Emulsions constitute a wide range of natural as well as processed products. Pharmaceutical applications of emulsions include oral administration, parenteral delivery, ophthalmic medicine, topical and transdermal creams, and fluorocarbon-in-water emulsions for blood oxygenation. In the foods area many of the products like mayonnaise, margarine, ice-creams are emulsions by nature and some products can also be used for delivery of active ingredients (e.g. nutraceuticals) with potential health benefits. Emulsions are also encountered at many stages of petroleum recovery, transportation, and processing. Typically, emulsions are manufactured in a two-step process. First a coarse emulsion called a premix is made which is passed through a high-pressure homogenizer. Intense energy supplied in the high pressure homogenizer causes breakage of the coarse emulsion to a fine one with a tighter distribution. Population balance equation (PBE) models are useful for emulsions since they allow prediction of the evolution of the drop size distribution on specification of the two rate processes i.e., breakage of drops due to the flow field and coalescence of colliding drops. In our work, we developed a PBE model to describe emulsion breakage in a high pressure homogenizer. The focus of the work was breakage and conditions to keep coalescence to minimum were implemented. Two breakage rates representing two mechanisms i.e., turbulent inertial and turbulent viscous breakage were necessary for reproducing the bimodal nature of the distributions. We used mechanistic functions in the PBE model to develop a predictive model which could be extended to changes in formulation variables as well as process variables. Starting with the assumption of binary breakage, the model was refined to include multiple drop breakage. The developed model was found to be extensible to reasonable changes in oil concentration, surfactant concentration, continuous phase viscosity and constant ratio of oil to surfactant. Anomalies in pressure prediction encountered earlier were also corrected for by including some additional features like heating, maximum stable diameter, and number of daughter drops. A preliminary attempt was also made to use the developed model for designing experiments for making target emulsions with pre-specified properties.

Fossil fuel is an important energy source, but is unavoidabiy running out. Since the cellulosic material is the most abundant source of organic matter, the ethanol, which is produced from cellulosic waste materials, is gaining more and more attention. These materials are cheap, renewable and their availability makes them superior compared to other raw materials. The cellulose must be hydrolyzed to glucose before it can be fermented to ethanol. The enzymatic hydrolysis of cellulose using cellulase enzymes is the most widely used method. The production cost of cellulase enzymes is the major cost in ethanol manufacture. To optimize the cost of ethanol production, enzyme stability needs to be improved through maintaining the activity of the enzymes and by optimizing the production of the cellulase. The aim of researchers, engineers and industrials is to get more biomass for the same cost. The filamentous fungus Trichoderma reesei has a long history in the production of the cellulase enzymes. This production can be influenced strongly by varying the growth media and culture conditions (pH, temperature, DO, agitation,... ). At present, it is my opinion that no modelling study has included both the hydrodynamic and kinetic aspects to investigate the effect of shear and mass transfer on the morphology of microorganisms that influence the rheology of the broth and production of cellulase. This thesis presents the development of a mathematical model for cellulase production and the growth of biomass in an airlift bioreactor. The kinetic model is coupled with the methodology of two-phase flow using mathematical models based on the bubble break-up and coalescence to predict mass transfer rate, which is one of the critical factor in the fermentation. A comparison between the results obtained by the developed model and the experimental data is given and discussed. The design proposed for the airlift geometry by Ahamed and Vermette enables us to get a high mass transfer and production rate. The results are very promising with respect to the potential of such a model for industrial use as a prediction tool, and even for design.

