Validation of turbulence models in a simulated air-conditioning unit

Details are given of a study to obtain experimental data in an idealized environment for the purpose of evaluating the corresponding computational predictions and which supplement parallel measurements made in actual packaged air-conditioning units. The system consisted of a purpose-built low-speed wind tunnel with a working section constructed to reproduce particular features of the real units. In the experiment, both the mean velocity profiles and turbulence properties of the flow are obtained from triple-hot-wire anemometry measurements. A numerical model, based on finite volume methodology, was used to obtain the solution of the Reynolds-averaged Navier–Stokes equations for incompressible isothermal flow. The Reynolds stress terms in the equations are calculated using the standard k–ϵ model and second-moment closure (Reynolds stress) models. The accuracy of the two models was evaluated against the experimental measurements made 10 mm downstream of a baffle. The results show that the standard k–ϵ model gave the better agreement except in regions of strong recirculation. © 1997 John Wiley & Sons, Ltd.     

Predictions of axisymmetric and two-dimensional impinging turbulent jets

The k − turbulence model and a version of a second-moment closure, modified to include the effect of pressure reflections from a solid surface, have been used as the basis of predictions of the flow that results from the orthogonal impingement of circular and two-dimensional (2-D) jets on a flat surface. Comparison of model predictions has been made with velocity measurements obtained in the stagnation and wall jet regions of the impinging flows. Results, in general, confirm the superiority of the Reynolds stress transport equation model for predicting mean and fluctuating velocities within the latter regions of such flows. In particular, modifications to the second-moment closure to account for the influence of the surface in distorting the fluctuating pressure field away from the wall successfully predict the damping of normal-to-wall velocity fluctuations throughout the impinging flows. In contrast, results derived from the eddy-viscosity-based approach do not, in general, accurately reproduce experimental observations.     

LES Study of Turbulent Boundary Layer Over a Smooth and a Rough 2D Hill Model

Large eddy simulation (LES) is carried out to investigate the turbulent boundary-layer flows over a hill-shaped model with a steep or relatively moderate slope at moderately high Reynolds numbers (Re = O(10³)) defined by the hill height and the velocity at the hill height. The study focuses on the effects of surface roughness and curvature. For Sub-grid Scale (SGS) modeling of LES, both the dynamic Smagorinsky model (DSM) and the dynamic mixed model (DMM) are applied. The behavior of the separated shear layer and the vortex motion are affected by the oncoming turbulence, such that the shear layer comes close to the ground surface, or the size of a separation region becomes small because of the earlier instability of the separated shear layer. Appropriate measures are required to generate the inflow turbulence. The methods of Lund et al. (J. Comput. Phys., 140:233–258, 1998) and Nozawa and Tamura (J. Wind Eng. Ind. Aerodyn., 90:1151–1162, 2002; The 4th European and African Conference on Wind Engineering, 1–6, 2005) are employed to simulate the smooth- and rough-wall turbulent boundary layers in order to generate time-sequential data of inflow turbulence. This paper discusses the unsteady phenomena of the wake flows over the smooth and rough 2D hill-shaped obstacles and aims to clarify the roughness effects on the flow patterns and the turbulence statistics distorted by the hill. Numerical validation is conducted by comparing the simulation results with wind tunnel experiment data for the same hill shape at almost the same Re. The applicability of DSM and DMM are discussed, focusing on the recirculation region behind a steep hill.     

Application of a second-moment closure model to simulate the turbulent dispersion from an elevated line source

In this study, a second-moment closure model and an algebraic stress and flux model are used to simulate the turbulent dispersion from an elevated line source located in a turbulent boundary layer. In the modeling of the Reynolds stress equation, several models for the triple velocity correlation are examined by comparing their predicting results with the direct numerical simulation data. Predicted mean and fluctuating correlation quantities of velocity and temperature by the second-moment closure model and the algebraic model are compared with the experimental data. From the comparisons, it is found that the second-moment closure is capable of reproducing the experimental results satisfactorily. However, predictions of the algebraic stress and flux model are quite poor. Received on 12 June 1998     

