven to be a difficult challenge. In this work, we investigate the applicability of convolutional neural networks to perform a direct image-to-image translation between the height map of a rough surface and its detailed local skin friction factors and Nusselt numbers. Additionally, we propose the usage of separable convolutional modules to reduce the total number of trainable parameters, and PReLU activation functions to increase the expressivity of the neural networks created. Our final predictions are improved by a new filtering methodology, which is able to combine the results of multiple neural networks while discarding non-physical oscillations likely caused by over-fitting. The main study is based on a new DNS database formed by 80 flow cases at a friction Reynolds number of Reτ=180 obtained by applying random shifts to the Fourier spectrum of the grit-blasted surface originally scanned by Busse et al. (2015). The results show that machine learning can accurately predict the skin friction values and Nusselt numbers for a rough surface. A detailed comparison with existing correlations in the literature revealed that the maximum errors generated by deep learning were only 8.1% for the global skin friction factors Cf¯ and 2.9% for the Nusselt numbers Nu¯, whereas the best classical correlations identified reached errors of 24.9% and 13.5% for Cf¯ and Nu¯ respectively. The deep learning results also proved stable with respect to rough surfaces with abrupt changes in their roughness elements, and only presented a minor sensitivity with respect to variations in the dataset size.

A Priori Validation of a Multidimensional Turbulent Premixed Combustion Model for High-Pressure Bunsen Flames

ABSTRACT Recently, the authors proposed a model for the filtered reaction source term conditioned on Favre filtered progress variable in turbulent premixed flames based on filtering of one-dimensional laminar flame profiles. Multidimensional effects from flame propagation at oblique angles through LES filter volumes and from subgrid flame folding were considered by a Gaussian filter kernel and through a wrinkling factor, respectively. Using a DNS database (one-step Arrhenius chemistry, unity Lewis number and constant transport) of statistically flat turbulent flames at different turbulence intensities and at ambient pressure, the model had been validated a priori and some model constants were adapted. Here, the model is validated for a Bunsen flames at varying inlet velocity, turbulence level, turbulent length scale and at higher pressure. In the Bunsen flames at bar, regions of strong curvature reminiscent of Landau-Darrieus instability appear, causing a faster increase of the wrinkling factor at small filter sizes than for the lower pressure flames. The effect is captured by a simple extension of the wrinkling factor model. Very good agreement between model and DNS filtered source term distributions is achieved using this adapted combustion model.

Linear logistic regression with weight thresholding for flow regime classification of a stratified wake

• Build a DNS database with O(100) realizations for stratified wake. • Develop a machine learning tool to classify the flow regimes in a stratified wake. • The physics governing stratified wake is universal with respect to Re and Fr. • The classifier captures the universal physics and generalize well. A stratified wake has multiple flow regimes, and exhibits different behaviors in these regimes due to the competing physical effects of momentum and buoyancy. This work aims at automated classification of the weakly and the strongly stratified turbulence regimes based on information available in a full Reynolds stress model. First, we generate a direct numerical simulation database with Reynolds numbers from 10000 to 50000 and Froude numbers from 2 to 50. Order (100) independent realizations of temporally evolving wakes are computed to get converged statistics. Second, we train a linear logistic regression classifier with weight thresholding for automated flow regime classification. The classifier is designed to identify the physics critical to classification. Trained against data at one flow condition, the classifier is found to generalize well to other Reynolds and Froude numbers. The results show that the physics governing wake evolution is universal, and that the classifier captures that physics.

A Differential Subgrid Stress Model and Its Assessment in Large Eddy Simulations of Non-Premixed Turbulent Combustion

