A radiation code based on method of lines (MOL) solution of discrete ordinates method (DOM) for the prediction of radiative heat transfer in nongray absorbing-emitting media was developed by incorporation of two different gas spectral radiative property models, namely wide band correlated-k (WBCK) and spectral line-based weighted sum of gray gases (SLW) models. Predictive accuracy and computational efficiency of the developed code were assessed by applying it to the predictions of source term distributions and net wall radiative heat fluxes in several one- and two-dimensional test problems including isothermal/non-isothermal and homogeneous/non-homogeneous media of water vapor, carbon dioxide or mixture of both, and benchmarking its steady-state predictions against line-by-line (LBL) solutions and measurements available in the literature. In order to demonstrate the improvements brought about by these two spectral models over and above the ones obtained by gray gas approximation, predictions obtained by these spectral models were also compared with those of gray gas model. Comparisons reveal that MOL solution of DOM with SLW model produces the most accurate results for radiative heat fluxes and source terms at the expense of computation time when compared with MOL solution of DOM with WBCK and gray gas models. In an attempt to gain an insight into the conditions under which the source term predictions obtained with gray gas model produce acceptable accuracy for engineering applications when compared with those of gas spectral radiative property models, a parametric study was also performed. Comparisons reveal reasonable agreement for problems containing low concentration of absorbing-emitting media at low temperatures. Overall evaluation of the performance of the radiation code developed in this study points out that it provides accurate solutions with SLW model and can be used with confidence in conjunction with computational fluid dynamics (CFD) codes based on the same approach.

Climate models are the primary tools used for understanding past climate variations and for future projections. The atmospheric radiation is the key component of these models. Accurate modeling of atmosphere necessitates reliable evaluation of the medium radiative properties and accurate solution of the radiative transfer equation in conjunction with the time-dependent multi-dimensional governing equations of atmospheric models. Due to difficulty in solving the equations of atmospheric and radiation models simultaneously, radiation equations have been solved when input data such as concentration, temperature etc. were made available upon solution of equations of atmospheric models. Generally, time step of conservation equations are 10-30 minutes but radiative transfer equation is called only once every 1-3 hours. However, there is inaccuracy due to the fixed radiation fluxes over the intervening time steps. To overcome this problem, the equations of atmospheric and radiation models have to be solved simultaneously and the solution methods have to be compatible. For this purpose, a radiative-convective model with radiation model based on method of lines (MOL) solution of discrete ordinate method (DOM) with wide band correlated-k (WBCK) was developed. To achieve this objective, a previously developed MOL solution of DOM with WBCK model was adapted to 1-D longwave clear sky atmosphere and its predictive accuracy and computational efficiency was examined on the test problem by using benchmark solution obtained from Line-by-line Radiative Transfer Model (LBLRTM). The radiation code was then coupled with radiative-convective model and the predictive accuracy of this model was examined for several coupling intervals. Comparisons reveal that as coupling interval increases, although the computation time of the model decreases, the predicted temperature profiles diverge from the one obtained when equations of radiative-convective model and the radiation model are solved simultaneously and percentage relative error in temperature increases an order of magnitude when coupling time between radiative-convective model and the radiation model increases from 2 to 10 hours. Therefore, it can be concluded that the equations of the radiation model have to be solved simultaneously with the equations of the climate model. Overall evaluation of the performance of the radiation model used in this study points out that it provides accurate and computationally efficient solutions and can be used with confidence in conjunction with the climate models for simultaneous solution of governing equations with radiation transfer equation.

This work has been aimed at developing a mechanistic, transient, 3-D numerical model to predict the behavior of an evaporating thin liquid film on a non-uniformly heated cylindrical rod with simultaneous parallel and cross flow of vapor. Interest in this problem has been motivated by the fact that the liquid film on a full-length boiling water reactor fuel rod may experience significant axial and azimuthal heat flux gradients and cross flow due to variations in the thermal-hydraulic conditions in surrounding subchannels caused by proximity to inserted control blade tip and/or the top of part-length fuel rods. Such heat flux gradients coupled with localized cross flow may cause the liquid film on the fuel rod surface to rupture, thereby forming a dry hot spot. These localized dryout phenomena can not be accurately predicted by traditional subchannel analysis methods in conjunction with empirical dryout correlations. To this end, a numerical model based on the Level Contour Reconstruction Method was developed. The Standard k- turbulence model is included. A cylindrical coordinate system has been used to enhance the resolution of the Level Contour Reconstruction Model. Satisfactory agreement has been achieved between the model predictions and experimental data. A model of this type is necessary to supplement current state-of-the-art BWR core thermal-hydraulic design methods based on subchannel analysis techniques coupled with empirical dry out correlations. In essence, such a model would provide the core designer with a "magnifying glass" by which the behavior of the liquid film at specific locations within the core (specific axial node on specific location within a specific bundle in the subchannel analysis model) can be closely examined. A tool of this type would allow the designer to examine the effectiveness of possible design changes and/or modified control strategies to prevent conditions leading to localized film instability and possible fuel failure.