Recent progress in nonlinear eddy-viscosity turbulence modeling

This article presents recent progresses in turbulence modeling in the Unit for Turbulence Simulation in the Department of Engineering Mechanics at Tsinghua University. The main contents include: compact Non-Linear Eddy-Viscosity Model (NLEVM) based on the second-moment closure, near-wall low-Re non-linear eddy-viscosity model and curvature sensitive turbulence model. The models have been validated in a wide range of complex flow test cases and the calculated results show that the present models exhibited overall good performance. The project supported by the National Natural Science Foundation of China (10232020), the National Key Basic Research Special Fund of China (2001CB409600) and the National High-Tech Development Program (2002AA311240)     

Second-moment closure modelling of turbulence in a non-inertial frame

This article investigates the performance of second-moment closures in a rotating reference frame and presents a new closure for the rapid pressure-strain rate correlation based on the recently developed materially frame indifference principle. It is observed here that the existing second-moment closures with appropriate near-wall treatment can adequately predict flows in a rotating channel and in an axially rotating pipe for moderate rotation rate. Analysis of the newly proposed model indicates that it is capable of reflecting flow features under strong rotation.     

Computational fluid dynamics applied to internal gas-turbine blade cooling: a review

This paper reviews current capabilities for predicting flow in the cooling passages and cavities of jet engines. Partly because of the need to enhance heat transfer coefficients, these flow domains entail complicated passage shapes where the flow is turbulent, strongly three-dimensional (3-D) and where flow separation and impingement, complicated by strong effects of rotation, pose severe challenges for the modeler. This flow complexity means that more elaborate models of turbulent transport are needed than in other areas of turbine flow analysis. The paper attempts to show that progress is being made, particularly in respect to the flow in serpentine blade-cooling passages. The first essential in modeling such flows is to adopt a low Reynolds number model for the sublayer region. The usual industrial practice of using wall functions cannot give a better than qualitative account of effects of rotation and curvature. It is shown that Rayleigh number effects can modify heat transfer coefficients in the cooling passages by at least 50%. The use of second-moment closure in the modeling is shown to be bringing marked improvements in the quality of predictions. Areas where, at present, more computational fluid dynamics (CFD) applications are encouraged are impingement cooling and pin-fin studies.     

On Modelling Periodic Motion with Turbulence Closures

In periodic flows near walls transport effects may be considerably larger than in a steady turbulent boundary layer. The question explored in this contribution is, therefore, whether providing separate transport equations for each of the Reynolds stresses consequently leads to a better modelling of a periodic flow's evolution than an eddy viscosity scheme whose constitutive equation is inherently linked to the generation and dissipation terms being in balance (local equilibrium). Our conclusion is that, while a stress-transport scheme is indeed better equipped to reproduce the phenomena, it does not consistently out-perform the EVM over the range of flows studied. In some cases it is suggested that more attention must be paid to modelling diffusive transport in order to secure the benefits of second-moment closure. To illustrate sensitivity to diffusive transport, two different diffusion models are tested, one of which leads to different effective transport coefficients in each stress component. This revised version was published online in July 2006 with corrections to the Cover Date.     

Assessment of Turbulence Models for Predicting the Interaction Region in a Wall Jet by Reference to LES Solution and Budgets

Highly-resolved LES and experimental data for a plane wall jet are used to study the characteristics of turbulence-closure proposals, mainly within the framework of second-moment-transport modelling. The study is motivated by the observed importance of diffusive Reynolds-stress transport in the interaction region between the outer shear layer and the near-wall layer of the wall jet, which gives the near-wall flow characteristics that are very different from those of a conventional boundary layer. Comparisons are presented for mean-flow quantities, second moments and budgets. Also included are a priori studies of approximations for the pressure-velocity interaction, pressure-fluctuation-driven transport and turbulent transport of the Reynolds stresses by triple correlations, the last observed to contribute significantly to the stress budgets. The study reveals major defects in the closure approximations for the pressure-velocity interaction terms, especially in the near-wall region. These defects result in a poor representation by the particular second-moment closures investigated of even the integral and mean-flow characteristics of the wall jet.     

