Search

An in-pond raceway system (IPRS) is an effective intensive aquaculture practice for regions with high water consumption and limited land resources. Water flow and dissolved oxygen (DO) are important for sustainable aquaculture. Several innovations have been made in IPRS design and operation to increase water exchange and DO concentration; one of these is the aeration plug-flow device (APFD). The APFD is commonly used in China as the only power source for water recirculation in aquaculture ponds. Understanding of the hydrodynamics of the system is necessary to improve the design of the IPRS with APFD. To this end, we performed experimental studies on a model system. We measured three-dimensional velocity at various locations using an Acoustic Doppler Velocimeter. Velocity distribution and turbulence characteristics were assessed, and plug-flow characteristics were analysed. Two patterns of velocity and turbulence in horizontal sections were observed: near the APFD, the water flow was intensively pushed downstream and simultaneously recirculated; farther away, the reflux area gradually decreased and the velocity and turbulence distribution trended towards uniform. Secondary flows occurred in different directions, which improved the diffusion of materials and DO retention. The system is effectively self-circulating, and the plug-flow capability may be scaled up for commercial application.

Flow recirculation in groyne fields promotes accumulation of fine sediments and development of riparian and aquatic vegetation. Although flow hydrodynamics in groyne fields is subject of intensive ongoing research, knowledge of vegetation effects on recirculating flow remains lacking. This paper reports the results of field experiments in a natural river comprising large-scale groyne models populated with rigid and flexible artificial vegetation. In the experimental runs population density, submergence ratios, and vegetation types were varied. The results show that vegetation rearranges twin-circulation flow into slow backwater flow. The magnitude of backwater flow is controlled by the bio-mechanical properties of vegetation. The hydrodynamics of the flow interface between the main river and groyne fields was only slightly altered by vegetation. Analytical models of flow at the interface and inside vegetated areas agree well with these experimental results.

A conventional box-type commercial kitchen hood and its improved version (termed the "IQV commercial kitchen hood") were studied using the laser-assisted smoke flow visualization technique and tracer-gas (sulfur hexafluoride) detection methods. The laser-assisted smoke flow visualization technique qualitatively revealed the flow field of the hood and the areas apt for leakages of hood containment. The tracer-gas concentration detection method measured the quantitative leakage levels of the hood containment. The oil mists that were generated in the conventional box-type commercial kitchen hood leaked significantly into the environment from the areas near the front edges of ceiling and side walls. Around these areas, the boundary-layer separation occurred, inducing highly unsteady and turbulent recirculating flow, and leading to spillages of hood containment due to inappropriate aerodynamic design at the front edges of the ceiling and side walls. The tracer-gas concentration measurements on the conventional box-type commercial kitchen hood showed that the sulfur hexafluoride concentrations detected at the hood face attained very large values on an order of magnitude about 10(3)-10(4) ppb. By combining the backward-offset narrow suction slot, deflection plates, and quarter-circular arcs at the hood entrance, the IQV commercial kitchen hood presented a flow field containing four backward-inclined cyclone flow structures. The oil mists generated by cooking were coherently confined in these upward-rising cyclone flow structures and finally exhausted through the narrow suction slot. The tracer-gas concentration measurements on the IQV commercial kitchen hood showed that the order of magnitude of the sulfur hexafluoride concentrations detected at the hood face is negligibly small--only about 10(0) ppb across the whole hood face.

Enhancement of bottom wall forced convection heat transfer rates in a ribbed cooling channel combined with differently shaped deflectors is investigated numerically. Four deflectors, including the sloping board (Case A), convex curved (Case B), concave curved (Case C), and cylindrical (Case D) deflectors, are mounted on the upper-downstream side of the ribs. The heat transfer and flow features are analyzed and compared based on the validation of the turbulence model and careful study of the grid independence. The results show that the flow structures are deeply affected by the deflectors, accompanied by changes in the heat transfer and friction factor. Compared with the ribbed channel with sloping board deflectors, Case B can guide the mainstream fluid to compress the recirculating flow region located on the lee-side regions downstream from the rib and can enhance the heat transfer of the bottom surface, despite the increasing friction factor. It is noted that the decrement ratio of the friction factor in Case D is 7.0-12.4%. Based on this remarkable expression, Cases B and D contribute to better overall performance. This is due to their prominent performance in the heat transfer enhancement and diminished flow resistance. (Less)

Modeling the pressure-diffusion process is discussed to improve the prediction of turbulent recirculating flows by a second moment closure. Since the recent DNS research of a turbulent recirculating flow suggeted that the pressure-diffusion process of the turbulence energy was significant in the recirculating region, the present study focuses on the rapid part of the process consisting of the mean shear. This rapid pressure-diffusion model is developed for the Reynolds stress equation using the two-component-limit turbulence condition and added to a low Reynolds number two-component-limit full second moment closure for evaluation. Its effects are discussed through applications of turbulent recirculating flows such as a trailing-edge and a back-step flows. Encouraging results are obtained though some margins to be improved still remain.