We present a new subgrid stress model for the large eddy simulation of turbulent flows based on the solution of transport equations for stress tensor components. The model was a priori term-by-term calibrated against an open DNS database on forced isotropic turbulence (Johns Hopkins University database). After that, it was applied in a large eddy simulation of non-premixed turbulent combustion. To demonstrate the impact of the new subgrid stress model on scalar fields, we excluded the backward effect of heat release on the subgrid stresses, considering an isothermal reaction (i.e., diluted mixture; the density variations associated with chemical heat release can be neglected) and a Burke–Schumann reaction sheet approximation. A periodic box filled with a homogeneous turbulent velocity field and a three-layer top-hat mixture fraction field was studied. Four simulations were performed in which a fixed model for mixture fraction and its variance was combined with either the proposed subgrid stress model or one of the standard models, including Smagorinsky, dynamic Smagorinsky and WALE. Qualitatively correct backscatter was observed in a simulation with the new model. The differences in the statistics of the mixture fraction and reactive component fields caused by the new subgrid stress model were analyzed and discussed. The importance of using an advanced subgrid stress model was highlighted.

Wavelet-Based Optical Flow For High-Resolution Velocimetry In Wall-Bounded Fluid Flows

The performance of a wavelet-based optical flow velocimetry (wOFV) algorithm in extracting high accuracy and high resolution velocity fields from tracer particle images in wall-bounded flows is as sessed. wOFV is first evaluated using a DNS database of a turbulent boundary layer, which served as ground truth synthetic data. The sensitivity of wOFV to the regularization parameter (λ) is quantified and results are compared to commercially available PIV crosscorrelation-based processing. Results on synthetic data indicate a slight preference towards under-regularization in the viscous sublayer region to minimise error. Nonetheless, tests on synthetic data reveal that wOFV can modestly outperform PIV in terms of accuracy for a broad λ range. For the synthetic dataset, wOFV revealed clear advantages over PIV in terms of vector resolution, being able to resolve 29 vectors in the viscous sublayer (versus 5 for PIV) and provide 13 vectors closer to the wall than PIV. wOFV was also applied to experimental data of a turbulent boundary layer near the leading edge of a side-wall quenching burner setup and compared to PIV. Good agreement between the two velocimetry techniques was found, with wOFV exhibiting superior vector resolution. Based on theoretical relations of the viscous sublayer, a physically justified λ weighting for wOFV was d etermined. With this setting, wOFV was able to resolve further and more accurately into the viscous sublayer. Analysis of the turbulent velocity fluctuations revealed some spurious results for PIV in close proximity to the wall, leading to significantly exaggerated turbulence intensity in the viscous sublayer region. wOFV did not exhibit this same effect, revealing that it is less biased by the presence of the wall and decreased image signal-to-noise ratio. These aspects show that wOFV can provide improved accuracy and reliability in resolving turbulent motion occurring in the vicinity of physical boundaries.

A Priori Analysis on Deep Learning of Filtered Reaction Rate

A filtered reaction rate model driven by deep learning is proposed and analyzed a priori in the context of large eddy simulation (LES). A deep artificial neural network (ANN) is trained on the explicitly filtered reaction rate source term extracted from a database comprised of turbulent premixed planar flame direct numerical simulations (DNSes) employing single-step chemistry. The filtered DNS database to be used for the training of the ANN covers a wide range of turbulence intensities and LES filter widths. An interpretation technique of deep learning is employed to search the principal input parameters in the high dimensional database to alleviate the model complexity. The deep learning filtered reaction rate model is then tested on the unseen filtered planar flames featuring untrained turbulence intensities and LES filter widths, in conjunction with another canonical type of flame configuration that it has not been trained on. The deep learning filtered reaction rate model achieves good agreement with the filtered DNS results and also provides a quantitatively accurate surrogate model when compared to existing algebraic models and other combustion models from the literature.

Studies on a swirling heptane spray flame by large-eddy simulation

Swirling spray flames are widely encountered in gas-turbine combustors, rocket and ram-jet engines and internal combustion engines. Large-eddy simulation (LES) with different sub-grid scale (SGS) stress and combustion models is frequently used for studying the combustion mechanism of liquid fuel and design of combustors. In this paper, LES with a statistical-moment combustion model was used for studying a swirling spray flame. The statistical results for velocity, temperature and species concentration were validated by both experimental results and DNS database, showing the reliability of the simulation results. The instantaneous results give the spiral structures of flames formed by swirl, which can intensify combustion.