Vegetation is commonly modelled as emergent arrays of rigid, circular cylinders. However, the drag coefficient (CD) of real stems or trunks is closer to that of cylinders with a square cross-section. In this thesis, vegetation has been idealised as square cylinders in laboratory experiments with a turbulence intensity of the order of 10% which is similar to that of typical river flows. These cylinders may also represent other obstacles such as architectural structures. This research has determined CD of an isolated cylinder and cylinder pairs as a function of position as well as the average drag coefficient (CDv) of larger arrays. A strain gauge was used to measure CD whilst CDv was computed with a momentum balance which was validated by strain gauge measurements for a regularly spaced array. The velocity and turbulence intensity surrounding a pair of cylinders arranged one behind the other with respect to mean flow (in tandem) were also measured with an Acoustic Doppler Velocimeter. The isolated cylinder CD was found to be 2.11 in close agreement with other researchers. Under fixed flow conditions CD for a cylinder in a pair was found to be as low as -0.40 and as high as 3.46 depending on their relative positioning. For arrays, CDv was influenced more by the distribution of cylinders than the flow conditions over the range of conditions tested. Mean values of CDv for each array were found to be between 1.52 and 3.06. This new insight therefore suggests that CDv for vegetation in bulk may actually be much higher than the typical value of 1 which is often assumed to apply in practice. If little other information is available, a crude estimate of CDv = 2 would be reasonable for many practical applications. The validity of a 2D realizable k-epsilon turbulence model for predicting the flow around square cylinders was evaluated. The model was successful in predicting CD for an isolated cylinder. In this regard the model performed as well as Large Eddy Simulations by other authors with a significant increase in computational efficiency. However, the numerical model underestimates CD of downstream cylinders in tandem pairs and overestimates velocities in their wake. This suggests it may be necessary to expand the model to three-dimensions when attempting to simulate the flow around two or more bluff obstacles with sharp edges.

The flow pattern in the uphill teeming process has been found to be closely related to the quality of ingots and further to affect the yield of ingot production, which is crucial for the steel making process. The formation of non-metallic inclusion and entrapment of mold flux has been considered to be affected by the flow pattern in the gating system and molds by many previous researchers. The aim of this study is to investigate the flow pattern of steel in the gating system and molds during the initial filling stage. In addition, to study the utilization of swirl blade implemented at the bottom of the vertical runner on the improvement of initial filling condition in the mold. A three dimensional model of two molds gating system for 6.2 ton ingots from Scana Steel was adopted in the present work. A reduced geometry model including one mold and a runner, based on the method from previous researchers, was also used for comparison with the current more extensive model. Moreover, a reduced geometry model including one swirl blade and a runner was simulated to find effects of an increased-length vertical runner on the flow pattern improvement at the vertical runner outlet. Flow pattern, hump height and wall shear stress were respectively studied. A reduced geometry with homogenous inlet conditions fails to describe the fluctuating conditions present as the steel enters the mold. However, the trends are very similar when comparing the (hump height-surface height) evolution over time. The implementation of swirl blades gives a chaotic initial filling condition with a considerable amount of droplets being created when steel enters the molds during the first couple of seconds. However, a more calm filling condition with less fluctuation is achieved at the molds after a short while. Moreover, the orientation of the swirl blades affects he flow pattern of the steel. A proper placement of a swirl blade improves the initial filling conditions. The utilization of swirl blades might initially result in larger hump height. However, it gives fewer fluctuations as the casting proceeds. In the model without swirl blades, the maximum wall shear stress fluctuates with a descending trend as the filling proceeds. An implementation of swirl blades can decrease and stabilize the wall shear stress in the gating system. A special attention should be made in choosing refractory at the center stone, the horizontal runner near center stone and the vertical runner at the elbow. This is where the wall shear stress values are highest or where the exposure times are long. / QC 20120203