The main objective of the following work is to simulate flow in the secondary side of a steam generator (SG) tube bundle during tube rupture sequence using approaches for turbulence simulation. The interest in this topic stems from experimental findings at PSI in the framework of the ARTIST program (Aerosol Trapping in a Steam Generator) which show that particles that are injected into a dry SG tube bundle through a tube breach are principally retained in the region close to breach (the so-called “break-stage”). This region is characterized by highly non-uniform flow, with very high velocities near the breach, and low-velocity recirculating flow away from the breach. Owing to complexity of the flow, 3D simulations with highly resolved computational mesh near the breach were done. First, the flow inside an isolated tube with a guillotine tube breach has been studied. The next part is devoted to simulation of a gas jet entering the tube bundle SG via the guillotine tube breach. Comparison of the calculated flow with the experimental data of axial velocity distribution at different vertical levels have been performed, and we have found good agreement of the obtained results with the experiments.Copyright © 2014 by ASME

Calculations of mean velocities and Reynolds stresses are reported for the recirculating flow established in the wake of two-dimensional polynomial-shaped obstacles that are symmetrical about a vertical axis and mounted in the water channel downstream of a fully developed channel flow for Re = 6 x 10 4 . The study involves calculations of mean and fluctuating flow properties in the streamwise and spanwise directions and include comparisons with experimental data [Almeida GP, Durao DFG, Heitor MV] for flow around a single obstacle with data resulting from the interaction of consecutive obstacles, using two versions of the low-Reynolds number differential second-moment (DSM) closure model. The results include analysis of the turbulent stresses in local flow co-ordinates and reveal flow structure qualitatively similar to that found in other turbulent flows with a reattachment zone

Introduction of a thermal field as a passive scalar seeding has been made in a channel with significant wall-injection convection and presence of a predominant obstacle. In such a complex flow, the study of the temperature field provides a good description of the whole dynamic organization of the flow: development of a highly sheared flow, reversal region back to the obstacle… The scalar and the dynamic fields are close together, i.e. the study of the mean temperature and of its fluctuations allows analysis of the scales of flow turbulence and description of the creation of vortices in this sheared flow.

Velocity data have been obtained using a bi-directional triple-split hot-film probe placed in the mixing layer, the recirculating region and the relaxing boundary layer of turbulent flow over a backwards-facing step. The probe is fully validated and it is shown that it provides measurements both of velocity and of angle to within an acceptable accuracy. Although these probes have a fairly poor frequency response ( Hz) through the effects both of spatial averaging and of thermal inertia, they are invaluable in providing bi-directional velocity data. The probe is not suitable for measurements in low-intensity flows. The calibration has been found to be surprisingly invariant with respect to use.

Two cyclone dust separators with geometric swirl numbers of 3.324 and 3.043 were used to analyze the motion of the complex three-dimensional time dependent motion set up in the free exhaust. A quantitative analysis of the flow was carried out, obtaining time dependent velocity measurements with the use of laser Doppler anemometry (L.D.A.) techniques. The investigations highlighted a eddy or vortex shedding mechanism in two distinct areas of the flow. This was in part caused by a reverse flow zone and a precessing vortex core within the exhaust region of the separator. Changes in the Reynolds number by a factor of 2 were observed to have no effect on the main characteristics of the flow. Some changes were seen in the flow structure with change in swirl number, particularly the size of the reverse flow zone and the position of the large engulfment vortices.