Active and Passive Collection of SSH Key Material for Cyber Threat Intelligence

This article describes a system for storing historical forensic artifacts collected from SSH connections. This system exposes a REST API in a similar fashion as passive DNS databases, malware hash registries, and SSL notaries with the goal of supporting incident investigations and monitoring of infrastructure.

Implicit Subgrid-Scale Modeling of a Mach 2.5 Spatially Developing Turbulent Boundary Layer.

We employ numerically implicit subgrid-scale modeling provided by the well-known streamlined upwind/Petrov–Galerkin stabilization for the finite element discretization of advection–diffusion problems in a Large Eddy Simulation (LES) approach. Whereas its original purpose was to provide sufficient algorithmic dissipation for a stable and convergent numerical method, more recently, it has been utilized as a subgrid-scale (SGS) model to account for the effect of small scales, unresolvable by the discretization. The freestream Mach number is 2.5, and direct comparison with a DNS database from our research group, as well as with experiments from the literature of adiabatic supersonic spatially turbulent boundary layers, is performed. Turbulent inflow conditions are generated via our dynamic rescaling–recycling approach, recently extended to high-speed flows. Focus is given to the assessment of the resolved Reynolds stresses. In addition, flow visualization is performed to obtain a much better insight into the physics of the flow. A weak compressibility effect is observed on thermal turbulent structures based on two-point correlations (IC vs. supersonic). The Reynolds analogy ( vs. ) approximately holds for the supersonic regime, but to a lesser extent than previously observed in incompressible (IC) turbulent boundary layers, where temperature was assumed as a passive scalar. A much longer power law behavior of the mean streamwise velocity is computed in the outer region when compared to the log law at Mach 2.5. Implicit LES has shown very good performance in Mach 2.5 adiabatic flat plates in terms of the mean flow (i.e., and ). iLES significantly overpredicts the peak values of , and consequently Reynolds shear stress peaks, in the buffer layer. However, excellent agreement between the turbulence intensities and Reynolds shear stresses is accomplished in the outer region by the present iLES with respect to the external DNS database at similar Reynolds numbers.

Multiscale Analysis of Anisotropy of Reynolds Stresses, Subgrid Stresses and Dissipation in Statistically Planar Turbulent Premixed Flames

The characterisation of small-scale turbulence has been an active area of research for decades and this includes, particularly, the analysis of small-scale isotropy, as postulated by Kolmogorov. In particular, the question if the dissipation tensor is isotropic or not, and how it is related to the anisotropy of the Reynolds stresses is of particular interest for modelling purposes. While this subject has been extensively studied in the context of isothermal flows, the situation is more complicated in turbulent reacting flows because of heat release. Furthermore, the landscape of Computational Fluid Dynamics is characterised by a multitude of methods ranging from Reynolds-averaged to Large Eddy Simulation techniques, and they address different ranges of scales of the turbulence kinetic energy spectrum. Therefore, a multiscale analysis of the anisotropies of Reynolds stress, dissipation and sub-grid scale tensor has been performed by using a DNS database of statistically planar turbulent premixed flames. Results show that the coupling between dissipation tensor and Reynolds stress tensor is weaker compared to isothermal turbulent boundary layer flows. In particular, for low and moderate turbulence intensities, heat release induces pronounced anisotropies which affect not only fluctuation strengths but also the characteristic size of structures associated with different velocity components.

Turbulence Measurement of Two Velocity Components in a Boundary Layer by a Two Parallel Hot-Wire Anemometer

A novel hot-wire technique consisting of two parallel hot-wires is proposed. The gap between the two parallel wires is filled with an adhesive, which can be expected to improve the sensitivity to the flow angle at high velocity regions. In the present study, measurement accuracy of the two component velocity fluctuations is appraised experimentally in a flat-plate turbulent boundary layer. As a result, it was verified that the present probe can reproduce not only the DNS database but also the measurements obtained by the standard I- and X-type hot-wire probes. In addition, the proposed probe can be brought closer to the wall than the X-type one. However, it is observed that the turbulent intensity measurements in the streamwise direction are slightly reduced, and the power spectrum of the streamwise velocity fluctuations is attenuated accordingly in a high-frequency range. Thus, fluid-dynamical and thermal effects of the adhesive need to be investigated in more detail in the future.