More and more attention has been paid to decreasing the number and size of non-metallic inclusions existing in the final products recently in steel industries. Therefore, more efforts have been made to monitor the inclusions' size distributions during the metallurgy process, especially at the secondary steelmaking period. A liquid sampling procedure is one of the commonly applied methods that monitoring the inclusion size distribution in ladles, for example, during the secondary steelmaking. Here, a crucial point is that the steel sampler should be filled and solidified without changing the inclusion characteristics that exist at steel making temperatures. In order to preserve the original size and distributions in the extracted samples, it is important to avoid their collisions and coagulations inside samplers during filling. Therefore, one of the first steps to investigate is the flow pattern inside samplers during filling in order to obtain a more in-depth knowledge of the sampling process to make sure that the influence is minimized. The main objective of this work is to fundamentally study the above mentioned sampler filling process. A production sampler employed in the industries has been scaled-up according to the similarity of Froude Number in the experimental study. A Particle Image Velocimetry (PIV) was used to capture the flow field and calculate the velocity vectors during the entire experiment. Also, a mathematical model has been developed to have an in-depth investigate of the flow pattern in side the sampler during its filling. Two different turbulence models were applied in the numerical study, the realizable k-ε model and Wilcox k-ω model. The predictions were compared to experimental results obtained by the PIV measurements. Furthermore, it was illustrated that there is a fairly good agreement between the measurements obtained by PIV and calculations predicted by the Wilcox k-ω model. Thus, it is concluded that the Wilcox k-ω model can be used in the future to predict the filling of steel samplers.

Rotating bed reactors are used throughout the process industry. They are usedboth in the chemical industry and other industrial sectors, such as pharmaceuticals and the textile industry in decolorization due to by-products or contaminants.SpinChem AB manufactures rotary bed reactors (RBRs) to perform chemical reactions between liquids and solids. The solid material consists of spherical particles0.1 mm - 1 mm in diameter that are packed between two cylindrical spaces in theRBR. The goal of this project work is to determine the power number, the axial force thatthe RBRn experiences, the torque requirement on the motor and power consumptionof the the RBR when a fully developed turbulent flow is achieved. The purpose ofthe work is to optimize the technology from the energy usage point of view, makethe product simple and easily accessible for chemical and industrial processes as acontribution to the development of sustainable society. In order to achieve the purpose and goal of the projects, Computational Fluid Dynamics (CFD) combined withexperimental models were used. Computation were made in COMSOL Multiphysicsfor two turbulence models. In it, the rotating machinery was used with moving meshtechnique for both the standard k−ε model and the SST k−ω turbulence models.The result is then compared with the empirical models. Investigation were done for two models of the rotating bed reactors (RBRs). Onemodel is called RBR S2 with relatively small size and RBR S14 which is a muchlarger version. For RBR S2 the experimental results turned out to be, an output ofpower number which is 3.4, torque requirement of 0.03 Nm, power consumption of3 W and a thrust force of 0.11 N. While the simulation results turned out to bean output of power number which is about 1.2, torque requirement of 0.013 Nm, apower consumption of 2 W and thrust force of 0.8 N. Similarly, the experimentalresult for RBR S14 was as follows. A power number of 0.53, torque requirement of0.41 Nm, power consumption of 6 W and a thrust force of 4.16 N. The simulationresults turned out to be, a power number of 0.34, torque requirement of 0,40 Nm,a power consumption of 4.14 W and thrust force of 3.61 N. With the help of the calculated power numbers, the power required to rotate theRBR can then be determined. Power number is determined when a fully developedturbulent flow is achieved. For RBRS2, a fully developed turbulent flow is achievedat Re = 2.8·104 and the angular velocity at that Reynolds number is about 830RPM. At that speed, the power is shown to be about 