This thesis investigates applications of the gauge/gravity duality to strongly coupled quantum field theories. After a review of the duality and of correlators and transport in quantum systems, we present our results on second-order non-conformal hydrodynamics. We derive new Kubo formulae for five second-order transport coefficients in non-conformal relativistic fluids. We then apply these Kubo formulae to a class of non-conformal holographic fluids at infinite coupling. We find strong evidence that the Haack-Yarom identity, known to relate second-order coefficients in conformal holographic fluids at infinite coupling, continues to hold in holographic fluids without conformal symmetry: Within our class of models, we prove that it still holds when leading non-conformal corrections are taken into account, and we show numerically that it is also obeyed beyond leading order. This provides further evidence that the identity may be a universal feature of strongly coupled fluids. Next, we present our results on magnetic spin impurities in strongly correlated systems. We build a holographic two-impurity Kondo model, identifying the inter-impurity interaction as double-trace deformation. Our numerical results for the phase diagram suggest a quantum phase transition between a trivial phase with uncorrelated spins and no Kondo screening, and a non-trivial phase with anti-ferromagnetic correlations and simultaneous Kondo screening. Computing the spectrum in the single-impurity case, we observe Fano resonances, which at low temperatures we identify with the Kondo resonance.

This thesis investigates tsunami impact on shore-based, low-rise structures in coastal areas. The aims are to investigate tsunami wave inundation in built-up coastal areas with reference to structural response to wave inundation, to assess the performance of current design codes in comparison with validated state-of-the-art numerical models and to improve structural design of residential buildings in tsunami risk areas. Tsunami events over the past few decades have shown that a significant proportion of fatalities can be attributed to the collapse of building infrastructure due to various actions of the incident waves. Although major tsunami events have demonstrated the potential catastrophic effects on built infrastructure, current building codes have no detailed or consistent guidance on designing structures in tsunami-prone regions. Furthermore, considerable differences in existing empirical formulae highlight that new research is necessary to appropriately address the particularities of the tsunami-induced forces and structure response into the design standards. In this thesis, numerical modelling methods are used to simulate hydrodynamic impact on shore-based coastal structures. The hydrodynamic simulations were conducted using a novel meshless numerical method, smoothed particle hydrodynamics (SPH), which is coupled with the finite element (FE) method to model structural behaviour. The SPH method was validated with experimental data for bore impact on an obstacle using a convergence study to identify the optimum particle size to capture the hydrodynamics. The FE model was validated against experimental data for plates under transient blast loads which have similar load characteristics with impulsive tsunami-induced bore impacts. One of the contributions of the thesis is the use of a new coupling method of the SPH-based software DualSPHysics and FE-based software ABAQUS. Using SPH particle spacing of the same size as the FE mesh size, enables the SPH output pressure to be directly applied as an input to the structural response model. Using this approach the effects of arrangement and orientation of single and multiple low rise structures are explored. Test cases were performed in 2-D and 3-D involving a discrete structure and multiple structures. The 3-D SPH simulations with single and multiple structures used an idealised coastal structure in the form of a cube with different on-plan orientations (0°, 30°, 45° and 60°) relative to the oncoming bore direction. The single structure cases were intended to study the improvement of the resilience of coastal structures by reducing the acting pressures on the vertical surfaces by changing the structure’s orientation. It was found the pressure exerted on the vertical surface of structure can be reduced by up to 50% with the 60° orientation case. The multiple structure models were conducted to examine shielding and flow focusing phenomena in tsunami events. The results reveal that the distance between two adjacent front structures can greatly influence the pressure exerted on the rear structure. This thesis also demonstrates the capability of SPH numerical method in simulating standard coastal engineering problems such as storm waves impact on a recurve wall in 2-D. The idealised structures were represented as standard timber construction and the finite element modelling was used to determine the corresponding stress distributions under tsunami impact. Following the comparison of the method used in this thesis with commonly used design equations based on the quasi-static approach, large differences in stress prediction were observed. In some cases the loads according to the design equations predicted maximum stresses almost one order of magnitude lower. This large discrepancy clearly shows the potential for non-conservative design by quasi-static approaches. The new model for the simulation of tsunami impact on discrete and multiple structures shows that the resilience of a coastal structure can be improved by changing the orientation and arrangement. The characteristics of tsunami waves during propagation and bore impact pressures on structures can be assessed in great detail with the combined SPH and FE modelling strategy. The techniques outlined in this thesis will enable engineers to gain a better insight into tsunami wave-structure interaction with a view towards resilience optimisation of structures vulnerable to tsunami impact events.