Particle subgrid scale modeling in hybrid RANS/LES of turbulent channel flow at low to moderate Reynolds number

Particle dispersion in a periodic channel is studied using the elliptic relaxation hybrid RANS/LES (ER-HRL) model. This approach employs a four-equation linear eddy viscosity (LEV) model while in Reynolds Averaged Navier-Stokes (RANS) mode near the wall, and switches to the LES Smagorinsky dynamic model in the outer flow region. We perform a systematic analysis of the dispersion of six sets of particles having Stokes numbers St = 0.2, 1, 5, 15, 25, 125 at shear Reynolds numbers of Reτ=150, 590. To account for the effect of the unresolved scales on particle dispersion, a novel subgrid-scale model (SGS) is proposed based on the wall-normal RMS of the velocity transport equation. The ER-HRL model is validated against DNS and LES databases, with a globally good agreement. For higher Reynolds number i.e. Reτ = 590, the model, with a much coarser grid, outperforms the LES subgrid stochastic acceleration (LES-SSAM) approach.

Scalar structures around the mixing interface of a thermal mixing layer in multiscale-generated decaying turbulent flows visualized using DNS database

This study shows the results for the visualization of a scalar mixing layer in decaying turbulence produced by a multiscale-generated turbulence grid. For this purpose, a thermal mixing layer in the air is used as the scalar mixing layer. A fractal generating grid proposed in the previous study is used for the generation of multiscale-generated turbulence. In addition, a parameter for determining the grid shape is varied. The scalar field is then visualized using a database given by direct numerical simulations. The instantaneous scalar field, the instantaneous scalar fluctuation field, the instantaneous scalar variance field, and the scalar dissipation field are visualized, and the visualization results are discussed using the wake interaction length scale.

Triple Decomposition of Velocity Gradient Tensor in Compressible Turbulence

The decomposition of the local motion of a fluid into straining, shearing, and rigid-body rotation is examined in this work for a compressible isotropic turbulence by means of direct numerical simulations. The triple decomposition is closely associated with a basic reference frame (BRF), in which the extraction of the biasing effect of shear is maximized. In this study, a new computational and inexpensive procedure is proposed to identify the BRF for a three-dimensional flow field. In addition, the influence of compressibility effects on some statistical properties of the turbulent structures is addressed. The direct numerical simulations are carried out with a Reynolds number that is based on the Taylor micro-scale of Reλ=100 for various turbulent Mach numbers that range from Mat=0.12 to Mat=0.89. The DNS database is generated with an improved seventh-order accurate weighted essentially non-oscillatory scheme to discretize the non-linear advective terms, and an eighth-order accurate centered finite difference scheme is retained for the diffusive terms. One of the major findings of this analysis is that regions featuring strong rigid-body rotations or straining motions are highly spatially intermittent, while most of the flow regions exhibit moderately strong shearing motions in the absence of rigid-body rotations and straining motions. The majority of compressibility effects can be estimated if the scaling laws in the case of compressible turbulence are rescaled by only considering the solenoidal contributions.

Spectral and modal energy transfer analyses of LES using the discontinuous Galerkin method and their application to the Variational Multiscale approach

In this work we perform a-priori analyses of the Discontinuous Galerkin (DG) Variational Multiscale (VMS) method for Large Eddy Simulation (LES). An analytical framework is introduced to study the ideal energy transfer between resolved and unresolved scales. The proposed framework is consistent with the discretization employed for the DG-LES simulations. The concept of modal eddy viscosity is also introduced which can be employed for the a-priori analysis of the DG-VMS method or spectral vanishing viscosity approaches. The developed framework is then applied to the analysis of the energy transfer in DG-LES by employing a DNS database of the Taylor-Green Vortex (TGV) at Re = 5000, 20000 and 40000. A-priori analyses are carried out for the Vreman [1] and all-all [2] variants of the DG-VMS approach. The performed analysis demonstrates that when the DG-LES resolution limit falls at the beginning of the dissipation range the assumption of large scales free of interaction with the unresolved scales is valid and the DG-VMS approach can replicate the ideal SGS dissipation spectrum. For coarser resolutions, typical of LES at high Reynolds numbers, the DG-VMS approach is unable to replicate the ideal energy transfer mechanism at the large-resolved scales. It is shown, a-priori, that a more accurate agreement can be obtained by employing a mixed Smagorinsky and DG-VMS approach with a fixed value of the scale-fraction parameter.