The flow patterns in semi-enclosed water bodies with free surfaces generally consist of free or forced recirculations which can have significant effect on the water quality and fluid exchange. To accurately predict the flow structures in these water bodies a three-dimensional layer-integrated numerical model has been developed. The turbulence (consisting of the free shear component and the bottom friction component) is parametrized by the k–ω equations. The governing equations are split into three parts in the finite difference solution: advection, dispersion and propagation. The advection part is solved by the four-node minimax-characteristics scheme. The dispersion part is solved by the central difference method and the propagation part is solved implicitly by using the Gauss–Seidel iteration method to overcome the restrictive Courant–Friedrich–Lewy (CFL) stability constraint. The model has been verified against the laboratory data on forced recirculating flow in a physical model harbour and on free recirculating flow in an expanded channel.

Fuel droplet characteristics in a flame holding region of an oil furnace (0.1MW) were investigated both in experiments and simulations. The size and shape of the recirculation flow under different conditions were examined and it was found that the simulation underestimated the size of the recirculation zone due to a κ-e turbulent model. The size classification technique of fuel droplets was implemented in order to understand the followable characteristics of droplet in a high-turbulence region. A droplet under 30μm can follow the recirculating flow, but that over 50μm penetrates the recirculation zone due to large momentum. A conceptual flow illustration was demonstrated.

A numerical study of a turbulent penetrative and recirculating flow induced by the energy input from a fire at the bottom boundary in a partially open rectangular enclosure is carried out. The compartment is connected through an opening to a long corridor, which opens into a stably stratified environment. The condition that is of interest is a stable, two-layered, temperature stratification, which is assumed to be caused by fire in an adjacent enclosure. In this study, attention is focused on the interaction between the cavity and its surrounding ambient medium through the opening. The influence of the stratification is examined in the turbulent flow regime by considering a range of stratification levels for given opening height and initial interface location. It is found that, depending on the stratification, the thermal plume above the fire may never reach the ceiling. Small penetration distances occur at large stratification levels. The flow field reveals a multicellular pattern: a strong main convective cell at the bottom and a weak counterrotating cell at the top. The stable thermal stratification can cause a destruction of the turbulence. This results in the relaminarization of the flow in the upper region of the cavity and may significantly affect the transport processes in the enclosure. This could distort the simplistic concept of two homogeneous gas layers, which forms the basis of zone modeling for compartment fires.

An experimental study has been presented which deals with the deflection and reattachment of a radial turbulent jet (main jet) discharged from a cylindrical nozzle onto an adjacent disk plate in the presence of a lateral controled flow from an annular nozzle, Static pressure distributions along the side wall and across the jet flow were measured using a disk-type pressure hole. The relation between the deflection of the main jet, the reattachment point on the side wall, the mean negative pressure in a recirculating flow region and the curvature of the jet center axis were obtained for various values of the control flow rate and the step height, which describe the length of and the pressure within the recirculating flow region.

The art of “modeling turbulence” is a needed tool in the construction of computer codes for turbulent flows. The state to which this art has been developed is inadequate, and quotations from authoritative sources support this point of view. The energy contained in the turbulent fluctuations, i.e., the turbulent energy, is often used as a parameter in the modeling process. The present article attempts to examine this quantity as it is being created, transported, and dissipated. For this purpose experimental evidence from the author's own experiments (free jets), as well as theoretical conclusions from the elementary deductions of the basic equations, the concept of turbulent potential flow, and a general solution to the Navier-Stokes-Reynolds equations, is drawn to attention. Recirculating flow is given special attention. The paper concludes with recommendations for principles that must be satisfied if improved modeling is to be achieved. These principles are necessary; whether they are also sufficient is open to question.