4 W for RBRS2. For RBRS14,a fully developed turbulent flow is achieved at Re = 1.5 · 105 and then the speed atthat Reynols number is about 83 RPM. The power need at that stage is shown tobe about 20 W. / Roterande bäddreaktorer används inom hela processindustrin. De används bådeinom den kemiska industrin och andra industriella sektor såsom, läkemedel och textilindustrin vid avfärgning på grund av biprodukter eller föroreningar. SpinChemAB tillverkar roterande bed reaktorer (RBR) för att utföra kemiska reaktioner mellan vätska och fasta material. Det fasta materialet består av sfäriska partiklar på0,1 mm - 1 mm i diameter som packas mellan två cylindrar i RBRn. Målet med detta projektarbete var att bestämma effekt nummer, effekt som krävsvid det effekt nummer, kravet på vridmoment från motorn samt den axiella kraftensom den roterande bäddreaktorn upplever när ett fullt utvecklat turbulent flöde uppnåtts. Syftet med arbetet var optimera teknologin ur energianvändningssynpunkt, göra den enkel och lättillgänglig för kemiska och industriella processer som ett bidragför hållbar samhällsutveckling. För att kunna uppnå syftet och målet med projekten användes, avancerade beräkningsmetoder i födes mekanik (CFD) i kombinationmed experimentella modeller. Beräkningar gjordes i COMSOL Multiphysics för tvåturbulenta modeller. I de användes roterande maskineriet med en medföljande mesh (moving mesh) för både standard k-ε modellen och SST k-ω modellen. Resultatet jämfördes sedan med de empiriska modellerna. Undersökningarna gjordes för två modeller av RBR. Ena modellen heter RBR S2med relativt små tillstorlek och RBR S14 som är mycket större version. För RBR S2visar den experimentella resultaten ett effekt nummer på 3,4, vridmoment på 0,03Nm, effekt förbrukning på 3 W och en axiellkraft ("thrust force") på 0,11 N. Simuleringsresultatet visar ett effekt nummer på 1,2, vridmoment på 0,013 Nm, effektförbrukning på 2 W och en axiellkraft på 0,8 N. För RBR S14 visade det experimentella resultatet ett effekt nummer på 0,53, vridmoment på 0,41 Nm, effektförbrukning på 6 W och en axiellkraft ("thrust force") på 4,16 N. Simuleringsresultatetvisade att effekt nummer var 0,34, vridmoment på 0,40 Nm, effektförbrukning på4,14 W och en axiellkraft på 3,61 N. Med hjälp av de framräknade effektnummer kan effekten som behövs rotera RBRnbestämmas. Effektnummer bestäms när ett fullt utvecklat turbulent flöde uppnåtts. För RBRS2 uppnås ett fullt utvecklat turbulent flöde vid Re = 2,8·04 och vinkelhastigheten är 830 RPM vid det Reynolds nummer. Effekten som krävs för att drivaRBRn vid det läge är ca 4 W för RBRS2. För RBRS14 uppnås ett fullt utvecklatturbulent flöde vid Re = 1,5·105 och då har vi en hastighet på 83 RPM. Vid denhastighet visas effekten vara ca 20 W.

Este trabalho tem como objetivo desenvolver um método numérico para simular escoamentos incompressíveis, isotérmicos, confinados ou com superfícies livres, de fuidos viscoelásticos governados pelos modelos integrais de Maxwell e K-BKZ (Kaye-Bernstein, Kearsley e Zapas). A técnica numérica apresentada é uma extensão do método GENSMAC (Tomé McKee - J. Comp. Phys., (110), pp 171--186, 1994 ) para a solução das equações de conservação, juntamente com as equações constitutivas integrais de Maxwell e K-BKZ. As equações governantes são resolvidas pelo método de diferenças finitas em uma malha deslocada. O tensor de Finger, B_t\'(t) é calculado com base nas idéias do método de campos de deformação (Peters et al. - J. Non-Newtonian Fluid Mech. (89), de maneira que não há a necessidade de seguir a trajetória da partícula de fuido para descrever a história de deformação da partícula. Uma abordagem diferente para a discretização do instante passado é utilizada e o tensor de Finger e o tensor das tensões são calculados utilizando um método de segunda ordem. A validação do método numérico descrito nesse trabalho foi feita utilizando o escoamento em um canal bidimensional e a solução numérica obtida para a velocidade e para as componentes de tensão com o modelo de Maxwell foram comparadas com as respectivas soluções analíticas no estado estacionário, mostrando excelente concordância. Os resultados numéricos para a simulação do escoamento em uma contração planar 4 : 1 mostraram bons resultados, tanto qualitativos quanto quantitativos, quando comparados com os resultados experimentais de Quinzani et al. ( J. Non-Newtonian Fluid Mech. (52), pp 1?36, 1994 ). Além disso, utilizando os modelos Maxwel e K-BKZ, o escoamento em uma contração planar 4 : 1 foi simulado para vários números de Weissenberg e os resultados obtidos estão de acordo com os encontrados na literatura. Resultados numéricos de escoamentos com superfícies livres modelados pelas equações integrais de Maxwell e K-BKZ são