Search

International audience; Considering the increasing impact of the environmental concerns in the current worldwide policy, the application of biological processes, namely the bio-fixation of CO2 by microalgae, represents a promising solution and an increasingly attractive strategy. Indeed, these photosynthetic microorganisms have the great capacity to fix and tolerate high CO2 concentrations converting it to biomass and highly valuable molecules. Thus, modeling of microalgae growth represents an essential tool for the optimization of the carbon dioxide consumption in engineered systems such as photobioreactors. In this context, the main goal of this work is the identification of the growth model parameters of Chlorella vulgaris, the model organism used in this study. The growth model developed in this study takes into account the combined influence of light intensity and the total inorganic carbon available per cell. First, an experimental campaign of batch culture was carried out in a well-stirred lab-scale photobioreactor under optimal conditions. Finally, model results of biomass dynamics and total inorganic carbon evolution over time are compared with data of batch and continuous cultures, confirming the accuracy of the identified model parameters.

International audience; Microalgae consume carbon dioxide for their growth and to produce secondary metabolites. One of the control system challenges is to maximize this carbon dioxide consumption by controlling microalgae growth, and by maintaining its growth rate at an optimal value. However, there is no such thing as a microalgae growth rate sensor. This paper proposes a software sensor for the specific growth rate and biomass concentration of a microalgae culture in a continuous photobioreactor using the measurements of Total Inorganic Carbon in the medium. Numerical simulations are included to illustrate the performance of the proposed estimation strategy.

In order to optimise the membrane emulsification process, practical decision tools are needed to select the correct process parameters. The aim of this work is to further the understanding of how the main parameters – transmembrane pressure, wall shear stress and interfacial tension – influence the dispersed phase flux and the droplet size and size distribution. The influence of transmembrane pressure has been analysed by the point of view of the fraction of active pores. The higher the transmembrane pressure, the higher the fraction of active pores: this analysis agrees with our experimental results. We have obtained up to 60% of active pores with d p = 0.8 μm at P tm = 2 bar and up to 47% of active pores with d p = 0.5 μm at P tm = 2 bar. The influence of wall shear stress and interfacial tension have been studied in parallel with the objective to group these two parameters in a single non-dimensional number, the capillary number. The higher the capillary number, the smaller the drops formed: the mean Sauter diameter decreases rapidly when increasing the capillary number from 0.11 to 0.25 and then the mean Sauter diameter reaches a plateau above a capillary number of 0.5. Our results show that to produce small droplets a capillary number above Ca c = 0.5 is necessary, this critical capillary number being independent of the emulsifier used.

Microalgae biotechnology has been focusing on the use of algae in the production of high value compounds. During the last few decades, the intense research effort has been aiming at improving new controls and supervising tools as well as on a good process understanding. This requirement involves a large diversity and a better accessibility to process measurements. Probes or sensors are required to control the process. They are however relatively limited. Classical photobioreactors are usually equipped with temperature, dissolved oxygen and pH probes. These sensors actually provide very little online information on cell growth, viability, metabolic state and production. For the time being, sensors reliability cannot meet industrial bioprocessing requirements. In this context software sensors show numerous potentialities. The central axis of this work is the development of an extended Kalman filter (EKF) for the estimation of biomass concentration based on a dynamic process model in combination with total inorganic carbon measurement. A microalga Porphyridium purpureum was used as a model organism in this study. Numerical simulations and real-life experiments (batch and continuous mode) have been carried out and corresponding results are given in order to highlight the performance of the proposed estimator.

In this study, a model of a continuous pilot photobioreactor for microalgae production is proposed. Three aspects have been studied: the modelling of kinetic growth, the gas-liquid transfer and the hydrodynamics in the photobioreactor. The modelling of each aspect has been developed with the dynamic simulation software SpeedUp, after experimental studies, then validated step-by-step. The connection of these three aspects aims to predict and optimise biomass production of the pilot plant.

In this work, simple models for batch reactor runaway with one-phase and homogeneous pressure relief are examined. Five flow models are detailed for the simulation of one and two-phase flows. Results from the different flow models are then compared. The effect of phase composition on the predictions of a hybrid two-phase flow model are illustrated. Finally, the use of dynamic simulation for the design of pressure relief equipment is described.

A neural network has been used to study steel transformation curves. Some problems occurred like non-linearity of the phenomena, measurements uncertainties, lack of examples in the database or variables interdependencies. Solutions for these problems are successively presented, leading to a fractal neural network able to learn differential equations, qualitiative information and to treat noisy experimental measuremnts. Some validation results are presented.