Prediction of Flow and Heat Transfer through Stationary and Rotating Ribbed Ducts Using a Non-linear <Emphasis Type="Italic">k</Emphasis>−<Emphasis Type="Italic">ε</Emphasis> Model

The present paper deals with the prediction of three-dimensional fluid flow and heat transfer in rib-roughened ducts of square cross-section, which are either stationary, or rotate in orthogonal mode. The main objective is to assess how a recently developed variant of a cubic non-linear k−ε model (proposed by Craft et al. Flow Turbul Combust 63:59–80, 1999) can predict three-dimensional flow and heat transfer characteristics through stationary and rotating ribbed ducts. The present paper discusses turbulent air flow and heat transfer through two different configurations, namely: (I) a stationary square duct with “in-line” normal and (II) a square duct with normal ribs in a “staggered” arrangement under stationary and rotating conditions, with the axis of rotation normal to the flow direction and parallel to the ribs. In this paper the flow and thermal predictions of the linear k−ε model (EVM) are also included, as a set of baseline predictions. The mean flow predictions show that both linear and non-linear k−ε models can successfully reproduce most of the measured data for stream-wise and cross-stream velocity components. Moreover, the non-linear model is able to produce better results for the turbulent stresses. The heat transfer predictions show that both EVM and NLEVM2, the more recent variant of the non-linear k−ε, with the algebraic length-scale correction term, overestimate the measured Nusselt numbers for both geometries examined. While the EVM with the differential length-scale correction term underestimates heat transfer levels, the Nusselt number predictions with the NLEVM2 and the ‘NYP’ term are in close agreements with the measured data. Comparisons with our earlier work, Iacovides and Raisee (Int J Heat Fluid Flow, 20:320–328, 1999), show that the NLEVM2 thermal predictions are of similar quality to those of a second-moment closure.     

Prediction of Turbulence-Generated Secondary Mean Flow in a Square Duct

Near-wall second-moment closures have revealed a tendency to severely underpredict the strength of the turbulence-generated secondary flow in noncircular ducts. The aim of this study has been to elucidate the reasons for this failure that seems to be present at both high and low Reynolds numbers. Fully developed three-componential turbulent flow inside a straight square duct has been computed with the quasi-linear SSG second-moment closure and near-wall effects were modeled by elliptic relaxation. The results compared favourably with the reference DNS data, except that the strength of the turbulence-induced secondary flows is significantly underpredicted. A close examination of the generation mechanisms of the mean streamwise vorticity revealed that this discrepancy can be attributed to the secondary shear-stress component, the importance of which is generally overlooked. The normal-stress anisotropy was, however, adequately returned by the model and so was the variation of the friction velocity with Reynolds number over a wide Reynolds-number range. The present study supports the view that the wall-function approach should be abandoned in order to retain a physical appealing representation of the generation mechanisms of mean streamwise vorticity along internal corners. This revised version was published online in July 2006 with corrections to the Cover Date.     

Modelling of pressure–strain correlation in compressible turbulent flow

Previous studies carried out in the early 1990s conjectured that the main compressible effects could be associated with the dilatational effects of velocity fluctuation. Later, it was shown that the main compressibility effect came from the reduced pressure-strain term due to reduced pressure fluctuations. Although better understanding of the compressible turbulence is generally achieved with the increased DNS and experimental research effort, there are still some discrepancies among these recent findings. Analysis of the DNS and experimental data suggests that some of the discrepancies are apparent if the compressible effect is related to the turbulent Mach number, Mt. From the comparison of two classes of compressible flow, homogenous shear flow and inhomogeneous shear flow （mixing layer）, we found that the effect of compressibility on both classes of shear flow can be characterized in three categories corresponding to three regions of turbulent Mach numbers： the low-Mr, the moderate-Mr and high-Mr regions. In these three regions the effect of compressibility on the growth rate of the turbulent mixing layer thickness is rather different. A simple approach to the reduced pressure-strain effect may not necessarily reduce the mixing-layer growth rate, and may even cause an increase in the growth rate. The present work develops a new second-moment model for the compressible turbulence through the introduction of some blending functions of Mt to account for the compressibility effects on the flow. The model has been successfully applied to the compressible mixing layers.     