La mécanique des fluides numérique (CFD) est utilisée pour modéliser la précipitation du carbonate de baryum dans un réacteur à lit fluidisé. L’étude est divisée en deux parties : la modélisation de l’hydrodynamique du réacteur et la modélisation de la précipitation du carbonate de baryum. Pour ces deux parties, les modèles sont validés par l’expérience. Dans la première partie de l’étude, des réactions instantanées de neutralisation en absence et en présence de particules solides inertes sont mises en œuvre dans le réacteur à lit fluidisé. Pour représenter la réaction chimique dans la phase liquide, plusieurs modèles son testés : le modèle Eddy Dissipation (EDM), le modèle Eddy Dissipation Concept (EDC) et le modèle modifié Eddy Dissipation- Multiple Time Scale turbulent mixer (EDM-MTS). Le modèle qui donne la meilleure prédiction de la réaction chimique est choisi : il s’agit du modèle EDM-MTS, qui demande aussi le moins de temps de calcul. Dans la deuxième partie, l’équation de bilan de population est incorporée au code de calcul pour prédire la distribution de taille des particules (PSD). La méthode des classes est implémentée couplée avec le modèle des écoulements multiphasiques Eulérien-Eulérien, le modèle de turbulence k-e et le modèle EDM-MTS. Les cinétiques de précipitation de BaCO3 sont incluses dans le modèle. Des expériences de précipitation du BaCO3 en réacteur à lit fluidisé permettent de valider le modèle de CFD développé. La PSD donnée par le modèle de CFD est en bonne concordance avec les résultats expérimentaux / Computational Fluid Dynamics (CFD) techniques are used to model the precipitation of Barium Carbonate in a solid-liquid fluidized bed reactor. The study is divided in two sections: The hydrodynamic behavior and the barium carbonate precipitation. The CFD model is validated with experimental results for both cases. In the first part, a neutralization reaction in the fluidized bed column with and without solids is carried out. In order to simulate the liquid phase reaction, the Eddy Dissipation Model (EDM), the Eddy Dissipation - Multiple Time Scale turbulent mixer model (EDM-MTS) and the Eddy Dissipation Concept (EDC) reaction models are tested. The model EDM-MTS is chosen for giving the best approach and for being the less computationally expensive. In the second part, the population balance equation is added to the model in order to calculate the Particle Size Distribution (PSD) in the fluidized bed reactor. The discrete method is chosen to solve the population balance equation coupled with the multi-phase Eulerian-Eulerian model, the k-e turbulence model and the EDM-MST model. The nucleation, growth and aggregation kinetics of BaCO3 are included in the precipitation model. The experimental BaCO3 precipitations realized in a fluidized bed reactor allowed the CFD precipitation model validation. The PSD obtained by CFD are in good agreement with the experimental PSD

Un système consistant d'une simulation de mécanique des fluides numérique (MFN) couplée à un modèle de bilan de population (PBM) est développé afin d'étudier l'effet des paramètres variés sur la performance d'un procédé de polymérisation en émulsion qui conduit à la peoduction des particules de polymère dans un milieu aqueux continu.Comme une grande gamme des produits polymériques, des latexes sont les « produits par processus » (products-by-process), et leurs propriétés sont déterminés pendant la polymérisation. LA distribution de la taille des particules (PSD) est une des plus importants paramètres qui influence la qualité finale de latex. La modélisation d'évolution du PSD est généralement réalisée par l'addition un ensemble des PBEs au modèle cinétique. Le PBE fournit un moyen d'étudier la contribution des différents phénomènes dans l'évolution du PSD, comme la nucléation, la croissance des particules par la polymérisation, et la coagulation des particules à cause du mouvement brownien ou le mouvement du fluide (la coagulation Perikinetic et Orthokinetic, respectivement).Afin d'évaluer l'impact du mélange non homogène et les paramètres physiques du système sur l'évolution du PSD du latex, la simulation transitoire d'écoulement à été réalisé avec l'aide d'un progiciel commercial de MFN (Fluent® 15.0) pour munir dans chaque pas du temps, les concentrations locales des espèces ioniques (pour déterminer le taux de la coagulation Perikinetic modelé par le modèle de DLVO) ainsi que certains paramètres hydrodynamiques comme le taux de dissipation de la turbulence et le taux de cisaillement (afin de déterminer le taux de la coagulation Orthokinetic). Cette information est appliquée simultanément par le module complémentaire de PBM dans Fluent pour calculer le PSD pour le prochain pas du temps ; ainsi, un couplage complet entre le MFN et le PBM est assuré / A framework, consisting of a computational fluid dynamics (CFD) simulation model coupled to a population balance model (PBM) is developed to study the effect of various parameters on the performance of an emulsion polymerization process which leads to the production of a fine dispersion of polymer particles in a continuous aqueous medium.Like most polymer products, latexes are “products-by-process”, whose main properties are determined during polymerization. One of the main parameters influencing the final quality of the latexes is the particle size distribution (PSD). Modeling the evolution of PSD is usually accomplished through the addition of a set of PBEs to the kinetic model. PBE provides a means of considering the contribution of different phenomena in the PSD evolution, being nucleation, growth of polymer particles by polymerization, and coagulation of particles due to brownian or fluid motion (Perikinetic and Orthokinetic coagulation, respectively).To assess the impact of nonhomogeneous mixing and physical parameters of the system on the evolution of the latex PSD, the transient simulation of flow was performed with the aid of a commercial CFD Package (Fluent® 15.0) to provide in each time step, the local concentrations of ionic species (to determine the rate of perikinetic coagulation modeled by DLVO model) and certain hydrodynamic parameters such as turbulence dissipation rate and shear rate (to determine the rate of orthokinetic coagulation). This information is applied simultaneously by the PBE add-on module of Fluent to calculate the PSD for the next time step; thus a complete coupling between CFD and PBM is assured