apresentados. Em particular, a simulação numérica do jato oscilante para diferentes números de Weissenberg e diferentes números de Reynolds é apresentada. / The aim of this work is to develop a numerical technique for simulating incompressible, isothermal, free surface (also con¯ned) viscoelastic flows of fuids governed by the integral models of Maxwell and K-BKZ (Kaye-Bernstein, Kearsley and Zapas). The numerical technique described herein is an extension of the GENSMAC method (Tome and McKee, J. Comput. Phys., 110, pp. 171-186, 1994) to the solution of the momentuum and mass conservation equations together with the integral constitutive Maxwell and K-BKZ equations. The governing equations are solved by the finite difference method on a staggered grid using a Marker-and-Cell approach. The fluid is represented by marker particles on the fluid surface only. This provides the visualization and location of the fluid free surface so that the free surface stress conditions can be applied. The Finger tensor Bt0(t) is computed using the ideias of the deformation fields method (Peters et al. J. Non-Newtonian Fluid Mech., 89, pp. 209-228, 2001) so that it is not necessary to track a fluid particle in order to calculate its deformation history. However, in this work modifcations to the deformation fields method are introduced: the past time is discretized using a different formula, the Finger tensor Bt0(x; t) is obtained by a second order method and the stress tensor ? (x; t) is computed by a second order quadrature formula. The numerical method presented in this work is validated by simulating the flow of a Maxwell fluid in a two-dimensional channel and the numerical solutions of the velocity and the stress components are compared with the respective analytic solutions providing a good agreement. Further, the flow through a 4:1 planar contraction of a specific fuid studied experimentally by Quinzani et al. (J. Non-Newtonian Fluid Mech., 52, pp. 1-36, 1994) was simulated and the numerical results were compared qualitatively and quantitatively with the experimental results and very good agreement was obtained. The Maxwell and the K-BKZ models were applied to simulate the 4:1 planar contraction problem using various Weissenberg numbers and the numerical results were in agreement with those published in the literature. Finally, numerical results of free surface flows using the Maxwell and K-BKZ integral constitutive equations are presented. In particular, the numerical simulation of jet buckling using several Weissenberg numbers and various Reynolds numbers are presented

Este trabalho tem como objetivo desenvolver um método numérico para simular escoamentos incompressíveis, isotérmicos, confinados ou com superfícies livres, de fuidos viscoelásticos governados pelos modelos integrais de Maxwell e K-BKZ (Kaye-Bernstein, Kearsley e Zapas). A técnica numérica apresentada é uma extensão do método GENSMAC (Tomé McKee - J. Comp. Phys., (110), pp 171--186, 1994 ) para a solução das equações de conservação, juntamente com as equações constitutivas integrais de Maxwell e K-BKZ. As equações governantes são resolvidas pelo método de diferenças finitas em uma malha deslocada. O tensor de Finger, B_t\'(t) é calculado com base nas idéias do método de campos de deformação (Peters et al. - J. Non-Newtonian Fluid Mech. (89), de maneira que não há a necessidade de seguir a trajetória da partícula de fuido para descrever a história de deformação da partícula. Uma abordagem diferente para a discretização do instante passado é utilizada e o tensor de Finger e o tensor das tensões são calculados utilizando um método de segunda ordem. A validação do método numérico descrito nesse trabalho foi feita utilizando o escoamento em um canal bidimensional e a solução numérica obtida para a velocidade e para as componentes de tensão com o modelo de Maxwell foram comparadas com as respectivas soluções analíticas no estado estacionário, mostrando excelente concordância. Os resultados numéricos para a simulação do escoamento em uma contração planar 4 : 1 mostraram bons resultados, tanto qualitativos quanto quantitativos, quando comparados com os resultados experimentais de Quinzani et al. ( J. Non-Newtonian Fluid Mech. (52), pp 1?36, 1994 ). Além disso, utilizando os modelos Maxwel e K-BKZ, o escoamento em uma contração planar 4 : 1 foi simulado para vários números de Weissenberg e os resultados obtidos estão de acordo com os encontrados na literatura. Resultados numéricos de escoamentos com superfícies livres modelados pelas equações integrais de Maxwell e K-BKZ são apresentados. Em particular, a simulação numérica do jato oscilante para diferentes números de Weissenberg e diferentes números de Reynolds é apresentada. / The aim of this work is to