Bu çalısmada, eksenel simetrik anigenislemeli borularda sürekli, sıkıstırılamayan, kompleks çevrintili akısın kapsamlı bir çalısmasının sayısal hesaplama sonuçları sunulmaktadır. Hibrit yöntemiyle geleneksel sonlu hacim metodunu kullanarak, SIMPLE (Semi-Implicit Method for Pressure Linked Equations) algoritmasına dayanan bir bilgisayar programı gelistirilmistir. Standart yüksek Reynolds sayılı k-epsilon türbülans modeliyle beraber, kütle ve momentum korunum denklemlerinin sayısal çözümleri, iteratif bir sayısal çözüm tekniğini kullanarak sağlanmıstır. Katı cidarlar yakınında cidar fonksiyonları kullanılmıstır. Eksenel simetrik anigenislemeli borularda çesitli eksenel kesitlerde, eksenel hız, türbülans kinetik enerji, türbülans kinetik enerji kaybolma miktarı, efektif viskozite radyal profilleri, boru simetri ekseni boyunca eksenel hızın değisimi, geri akısın geometrik yeri ve eksenel simetrik anigenislemeli boruda cidar kayma gerilmesinin dağılımı için sayısal hesaplamalar sunulmus ve deneysel ölçümlerle karsılastırılmıstır. Sayısal hesaplama sonuçları deneysel bulgularla iyi uyum göstermektedir. In this study, the results of an extensive study of numerical computation of steady, complex turbulent recirculating flow in axisymmetric sudden expansions in pipes are presented. Employing the conventional finite-volume method with a hybrid scheme, a computer program based on the SIMPLE (Semi-Implicit Method for Pressure Linked Equations) algorithm has been developed. Numerical solutions of the conservation equations of mass and momentum, together with the standard high- Reynolds-number k-epsilon turbulence model, are obtained using an iterative numerical solution technique. Near the solid boundaries, wall functions are employed. Numerical computations for radial profiles of axial velocity, turbulence kinetic energy, turbulence kinetic energy dissipation rate, effective viscosity, centerline velocity decay, locus of flow reversal and distribution of wall shear stress along axisymmetric sudden expansion flow geometry are presented and compared with experimental measurements. The results of numerical predictions show good agreement with experimental data.

This investigation deals with the modification of streamline curvature effects in the k-e turbulence model for the case of recirculating flows. Based upon an idea that the modification of curvature effects in C2 should not be made in regions where the streamline curvature is small, a hybrid k-e model extended from the modification originally proposed by Srinivasan and Mongia is developed. A satisfactory agreement of model predictions with experimental data reveals that the hybrid k-e model can perform better simulation of recirculating turbulent flows.

The influence of a passive (no external energy input) and an active (via energy input) control method on the flow characteristics of a turbulent separated and reattaching flow is investigated. A surface-mounted fence and an oscillating blowing/suction jet are utilized as passive and active control method, respectively. The impact of these methods is analyzed here for a flow with (a freely evolving) separation along a smoothly contoured convex surface. The results of large-eddy simulations (LES) of the turbulent flow over a rounded step at Re h = 9100 reveal the importance of properly chosen control parameters which can lead to significant reductions of the size of the recirculating flow region.

An investigation of recirculating characteristics of circular turbulent confined jets with large density variations is presented. This numerical study aims at testing analytical predictions associated with the Craya-Curtet number. It investigates also its effect on the dynamic field and recirculation. It appears that for different density (He-air, CO2-air) and geometry ratios, the non-isothermal Craya-Curtet number [1] is not sufficient to describe the flow pattern or predict its recirculation. Depending on the momentum, aspect and density ratios of the flow, the centre of the recirculating flow (the eye) tends to reach the initial (non-self-preserving) region of the jet and influences the development of the jet. As a consequence, predictions are not in agreement with theory. The reason is that initial conditions do not satisfy the hypothesis used in the prediction of recirculation theory. Calculations are performed in three configurations: CO2, air and He. These configurations are fully developed pipe jets evolving in an air secondary turbulent flow. Validations are performed using experimental data [2] obtained in similar configurations for the three gases.

The mean velocity field of unventilated two parallel plane jets for large nozzle spacing ratios has been studied using a two-component laser Doppler anemometer. Results show that by comparison with small s/w (less than 5), the interactions between the inner shear layer and the recirculation zone for large s/w are weaker. The initial peaks of turbulence intensities and Reynolds shear stress corresponding to the nozzle inner and outer edges tend to persist downstream from the location of the vortex centre without being overwhelmed by the influences caused by the recirculating flow in the recirculation zone. The lateral turbulence intensities and Reynolds shear stress in the outer shear layer spread and decay more rapidly than in the inner shear layer. The attraction of the two individual jets towards each other is clearly illustrated by the mean velocity vector field and supported by static pressure measurements. Comparisons of the results of the unventilated jets with those of ventilated jets in the literature indicate there is no recirculation zone in the ventilated jets and the unventilated jets combine to develop into a single free jet much earlier than ventilated jets.