En France, 25% des déchets ménagers collectés sont stockés en Installation de Stockage de Déchets Non Dangereux (ISDND). Pour réduire leur impact sur l'environnement, le concept d'ISDND gérée en mode « bioréacteur » a été étudié et évalué depuis plus de dix ans en Europe. Ce concept est basé sur la réinjection des lixiviats, qui consiste à collecter les lixiviats en fond de casier et à les réinjecter sous la couverture de surface. L'optimisation de la répartition du lixiviat nécessite d'estimer les volumes et débits de réinjection et de dimensionner les dispositifs. Afin d'optimiser le dimensionnement des dispositifs de réinjection et d'améliorer la compréhension des écoulements de lixiviat dans les déchets, une voie possible est la modélisation hydrodynamique. Un modèle hydrodynamique regroupe une représentation conceptuelle du milieu poreux et un modèle mathématique permettant de décrire les écoulements. Pour simuler une infiltration, il est nécessaire de renseigner les paramètres hydrodynamiques intervenant dans les équations mathématiques du modèle. Des informations complémentaires sur le milieu sont donc requises pour contraindre les modèles hydrodynamiques et évaluer ces paramètres. Dans le cadre de cette thèse, nous nous sommes intéressés à la méthode géophysique de la tomographie de résistivité électrique (ERT) afin de contraindre les modèles hydrodynamiques à l'échelle d'un casier de déchets. En effet, de nombreuses études ont montré l'intérêt de cette méthode de mesure pour le suivi de la réinjection de lixiviat dans les massifs de déchets. Cependant, cette méthode ne permet pas de mesurer directement la teneur en eau et d'évaluer les paramètres hydrodynamiques. Dans la littérature, plusieurs auteurs ont proposé des méthodes de contrainte des modèles hydrodynamiques à partir des données ERT, basées sur la loi pétrophysique d'Archie. Afin de s'affranchir de l'utilisation de cette loi, peu adaptée aux milieux hétérogènes comme les déchets, il nous a semblé intéressant d'utiliser la forme de l'infiltration extraite de l'ERT pour contraindre les modèles hydrodynamiques.Ainsi, la problématique de cette thèse est d'améliorer la compréhension des écoulements de lixiviat en proposant une méthode de contrainte des modèles hydrodynamiques à partir de la forme de l'infiltration extraite des données ERT.Pour répondre à cet objectif, le travail de thèse a été divisé en trois étapes successives correspondant à chaque partie du manuscrit. La première partie a été consacrée à une synthèse bibliographique concernant la méthode ERT et la modélisation hydrodynamique. La seconde partie avait pour objectif de proposer une méthodologie permettant de délimiter l'infiltration de lixiviat sur les données ERT. La troisième partie a permis de développer une méthode de contrainte des modèles hydrodynamiques à partir de la délimitation de l'infiltration obtenue d'après les données ERT. / In France, approximately 25% of the total amount of collected household waste is stored in municipal solid waste landfills (MSWL). To reduce their impact on the environment, the bioreactor concept was studied and tested for more than a decade in Europe. This concept is based on leachate reinjection, which consists in collecting leachate at the bottom of the waste deposit cell and reinjecting it underneath the landfill cover. The optimization of leachate distribution requires tools to design leachate injection systems (LIS) and to assess volumes and flow rates for the reinjection. To design LIS and to improve the understanding of leachate flow in the waste medium, a subsurface flow modelling procedure may be considered. Hydrodynamic models are based on a conceptual approach to represent the porous medium and a mathematical model to describe flow. To simulate an infiltration into a porous medium, the assessment of the hydrodynamic parameters included in the mathematical equations of the model is required. Thus, additional information is needed to constrain hydrodynamic models and to assess hydrodynamic parameters.This thesis focused on the use of the electrical resistivity tomography (ERT) method to constrain hydrodynamic models at the landfill scale. Indeed, many studies have shown that this geophysical method is suitable to study leachate reinjection into the waste landfills. However, this geophysical method does not allow ERT users to directly measure water content and to assess hydrodynamic parameters. In the literature, several authors proposed constraint methods of hydrodynamic models from ERT measurements, based on the use of Archie's law. To