Turbulent penetrative convection with an internal heat source

An infinite fluid with a vertical cubic temperature profile in the absence of fluid motion is considered as a model for penetrative convection in which a central unstably stratified fluid layer is bounded above and below by stably stratified layers. Turbulence statistics from direct and large eddy numerical simulations for the mean temperature gradient, the velocity and temperature variances and the heat flux are presented for Rayleigh numbers R up to four orders of magnitude above critical. By means of a simplified second-moment closure, analytical scaling laws for the statistics are determined. For high Rayleigh numbers, the mean temperature gradient approaches zero in a central well-mixed layer, a reduced positive (stable) value in upper and lower partially mixed layers, and an unmixed value far above and below. The temperature variance is a factor of R^1/3 larger in the partially mixed layers compared to the well-mixed layer; the velocity variance and heat flux scales the same in both layers. Approximation of the three layers by a two layer model yields an estimate for the height of the mixed layer: the height decreases slowly with increasing Rayleigh number and at the highest Rayleigh number simulated is approximately 30% longer than the unstable layer in the absence of fluid motion.     

Toward a Turbulence Constitutive Relation for Geophysical Flows

Rapidly rotating turbulent flows are frequently in approximate geostrophic balance. Single-point turbulence closures, in general, are not consistent with a geostrophic balance. This article addresses and resolves the possibility of a constitutive relation for single-point second-order closures for classes of rotating and stratified flows relevant to geophysics. Physical situations in which a geostrophic balance is attained are described. Closely related issues of frame-indifference, horizontal divergence, and the Taylor–Proudman theorem are discussed. It is shown that, in the absence of vortex stretching along the axis of rotation, turbulence is frame-indifferent. Unfortunately, no turbulence closures are consistent with this frame-indifference that is frequently an important feature of rotating or quasi-geostrophic flows. A derivation and discussion of the geostrophic constraint which ensures that the modeled second-moment equations are frame-invariant, in the appropriate limit, is given. It is shown that rotating, stratified, and shallow water flows are situations in which such a constitutive relation procedure is useful. A nonlinear nonconstant coefficient representation for the rapid-pressure strain covariance appearing in the Reynolds stress and heat flux equations, consistent with the geostrophic balance, is described. The rapid-pressure strain closure features coefficients that are not constants determined by numerical optimization but are functions of the state of turbulence as parametrized by the Reynolds stresses and the turbulent heat fluxes as is required by tensor representation theory. These issues are relevant to baroclinic and barotropic atmospheric and oceanic flows. The planetary boundary layers in which there is a transition, with height or depth, from a thermally or shear driven turbulence to a geostrophic turbulence is a classic geophysical example to which the considerations in this article are relevant. Received 14 October 1996 and accepted 9 June 1997     

A model of stress dissipation in second-moment closures

The paper presents a modified expression for the dissipation rate tensor _ij in the second-moment closure models, which employs the dissipation flatness parameterE and the turbulenceRe number. The expression reproduced the distribution among the three diagonal components of _ij in agreement with the direct numerical simulation of a plane channel flow ofMansour, Kim and Moin, 1988. Implemented in a low-Re-number differentialRe-stress model the relationship yielded predictions of dissipative components better than other models, albeit spoiled by still unsatisfactory modelling of the equation for the energy dissipation rate . on leave from Mainski Fakultet, University of Sarajevo, Bosnia Hercegovina.     

结构可靠指标求解的一种新的迭代方法

由国际安全度联合委员会推荐的一次二阶矩可靠度方法是国际工程界公认的一种较好的可靠度分析方法（简称ＪＣ法）。这种方法具有迭代格式简单，收敛快的特点。但许多实际计算表明，这种方法只适用于功能函数非线性程度不高的情况，否则不收敛。本文提出一种新的迭代方法，这种方法迭代格式与ＪＣ法同样简单，一般情况下均能保证收敛．     

Joint Scalar versus Joint Velocity-Scalar PDF Simulations of Bluff-Body Stabilized Flames with REDIM