The two-dimensional turbulent recirculating flow fields behind a V-shaped bluff body have been investigated numerically. Similar bluff bodies are used in combustion chambers for flame stabilization. The governing transport equations in conservative form are solved by a pressure based predictor-corrector method. The standard k-ϵ turbulence closure model and a boundary fitted multi-block curvilinear grid system are used in the computation. The code is validated against turbulent flow over a backward facing step problem. The predicted flow field behind the bluff body is also compared with experiment. It is found that while the qualitative features of the flow are well predicted, there is quantitative disagreement between the measurement and prediction. This disagreement can be partially attributed to the k-ϵ turbulence model which is known to be inadequate for recirculating flows. Parametric investigation of the flow field by varying the shape and size of the bluff body is also performed and the results are reported.

Source: https://erdc-library.erdc.dren.mil/jspui/ A simple adjustment is proposed that improves the predictions of the standard k-? turbulence model in the presence of two-dimensional recirculating flow. An eddy Reynolds number calculated from velocity, vorticity, and eddy viscosity is introduced as a scaling parameter for reducing the decay term in the governing equation for the turbulence-energy dissipation rate. Using the adjusted k-? model to compute the flow past a backstep (channel expansion), the STEMR finite-difference code gives better predictions for streamlines, velocities, and turbulence energies than it does with the standard k-? model. Computed reattachment lengths compare well with experimental data for channel expansion ratios between 1 and 2.

Amelioration de la precision des previsions en incorporant un modele dependant de la courbure dans le modele k-e

Experimental measurements were taken in a closed, recirculating, turbulent separated flow. The experiments were made to elucidate the mass and momentum exchange mechanisms between the throughflow and the recirculating flow in a combustor that had previously been developed for burning heavy fuels without a flame tube. Mean velocity measurements are presented for the cold flow field and these are used to derive a detailed streamline pattern. The three-dimensional shear layer surface that divides the recirculating flow from the throughflow is analyzed in detail and the variation of momentum exchange coefficient along its development length is derived. Mass exchange data show that the variation of mass transport along the mixing layer is similar to the variation of momentum transport. Overall features of the exchange processes are shown to be similar to those in simple two-dimensional, nonducted mixing flows if allowance is made for the three dimensionality of the present layers. However, it is also shown that there are significant quantitative differences between the two types of flow. The overall mass exchange rate between the two flow regions was found to be one-third of the mass rate supplied to the combustor and this is a typical figure for conventional designs.

This paper proposes an integro-differential equation to describe the decay of particle gas-borne concentration in recirculating turbulent flows which are fully mixed. Its application to the decay of gas-borne concentration for particles injected into the primary circuit of an Advanced Gas Cooled Reactor is considered. The novel aspect of this work is the way the resuspension following initial deposition is modeled.

Recirculating flows in channel induction furnaces require investigation in order to understand and predict their mixing efficiency and their disadvantageous aspects, such as blocking of the channel by non-metallic materials and refractory erosion. These aspects reduce their lifetime. To understand better the physical mechanisms driving the flow, we have divided our study in two parts. In the first which is described in this paper, we analyse the two-dimensional flow in a laboratory model simulating one leg of the channel. The model consists of a 150 × 150 mm square tank containing 90 kg of mercury and powered from a 400 Hz generator. The distribution of the flux B was measured using search coils. Thence estimates of the distribution of the ( J × B ) and its curl ▽ × ( J × B ) were derived. The results were compared with computed solutions to Maxwell's equations obtained from a finite element code. This is a decoupled problem. In that first Maxwell's equations are solved, and then, using the results for J × B , the Navier-Stokes equations are solved. The mean flow and the turbulent intensity of the recirculating flow were measured and compared with solutions obtained from a (kk-ϵ) code, and approximate analyses based on a balance of the ▽ × (J × B) and inertia terms. Numerical experiments show how the flow pattern in the channel cross-section changes with alterations in either (i) the channel g eometry or (ii) the coil length and (iii) the position of the channel to the coil.

The recirculating flow characteristics in a confined space with axisymmetric sudden expansion followed by a conical contraction section were measured by means of an electrochemical method. Integral scales of space correlation in both longitudinal and circumferential directions were obtained. Also, the recirculating velocity of a large eddy was determined from the measurement of space-time correlation. Experimental results confirmed that there exists a large recirculating eddy near the wall of test chambers.

This paper presents a review of nitrogen-oxides emission control technology by combustion modification and by altering combution condition. The influence of factors-mixing, swirl, recirculating flow, turbulence, oscillation, high temperature, high pressurs reburning and additives-on NO formation and a cooleddisk burner as low NOx emission burner are reviewed. In addition nonuniformities and rates of mixin ghave important effects on NO formation, therefore, the influence of mixture ratio of reactants for premixed flame, and diffusion flame are also described.As NO formation in combustion is markedly affected by combustion condition, we can control a large amount of NO emission by using suitable burner parameters .