avoid the use of this relationship, which seems to be inappropriate for heterogeneous media such as waste, we found relevant to use the infiltration shape obtained on the ERT results to constrain hydrodynamic models.Thus, the aim of this thesis is to improve the understanding of leachate flow by proposing a constraint method of hydrodynamic models using the infiltration shape extracted from ERT measurements.This thesis was divided into three successive steps corresponding to each part of the manuscript. The first part corresponds to a bibliographic study concerning the ERT method and hydrodynamic modelling. The second part aimed at proposing an interpretation methodology to delimitate an infiltration area from the ERT measurements. The third part allowed us to develop a constraint method of hydrodynamic models using the infiltration delimitation obtained from the ERT measurements with the methodology developed in the second part.

In this thesis, a general methodology for generating representative load cycles for arbitrary monohull surface vessels is developed. The proposed methodology takes a hull geometry and propeller placement, vessel loading condition, vessel mission, and weather data (wind, waves, currents) and, from that, generates the propeller states (torque, speed, power) and steering gear states (torque, speed, power) necessary to accomplish the given mission. The propeller states, together with the steering gear states, thus define the load cycle corresponding to the given inputs (vessel, mission, weather). Some key aspects of the proposed methodology include the use of a surge-sway-yaw model for vessel dynamics as well as the use of surrogate geometries for both the hull and propeller(s). What results is a methodology that is lean (that is, it requires only sparse input), fast, easy to generalize, and reasonably accurate. The proposed methodology is validated by way of two separate case studies, case A and case B (both involving distinct car-deck ferries), with case A being a more ideal case, and case B being a less ideal case given the methodology proposed. In both cases, the load cycle generation process completed in greater than real time, achieving time ratios (simulated time to execution time) of 3.3:1 and 12.8:1 for cases A and B respectively. The generated propeller and steering gear states were then compared to data collected either at sea or from the vessels' documentation. For case A, the propeller speed, torque, and power values generated were all accurate to within +/- 3%, +/- 7%, and +/- 10% of the true values, respectively, while cruising, and accurate to within +/- 14%, +/- 36%, and +/- 42% of the true values, respectively, while maneuvering. In addition, the steering gear powers generated in case A were consistent with the capabilities of the equipment actually installed on board. For case B, the propeller speed, torque, and power values generated were all accurate to within +/- 2%, +/- 8%, and +/- 9% of the true values, respectively, while cruising, and accurate to within +/- 28%, +/- 45%, and +/- 66% of the true values, respectively, while maneuvering. In case B, however, the steering gear powers generated were questionable. Considering the results of the validation, together with the rapid process runtimes achieved and sparse inputs given, one may conclude that the methodology proposed in this thesis shows promise in terms of being able to generate representative load cycles for arbitrary monohull surface vessels. / Graduate

Le projet de recherche consiste à observer en laboratoire la radiation de Hawking, cette prédiction stupéfiante de l'astrophysicien anglais Stephen Hawking faite en 1974 : les trous noirs ne sont pas noirs. Autrement dit, ils n'absorbent pas tout ce qui est à leur portée mais émettent un rayonnement. En plus des complications du fait que ces objets célestes sont à des milliers d'années-lumière, ce rayonnement est tellement faible que cela reviendrait à essayer d'entendre un murmure dans un concert de rock. Mais William Unruh, en 1981, a proposé une solution : utiliser des systèmes hydrodynamiques qui présentent les mêmes équations mathématiques qu'en astrophysique. Plus précisément, dans notre cas, nous utilisons la correspondance entre la propagation des ondes lumineuses au voisinage d'un trou noir et celles des ondes de surface dans un contre-courant rendu inhomogène par la présence d'un obstacle immergé. Pour cela, une compréhension approfondie de la mécanique des ondes de surface est nécessaire (bathymétrie variable, vorticité, non-linéarités…). Du côté