develop a numerical technique for simulating incompressible, isothermal, free surface (also con¯ned) viscoelastic flows of fuids governed by the integral models of Maxwell and K-BKZ (Kaye-Bernstein, Kearsley and Zapas). The numerical technique described herein is an extension of the GENSMAC method (Tome and McKee, J. Comput. Phys., 110, pp. 171-186, 1994) to the solution of the momentuum and mass conservation equations together with the integral constitutive Maxwell and K-BKZ equations. The governing equations are solved by the finite difference method on a staggered grid using a Marker-and-Cell approach. The fluid is represented by marker particles on the fluid surface only. This provides the visualization and location of the fluid free surface so that the free surface stress conditions can be applied. The Finger tensor Bt0(t) is computed using the ideias of the deformation fields method (Peters et al. J. Non-Newtonian Fluid Mech., 89, pp. 209-228, 2001) so that it is not necessary to track a fluid particle in order to calculate its deformation history. However, in this work modifcations to the deformation fields method are introduced: the past time is discretized using a different formula, the Finger tensor Bt0(x; t) is obtained by a second order method and the stress tensor ? (x; t) is computed by a second order quadrature formula. The numerical method presented in this work is validated by simulating the flow of a Maxwell fluid in a two-dimensional channel and the numerical solutions of the velocity and the stress components are compared with the respective analytic solutions providing a good agreement. Further, the flow through a 4:1 planar contraction of a specific fuid studied experimentally by Quinzani et al. (J. Non-Newtonian Fluid Mech., 52, pp. 1-36, 1994) was simulated and the numerical results were compared qualitatively and quantitatively with the experimental results and very good agreement was obtained. The Maxwell and the K-BKZ models were applied to simulate the 4:1 planar contraction problem using various Weissenberg numbers and the numerical results were in agreement with those published in the literature. Finally, numerical results of free surface flows using the Maxwell and K-BKZ integral constitutive equations are presented. In particular, the numerical simulation of jet buckling using several Weissenberg numbers and various Reynolds numbers are presented

Depuis une dizaine d’années, l’évolution de la puissance des ordinateurs a permis de développer l’utilisation, dans les études d’ingénierie, des simulations 3D CFD (Computational Fluid Dynamics) pour l’étude de l’atmosphère à petite échelle, en particulier pour la dispersion de polluants sur des sites industriels et urbains complexes. Compte tenu de la complexité des domaines à étudier et des ressources de calcul généralement disponibles, ces études sont la plupart du temps réalisées à l’aide des modèles RANS (Reynolds Averaged Navier-Stokes), et particulièrement avec le modèle de fermeture k – e. Différents travaux de validation de l’approche RANS k – e ont mis en évidence quelques limitations à reproduire la dynamique de l’écoulement et de la dispersion dans des configurations géométriques complexes. Le travail de recherche réalisé dans le cadre de cette thèse a pour objectif une caractérisation expérimentale fine de l’écoulement et de la dispersion turbulente autour d’un obstacle bidimensionnel placé dans une couche limite de surface, afin d’évaluer la validité des modèles RANS en vue de leur application pour l’étude de la dispersion atmosphérique.Dans un premier temps, nous avons utilisé des techniques d’anémométrie à fil chaud, d’anémométrie laser Doppler et d’anémométrie par image de particules, pour déterminer le champ de vitesse dans une couche limite de surface rugueuse et autour d’un obstacle bidimensionnel de section carrée. Une attention particulière a été portée sur l’analyse des termes de l’équation évolutive de l’énergie cinétique turbulente (ECT) et sur la détermination de la viscosité turbulente vt. Différentes approches ont également été utilisées pour estimer le taux de dissipation e de l’énergie cinétique turbulente. Nous avons mis en évidence que ces différentes approches fournissent des résultats comparables dans le cas de la couche limite, tandis que seule la technique estimant e comme le résidu de l’ECT est applicable dans le sillage de l’obstacle. De plus, nos mesures ont permis d’évaluer les paramétrisations du modèle k – e et de montrer que la valeur du coefficient Cμ = 0.09 ne semble pas adaptée dans le cas de la couche limite, conduisant à une surestimation de vt, alors que cette valeur apparait plus adaptée dans le cas de l’obstacle. Une étude de sensibilité, portant la détermination de la