Two transported PDF strategies, joint velocity-scalar PDF (JVSPDF) and joint scalar PDF (JSPDF), are investigated for bluff-body stabilized jet-type turbulent diffusion flames with a variable degree of turbulence–chemistry interaction. Chemistry is modeled by means of the novel reaction-diffusion manifold (REDIM) technique. A detailed chemistry mechanism is reduced, including diffusion effects, with N ₂ and CO ₂ mass fractions as reduced coordinates. The second-moment closure RANS turbulence model and the modified Curl’s micro-mixing model are not varied. Radiative heat loss effects are ignored. The results for mean velocity and velocity fluctuations in physical space are very similar for both PDF methods. They agree well with experimental data up to the neck zone. Each of the two PDF approaches implies a different closure for the velocity-scalar correlation. This leads to differences in the radial profiles in physical space of mean scalars and mixture fraction variance, due to different scalar flux modeling. Differences are visible in mean mixture fraction and mean temperature, as well as in mixture fraction variance. In principle, the JVSPDF simulations can be closer to physical reality, as a differential model is implied for the scalar fluxes, whereas the gradient diffusion hypothesis is implied in JSPDF simulations. Yet, in JSPDF simulations, turbulent diffusion can be tuned by means of the turbulent Schmidt number. In the neck zone, where the turbulent flow field results deteriorate, the joint scalar PDF results are in somewhat better agreement with experimental data, for the test cases considered. In composition space, where results are reported as scatter plots, differences between the two PDF strategies are small in the calculations at hand, with a little more local extinction in the joint scalar PDF results.     

A three-dimensional simulation of a steady approach flow past a circular cylinder at low Reynolds number

The three-dimensional (3D) unsteady viscous wake of a circular cylinder exposed to a steady approach flow is calculated using a fractional-step finite-difference/spectral-element method. The calculated flow fields at Reynolds numbers of 100 (2D) and 200 (3D) are examined in detail. The flow field at Re = 100 is 2D as expected, while the flow field at Re = 200 has distinct 3D features, with spanwise wavelengths of about 3.75 cylinder diameters. The calculated results produce drag and lift coefficients and Strouhal numbers that agree extremely well with the experimental values. These 3D values at Re = 200 are in better agreement with experimental values than the results of a 2D calculation at Re = 200, which is expected. © 1998 John Wiley & Sons, Ltd.     

Computational Modelling of Turbulent Mixing in Confined Swirling Environment Under Constant and Variable Density Conditions

A high-intensity swirling flow in a model combustor subjected to large density variations has been examined computationally. The focus is on the Favre-averaged Navier–Stokes computations of the momentum and scalar transport employing turbulence models based on the differential second-moment closure (SMC) strategy. An updated version of the basic, high-Reynolds number SMC model accounting for a quadratic expansion of both the pressure–strain and dissipation tensors and a near-wall SMC model were used for predicting the mean velocity and turbulence fields. The accompanied mixing between the annular swirling air flow and the central non-swirling helium jet was studied by applying three scalar flux models differing mainly in the model formulation for the pressure-scalar gradient correlation. The computed axial and circumferential velocities agree fairly well with the reference experiment [So et al., NASA Contractor Report 3832, 1984; Ahmed and So, Exp. Fluids 4 (1986) 107], reproducing important features of such a weakly supercritical flow configuration (tendency of the flow core to separate). Although the length at which the mixing was completed was reproduced in reasonable agreement with the experimental results, the mixing activity in terms of the spreading rate of the shear/mixing layer, that is its thickness, was somewhat more intensive. Prior to these investigations, the model applied was validated by computing the transport of the passive scalar in the non-swirling (Johnson and Bennet, Report NASA CR-165574, UTRC Report R81-915540-9, 1981) and swirling (Roback and Johnson, NASA Contractor Report 168252, 1983) flow in a model combustor.     

Fourier neural operator approach to large eddy simulation of three-dimensional turbulence

Fourier neural operator (FNO) model is developed for large eddy simulation (LES) of three-dimensional (3D) turbulence. Velocity fields of isotropic turbulence generated by direct numerical simulation (DNS) are used for training the FNO model to predict the filtered velocity field at a given time. The input of the FNO model is the filtered velocity fields at the previous several time-nodes with large time lag. In the a posteriori study of LES, the FNO model performs better than the dynamic Smagorinsky model (DSM) and the dynamic mixed model (DMM) in the prediction of the velocity spectrum, probability density functions (PDFs) of vorticity and velocity increments, and the instantaneous flow structures. Moreover, the proposed model can significantly reduce the computational cost, and can be well generalized to LES of turbulence at higher Taylor-Reynolds numbers.