technique, une méthode de mesure de surface libre a été développée et optimisée. / The aim of the PhD is the observation the Hawking radiation in the laboratory, this astounding prediction of the English astrophysicist Stephen Hawking made in 1974: black holes are not black. In other words, they do not absorb anything within reach but emit a radiation. In addition to the complications of the fact that these celestial objects are thousands of light years away, this radiation is so weak that it would be like trying to hear a whisper in a rock concert. But William Unruh, in 1981, proposed a solution: to use hydrodynamic systems which have the same mathematical equations as in astrophysics. More precisely, in our case, we use the correspondence between the propagation of light in the vicinity of a black hole and surface waves propagation on a inhomogeneous countercurrent (due to the presence of a submerged obstacle). For this, a thorough understanding of the surface waves physics is necessary (variable bathymetry, vorticity, non-linearities ...). On the technical side, a free surface measurement method has been developed and optimized.

[pt] Neste trabalho a análise do comportamento de sistemas rotativos do tipo eixo-rotormancal é estendida para incluir os efeitos da presença de mancais hidrodinâmicos na resposta dinâmica. Estes efeitos estão associados à não-linearidade da força de reação exercida pelos suportes sobre o eixo e dependem dos deslocamentos, velocidades transversais e da rotação própia do rotor. A modelagem estrutural do sistema é obtida empregando-se o método dos elementos finitos. O eixo é representado pelo modelo de viga de Timoshenko com dois nós, quatro graus-de-liberdade por nó, e a interpolação do campo de deslocamentos é obtida utilizando-se as funções de Hermite. Os rotores são modelados empregando-se elementos de inércia concentrada associada aos graus-de-liberdade de um ponto nodal do modelo. E, na representação dos mancais hidrodinâmicos utilizou-se a equação de Reynolds, com as hipóteses simplificadoras para mancais curtos, obtendo-se a solução para a distribuição de pressão do filme de óleo em forma fechada. Essa distribuição de pressão permite a obtenção dos coeficientes das matrizes e rigidez e de amortecimento associadas aos graus de liberdade do eixo no ponto nodal de representação do mancal. Para a integração temporal do sistema de equações diferencias utiliza-se o procedimento passo-a-passo, tendo-se implementado os métodos implícitos de Newmark e Wilson – teta, na forma incondicionalmente estável. Devido à não-linearidade das equações obtidas com a presença dos mancais hidrodinâmicos, em cada intervalo de tempo utiliza-se o procedimento de Newton-Raphson modificado para a correção da solução numérica obtida com outros resultados analíticos/numéricos disponíveis na literatura. Também, uma representação numérica para mancais hidrodinâmicos segmentados é apresentada, utilizando-se o desenvolvimento teórico para mancais simples. Neste caso a avaliação do procedimento numérico é fornecida comparando-se a solução numérica com resultados experimentais obtidos dos rotores de usina hidrogenada avaliada pelo CEPEL. Em ambos os procedimentos o rotor idealizado de jeffcott é empregado no estudo de casos. Verifica-se que os principais resultados associados aos efeitos da precessão auto-excitada (oil whirl), de chicoteamento (oil whip), e da estabilização dinâmica do sistema são reproduzidos pelos modelos numéricos utilizados. / [en] In this work a formulation for the analysis of shaft-rotor-bearing type rotating systems is extendend to accommodate the effects of hydrodynamic bearings in its dynamic response. These effects, which are associated to the nonlinear force on the shaft at the bearings, are dependent of the transverse displacements, transverse linear velocities an the angular veolicty of the shaft. The structure behavior is modeled by employing the finite element method. The shaft is represented by the two node timoshenko model for bearns, with four desgrees-of-freedom per node and Hermite interpolation functions to represent the displacement fields along the bearn axis. Rotors are modeled by using concentrated inertia elements associated to the shaft degrees-of-freedom at the bearing nodal point. In the numerical analysis considering the time integration of the system differential equation, a step-by-step procedure was employed with the newmark technique in this unconditionally stable form. Due to the nonlearities associated with the hydrodynamic bearings, the solution of the system