Measurements of transient water flows are performed in a 1/15-scale Fast Flux Test Facility nuclear reactor outlet plenum. During a repeated inlet-flow coastdown transient, measurements are made of the time-dependent, ensemble-averaged values of the velocity, turbulence kinetic energy, and Reynolds stress fields. These data are compared to predictions from a fluid dynamic simulation which uses a two-equation turbulence model. In general the comparison is good, and most observed transient disagreements can be attributed to errors in the calculation of the initial conditions rather than transient effects.

Etude experimentale de la vitesse de fusion de barres de glace immergees dans un recipient plein d'eau agitee par un courant ascendant de bulles de gaz. Etablissement d'une correlation entre le nombre de Wussell et le nombre de Reynolds turbulent

This paper presents the results of an extensive study of subsonic separated flows using a laser Doppler velocimeter. Both a rectangular rearward facing step and cylindrical (axisymmetric) sudden expansion geometry were studied. The basic objectives were to resolve the question of whether a velocity bias error does, in fact, occur in LDV measurements in highly turbulent flows of this type and, if so, how it may be eliminated; map the velocity field (mean velocity, turbulence intensity, Reynolds stress, etc.) including the entire recirculation zone; and compare experimental results with numerical predictions based on the k-ε turbulence model. Measurements were carried out using a one-dimensional LDV operating in forward scatter with signal processing by means of a commercial counter-type processor. Results obtained show that velocity bias does occur in turbulent flows and that it can be overcome by proper data acquisition procedures. The results also indicate that the important mean velocity and turbulence quantities can be obtained with reasonable accuracy using a one-dimensional LDV system. Although the k-ε turbulence model provides a good qualitative picture of the flow field, it does not yield a completely adequate quantitative description. Results obtained here illustrate the discrepancies to be expected and provide a basis for further model development.

A mathematical model of turbulent density-driven flows is presented and is solved numerically. A form of the k–ϵ turbulence model is used to characterize the turbulent transport, and both this non-linear model and a sediment transport equation are coupled with the mean-flow fluid motion equations. A partitioned, Newton–Raphson-based solution scheme is used to effect a solution. The model is applied to the study of flow through a circular secondary sedimentation basin.

Calcul de la structure de la zone de recirculation se formant derriere une marche tournee vers l'aval pour des nombres de Reynolds de 50 a 5000 a l'aide de la methode des tourbillons aleatoires. Les resultats sont analyses en termes de lignes de courant instantanees, de profils de vitesse moyens et de statistiques de turbulence. On idenfie 4 regimes d'ecoulement distincts

Shortcomings and recommended corrections to the standard two-equation k-epsilon turbulence model suggested by previous investigators are presented. They are assessed regarding their applicability to turbulent swirling recirculating flow. Recent experimental data on swirling confined flows, obtained with a five-hole pitot probe and a six-orientation hot-wire probe, are used to obtain optimum values of the turbulence parameters C-mu, C2, and sigma-epsilon for swirling flows. General predictions of moderately and strongly swirling flows with these values are more accurate than predictions with the standard or previous simple extensions of the k-epsilon turbulence model.

Presentation d'une methode numerique par volume fini pour la resolution des equations de Navier-Stokes bidimensionnelles, incompressibles, et stationnaires, en coordonnees generales curvilignes. Application de la methode aux ecoulements turbulents sur des profils avec et sans separation au bord de sortie posterieur. Comparaison des calculs avec des donnees experimentales

Evaluation de la performance d'un modele stochastique pour le calcul de la dispersion des particules dans un champ d'ecoulement turbulent confine elliptique (avec recirculation)

Analysis of experimental data, obtained in the present work on the vortical structure of a recirculating flow formed by a backward-facing step, has allowed the development of an approximate calculation method for predicting velocity and temperature profiles, reattachment length and local friction and heat transfer coefficients. An analytical expression for heat transfer in the reattachment region is derived which correlates experimental heat transfer data for different types of flow with a sudden expansion.

Les resultats experimentaux montrent que deux cellules d'ecoulement secondaire se developpent autour des coins de maniere semblable a celle observee dans le cas incompressible. Les resultats numeriques sont bases sur le modele de Baldwin-Lomay