constante σk du modèle k – e, indique une contribution non négligeable des termes de corrélation entre la vitesse et la pression dans le sillage de l’obstacle.Dans un deuxième temps, nous avons étudié la dispersion d’un scalaire passif, en mesurant les différents moments statistiques de la concentration, au moyen d’un détecteur à ionisation de flamme. Nous avons également déterminé les flux turbulents de masse, par un couplage entre les mesures de vitesse et de concentration, en prenant soin de contrôler les influences réciproques des deux techniques de mesure. Ces mesures nous ont permis de tester la validité de différents modèles de fermeture de l’équation d’advection-diffusion pour estimer les flux dans le sens vertical et dans le sens longitudinal. Nous avons également pu déterminer expérimentalement le coefficient de diffusivité turbulente Dt, nous permettant d’évaluer un nombre de Schmidt turbulent Sct, afin de mettre en évidence que la valeur Sct = 0.7 est adaptée à la majorité des zones étudiées, excepté dans la zone de recirculation induite par l’obstacle. Enfin, nous nous sommes intéressés aux différents termes de l’équation de la variance de la concentration et plus particulièrement à son taux de dissipation. À nouveau, les mesures nous ont permis de tester un modèle de fermeture disponible dans la littérature et de montrer la bonne cohérence entre le modèle et l’expérience. / In the last decades, there has been an increasing use of Computational Fluid Dynamics (CFD)simulations to evaluate the impact of atmospheric pollutants dispersion in within industrial and urban sites. Given the high geometrical complexity of these sites, these simulations are mainly performed adopting a Reynolds Averaged Navier-Stokes (RANS) approach and using k−e closure models. As is well known from previous studies, RANS k−e simulations are affected by some limitations that prevent them correctly reproducing the dynamics of the flow and the pollutant dispersion in complex geometrical configurations. The aim of the PhD is to provide a detailed experimental characterization of the flow and the turbulent dispersion around an idealized two-dimensional obstacle placed within a boundary layer flow. This is subsequently used to analyse the reliability of RANS closure models as predictive tools for the atmospheric dispersion of airborne pollutants. Initially we focus on the flow dynamics of a boundary layer flow developing over a rough wall and in the wake of a 2D obstacle. The velocity field is investigated experimentally by means of different measurement techniques, namely Hot Wire Anemometry (HWA), Laser Doppler Anemometry (LDA) and Stereo-Particle Imagery Velocimetry (PIV). A particular attention was devoted to the estimate of the turbulent viscosity nt as well as of the terms composing the turbulent kinetic energy budget (TKE), including its rate of dissipation e which was determined adopting different approaches. These measurements allowed us to analyse the accuracy of the parameterizations included in a standard k−e closure model. Our analysis show that a value of the coefficient Cμ = 0.09 leads to significant overestimation of nt in a boundary layer flow. Conversely, adopting Cμ = 0.09 provides a good agreement between modeled and direct estimates of nt in the wake of the obstacle. As a second step, we studied the dispersion of a passive scalar emitted by a ground level line source. To that purpose we measured the first four order moments of the concentration probability density function by mean of a flame ionization detector (FID). Furthermore, the coupling of the FID system with the LDA or HWA system allowed us to directly measure the turbulent mass transfer, i.e. the correlation between velocity and concentration fluctuations. The combination of these two techniques was carefully analyzed, in order to quantify eventual mutual disturbances of one measurement technique on the other. The measurements of the velocity/concentration correlations allowed us to determine experimentally the turbulent diffusivity Dt and the turbulent Schmidt number Sct , and therefore to test the accuracy of different closure models of the advection-distribution equation. Our results show that the value of the turbulent Schmidt number is approximately equal to 0.7 in most of the domain, except in the recirculation zone on the wake of the obstacle. Experimental data provide also a complete description of the spatial distribution of the concentration variance, and of the term composing its budget (with a focus on its dissipation). As for the velocity field, we test the reliability of different closure model proposed in the literature of the turbulent mass transfer terms, enlightening the shortcomings of simple gradient-law closer models.