of equations is obtained using a modified Newton-Raphson precedure at each time step for solution convergence. In the evaluation of the proposed computacional system, comparison with solutions obtained from analytical/numerical results available in the literature are used. Also, a numeric represemtation of tilting-pad bearings is presented using the theory for plain journal bearings, under the same simplified conditions. In this case an evaluation of the numerical procedure is given by comparing calculated solutions with experimental results obtained from the evaluation of a hydrogenaration plant provided by CEPEL-Brazilian Research Center For Eletrobras. In both plain an tilting-pad journal bearing numerical procedures, the idealized Jeffcott rotor is employed as a case study for different operating conditions. As a result, it is shown that the solutions associated to the main oil whirl and oil whip effects and afterwards dynamic stabilization are represented by the proposed numerical procedures employed.

[pt] Na produção de petróleo, durante a recuperação secundária, a injeção de água no reservatório de petróleo com o objetivo de deslocar o óleo até o poço produtor pode levar a um regime de escoamento bifásico onde ocorre a formação de emulsões. As emulsões são um problema para a indústria do petróleo porque produzem perda de carga nas linhas de produção e tornam difíceis os processos de separação óleo-água, gerando altos custos. Este fenômeno ainda não é bem entendido e não é exclusivo do que ocorre no meio poroso porque também está presente nas diferentes etapas da produção de petróleo. Este trabalho foi focado na formação de emulsões no meio poroso de um reservatório, considerando especialmente o escoamento bifásico na escala de poros. Assim foi utilizada uma junção de micro canais em T para descrever o que poderia acontecer na união de duas gargantas de poros em um reservatório de petróleo. Neste caso utilizamos a técnica de formação de gotas por fluxo cruzado estudada e desenvolvida, nas últimas décadas, na área de micro-fluídica. Através da injeção de dois líquidos imiscíveis nos canais que formam a junção, foi estabelecido um regime estável de formação de gotas. Para estudar a influência das diferentes variáveis do processo na formação de gotas, foram variadas as vazões dos líquidos injetados e suas propriedades, como viscosidades e tensão interfacial. Os resultados mostram que os diferentes regimes de escoamento e formação de gotas observados não são só uma função do número de capilaridade da fase contínua, como sugere a literatura para junções micro-fluídicas T de seção retangular. Nos experimentos desenvolvidos neste trabalho, nos quais a seção reta dos canais é oval e a fase contínua é a fase aquosa, as características da fase dispersa, como vazão e viscosidade, tiveram uma grande influência no processo de formação de gotas. / [en] In oil production, during secondary recovery, water injection into the reservoir to displace oil towards the production well generates different phenomena, among which is emulsion formation. Emulsions are a problem for the oil industry, because they change the pressure drop in production lines and hardens the process of separating oil-water, generating high operating costs. Emulsion formation in oil production is not yet well understood and it does not only occur in porous media, emulsion formation also occurs during different stages of oil production. This work was focused on the emulsion formation in porous media specifically considering the pore scale two phase flow. A T-junction microfluidic device was used as a model of the union of two pore throats in a porous media. We studied droplet formation in cross-flowing streams, a technique studied and developed in the past decade in the area of microfluidics technologies. Through the injection of two immiscible liquids in the channels of the device, we established a stable droplet formation regime. To study the influence of different variables that govern drop formation, we varied the flow rates of both immiscible phases and their properties, such as viscosity and interfacial tension. The results show that the different regimes observed are not only a function of capillary number, defined based on the continuous phase, as suggested in the literature for microfluidic T junctions with rectangular cross section. In our experiments the cross section of the channel had an elliptic shape and the continuous phase was aqueous, and the characteristics of the dispersed phase had a great influence in the process of drop formation.