An extended flame index partitioning for partially premixed combustion

Compared to partially premixed combustion (or combustion of non-homogeneous reactants in general), fully premixed and diffusion flames represent only two asymptotic limits of combustion modes. However, the deep knowledge accumulated over the years on these two elementary and archetypal flame prototypes is such that they remain the cornerstone and reference building blocks of most combustion modelling proposals. Therefore, from a general point of view, being able to distinguish between premixed and non-premixed modes of combustion thanks to a flame index appears as a quite appealing but challenging task that still concentrates many research efforts. Indeed, the availability of such an index is not only appealing to proceed with the analysis of either experimental or computational data issued from DNS (or highly resolved LES) databases. It is also an essential ingredient to elaborate advanced flamelet-based multiregime combustion models on the basis of single regime tabulated flamelet databases. In the present study, a new definition of the premixedness index is proposed for partially premixed combustion. It is based on a weighted form of the cross-scalar dissipation rate of the mixture fraction and progress variable , i.e. quantities that have been previously identified as relevant parameters to describe partially premixed combustion regimes. The relevance of the corresponding index is assessed through a detailed computational procedure that includes three successive validation subsets: counterflow flames (including premixed, rich partially-premixed, and diffusion flames), (ii) stabilised triple flames for three distinct values of the inlet mixture fraction gradient, and finally (iii) unsteady flame kernel developments in non-homogeneous mixtures of fresh reactants, which are characterised by various initial levels of the segregation rate between the fuel and oxidiser. The proposed premixedness index and its counterpart are used as the weighting coefficients between tabulated premixed flamelets (TPF) and tabulated diffusion flamelets (TDF) data, which have been parameterised as functions of and . It is noteworthy that, in contrast to some previous proposals of the literature, the present flame index does not require the consideration of any other quantities in addition to those already used to parameterise the flamelets databases, i.e. and . The validation procedure makes use of steady and unsteady processes with a priori and a posteriori analyses. In both cases, the comparisons between the results obtained with the proposed flame partitioning and detailed chemistry (DC) computations lead to a satisfactory level of agreement and, from a general viewpoint, the level of agreement is better than the one obtained with either premixed or diffusion flamelet-based models.

On the validity of Damköhler's second hypothesis in statistically planar turbulent premixed flames in the thin reaction zones regime

The validity of Damköhler's second hypothesis, which was originally proposed for premixed combustion for conditions where the large scale turbulent timescale remains smaller than the chemical timescale and the integral length scale remains smaller than the flame thickness, has been assessed for a range of Damköhler and Karlovitz numbers (i.e. 0.2≤Da≤3.0 and 0.58≤Ka≤33.34) using a three-dimensional DNS database of statistically planar turbulent premixed flames subjected to forced unburned gas turbulence. It has been found that Damköhler's first hypothesis remains valid for all cases considered in this work irrespective of the combustion regime. By contrast, Damköhler's second hypothesis does not hold in the strict sense for the thin reaction zones regime flames with small values of Damköhler number. However, the ratio of the turbulent flame speed normalised by laminar burning velocity and the square root of the ratio of turbulent diffusivity to mass diffusivity converges to a value of the order of unity for the flames with high turbulence intensity and low Damköhler number in the thin reaction zones regime. Under this condition, Damköhler's second hypothesis, in an order of magnitude sense, can be considered to be valid alongside Damköhler's first hypothesis. A scaling analysis has been carried out using the equilibrium of the tangential strain rate and curvature terms in the Flame Surface Density transport equation to demonstrate that Damköhler's second hypothesis can be expected to be valid only in an order of magnitude sense for the thin reaction zones regime flames with small values of Damköhler number.