L’oxydation du cyclohexane est l’un des procédés les plus importants dans la chaîne de production du Nylon où l’oxygène et le cyclohexane entrent en contact pour former le cyclohexanol, la cyclohexanone puis l’acide adipique. Le rendement est influencé à la fois par le transfert de l’oxygène et par le mélange des réactifs en phase liquide. Des réacteurs de type colonne à bulles sont généralement utilisés pour fournir une aire interfaciale importante et garantir une agitation efficace en phase liquide. Cependant, la complexité des mécanismes impliqués (hydrodynamique, transfert, réaction, fort taux de vide) rend difficile la prédiction des performances des réacteurs. Ce travail est consacré à l’amélioration des lois de fermetures (quantité de mouvement et transferts) pour la modélisation Euler/Euler des réacteurs industriels utilisés pour le procédé d’oxydation du cyclohexane. Dans un premier temps, des expériences de laboratoire avec le système eau/air ont été réalisées jusqu’à de forts taux de vide (> 30%) pour mesurer les effets collectifs sur la force de traînée et le transfert de masse dans un essaim de bulles homogène. Les résultats ont confirmé que le coefficient de traînée des bulles augmente de manière significative avec le taux de vide alors que de manière surprenante l’effet est très faible sur le transfert. Dans un second temps, des expériences ont été réalisées avec le système cyclohexane/diazote dans des conditions industrielles (P = 1 - 20 bar, T = 30 - 150°C). L’analyse des résultats de transfert en condition industrielle a nécessité la simulation numérique directe du transfert à l’intérieur d’une bulle sphérique / Cyclohexane oxidation is one of the most important processes in the production line of Nylon, where oxygen and cyclohexane get in contact to produce cyclohexanol, cyclohexanone and then adipic acid. The production yield is influenced by both the oxygen transfer and the reactants mixing in liquid phase. Bubble column type reactors are usually used to provide a large interfacial area and efficient liquid phase agitation. However, the complexity of the mechanisms involved (hydrodynamic, transfer, reaction, high void fraction) makes it difficult to predict the performance of such reactors. This work is devoted to improve the associated closure laws of momentum and transfer equations used in Euler/Euler modelling of industrial reactors for cyclohexane oxidation. Bench-scale experiment for air-water system has been firstly carried out to measure the collective effects on the drag force and the mass transfer of a bubble in a homogenous bubble swarm with a high void fraction up to 30%. The results confirmed that bubble’s drag coefficient increases significantly with the void fraction. Meanwhile surprisingly, weak effect has been observed on the transfer. Nextly, pilot experiments with nitrogen-cyclohexane system have been performed under industrial conditions (P = 1 - 20 bar, T = 30 - 150°C). Analysis of the results of transfer under industrial conditions required finally direct numerical simulation of transfers inside a spherical bubble.

Ces travaux concernent l'étude d'un procédé de vapogazéification de biomasse en lit fluidisé circulant. En résultent un ensemble d'outils théoriques (modèles de simulation et de dimensionnement) et expérimentaux (pilote de vapogazéification en lit fluidisé circulant) dans le but de permettre la construction d'un procédé à dimension industrielle. Le dimensionnement, la conception et la construction d'un pilote de vapogazéification de 20 kg/h de biomasse en lit fluidisé circulant ont été réalisés. L'hydrodynamique de ce pilote a été étudiée de manière à isoler les paramètres clés permettant d'en assurer le contrôle. Les résultats obtenus lors de l'opération de pilote ont en outre permis de mieux comprendre les mécanismes d'attrition de particules polydisperses (olivine et char) en lit fluidisé circulant. Pour compléter cette étude, les phénomènes de mélange et de ségrégation des particules ont été étudiés par voie expérimentale et numérique (CFD-3D MFIX). Enfin, un certain nombre de modèles numériques (0D, 1D bulle-émulsion, bilan de population) ont été proposés pour répondre aux besoins techniques d'aide à la conception des lits fluidisés circulants appliqués à la vapogazéification de biomasse. / The vapogasification of biomass in a Circulating Fluidized Bed has been studied in order to enable the design and the erection of an industrial process. Theoretical and experimental tools have been designed to allow an engineering approach. In order to understand the hydrodynamic behaviour of the reactor, a pilot plant has been designed and erected to feed 20 kg/h of biomass. An exhaustive study of process parameters has been carried out, to determine predominant parameters for the process global control. Operating the built system has also permit to understand attrition phenomenon of polydisperse particles (olivine and char) in the Circulating Fluidized Bed. In order to complete those results, a study of mixing and segregation of those polydisperse particles has been carried out in fluidized bed with both experimental and numerical (CFD-3D MFIX) approaches. A significant contribution has also been done to numerical models assisting biomass vapogasification in Circulating Fluidized Beds design: 0D, 1D, bubble-emulsion, population balance).

L'intérêt croissant pour l'intensification des procédés a conduit à l'avènement d'un nombre conséquent de nouvelles technologies. Le projet ANR PROCIP qui a financé cette thèse a pour but de développer un logiciel d’aide à la décision pour aider l’utilisateur industriel dans son choix de technologie optimale pour une application donnée. La méthodologie globale de discrimination des technologies vis-à-vis d’un système réactif est basée sur l’utilisation d’une base de données technologique et sur le calcul de critères de choix. Elle nécessite une bonne connaissance du comportement hydrodynamique des appareils et de leurs performances en termes de transfert de chaleur et de matière. Dans cet objectif, les travaux présentés ici portent notamment sur l’étude du comportement hydrodynamique de milli-réacteurs échangeurs industriels. Des méthodologies expérimentales et numériques de caractérisation ont été mises au point. Elles ont permis d’obtenir des corrélations pour l’estimation des pertes de charge, des coefficients de dispersion axiale et des temps de mélange pour plusieurs milli-réacteurs en fonction des conditions opératoires et de la géométrie des appareils. En outre, une nouvelle méthode numérique est proposée pour la détermination des coefficients de dispersion axiale et des temps de mélange. Elle est appliquée pour prédire l’effet de l’extrapolation des caractéristiques géométriques des appareils sur ces propriétés. Dans une dernière partie, l’impact de la dispersion axiale sur le taux de conversion et la sélectivité de différents schémas de synthèses chimiques est discuté. / The interest for process intensification has leaded to the emergence of a wide panel of new technologies. The aim of the ANR PROCIP collaborative project which has funded this work is to develop a new software program including a methodology for process choice focused on intensified technologies. The global methodology of discrimination between the different technologies with respect to a given reactive system is based on the use of an equipment database and on the evaluation of criterion of choice. This methodology implies a good knowledge of the hydrodynamics of the different reactors and their mass and thermal transfer performances. The purpose of the present work is to develop experimental and numerical methodologies for the hydrodynamic characterization of different industrial milli-heat-exchangers reactors. Pressure drop, mixing time and axial dispersion coefficient correlations are given as function of the operating conditions and the geometrical parameters of the reactors. In particular, a new numerical method using CFD computation for the determination of axial dispersion coefficients and mixing times is presented. This method is used to predict the effect of the scale-up of the geometrical characteristics of an intensified reactor on its hydrodynamic performances. Finally, the impact of axial dispersion on the conversion rate and the selectivity for different chemical synthesis schemes is discussed