Arthrospira (Spirulina) platensis Toliara isolated from alkaline and salt lakes in the south-western area of Madagascar is a potential source of proteins that could efficiently fight against food deficiency in developing countries like Madagascar. Up to now, productivity in this country has been low, so a better understanding of the growth conditions of this species is needed to improve its production. Growth experiments were undertaken in bubble columns at laboratory scale. The influence of agitation of the culture, medium salinity (ranging from 13 to 35 g L−1) and CO2 addition (ranging from 0 to 2%, v/v) on growth and protein content was examined. Because Arthrospira cells are fragile, a bubble column without additional mixing gave the best growth. Arthrospira (Spirulina) platensis showed higher specific growth rate (μmax) and protein content for lower salinity. Addition of 1% of CO2 improved the productivity by near 60%. The feasability of semi-continuous culture was demonstrated and optimal culture conditions led to a mean productivity of 0.22 ± 0.03 g L−1 d−1, a mean specific growth rate of 0.015 ± 0.002 h−1 and a protein content of 53 ± 2% of total dry weight.

Physical analysis gives a powerful help in conditioning chemical engineering problems. The present paper demonstrates with transient problems that taking into account physical considerations in formulating them improves their solving, either analytically or numerically. It concludes that using extensive differential variables is more effective than intensive ones.

To assist the design engineer to more flexible and controllable, near optimum cost, heat exchanger networks, a computer program, called CAD-HEN (Computer Aided Design of Heat Exchanger Networks), based on thermodynamic principles, was developed. Dynamic simulation and response analysis of the heat exchanger network can be done by the SpeedUp dynamic process simulator by using a specific model. CAD-HEN handles supertargeting, synthesis, and optimization of a heat exchanger network. The targeting and optimization procedures are graphical and interactive. The user views various plots, such as Total Cost Target profile and Grand Composite Curves. An automatic synthesis algorithm, based on the vertical heat transfer model is used to generate in the stream grid environment a near minimum surface network which is optimized with a simplified loop breaking algorithm. A specific model for SpeedUp was developed in order to proceed at dynamic simulation and detailed control study of an existing heat exchanger network.

We have carried out simulations of steel transformation curves with Neural Networks. For this purpose, it was necessary to use special learning techniques because of the non-ideal nature of the industrial experimental results (insufficient in precision and quantity) that we used. The consideration of examples not used in the learning process led us to introduce knowledge of various types and sources in the Neural Network. It was noted that knowledge that an “expert” might use was particularly helpful for a Neural Network.

A Coupled-System approach (SyDCoD) for student training tool to fault diagnosis in a distillation column has been developed. The basic methodology based on the coupling between symbolic-numeric computing integration in which the symbolical computing is represented by a knowledge base of diagnosis and the numerical computing is represented by a simulation. The Knowledge Base approach involves a structure hierarchical diagnosis and structure pattern-matching by categorisation of component malfunction. The hierarchy is a collection of small individual components coordinated to achieve at an overall diagnosis. This fault hierarchy is particularly effective in handling multiple symptoms and multiple fault situations. This methodology give to the students or the student operators better initiation and better understanding on the failure diagnosis in the distillation column plant. The coupled system problem-solver strategy gives more posibilities to students how to choose the strategy to solve their problem.

The industrial crops of microalgae use processes calling upon the presence of parts of metal nature such as steel 316L type. The goal of this study is to test the electrochemical behavior of this material in a marine culture of microalgae. Porphyridium purpureum was used under a photoperiod of alternation darkness/light 12/12 h, in order to apprehend the problems of biocorrosion involved in the biofouling. The evolution of the free potential of corrosion, according to the position of the samples and for different surface roughness, observations of the surface quality under the electron microscope with sweeping were carried out. The results showed that, overall, the strain P. purpureum does not have a corrosive effect on the 316L. The free potential of corrosion lies between −0.307 and −0.005 V(SCE). The adhesion of the cells seems stronger on the interface air/solid of the half-plunged sample with surface grit polished 1,000, confirmed by the presence of biofilm on the air part. The photoperiod acts on the evolution of the generated free potential of corrosion of the one 24-h period oscillation. Furthermore, the samples plunged horizontally lead to a stabilizing effect on the potential of free corrosion.

International audience; The microalgae biotechnology is a very promising solution for environmental applications. In particular, these photosynthetic microorganisms have a great capacity to fix the carbon dioxide converting it into biomass and other secondary metabolites. Therefore the biological CO2 fixation by microalgae has attracted much attention. The optimization of this biological process by maintaining the algal culture under optimal growth conditions represents a major challenge. Thus microalgae growth models are needed to optimize the carbon dioxide consumption by microalgae in engineered systems such as photobioteactors. In this paper, a procedure to identify parameters of a microalgae growth model is described. First of all, experiments carried out in a lab-scale photobioreactor are presented. Thereafter, a description of the selected growth model, applied to cultures of Chlorella vulgaris, which allows the effective representation of the evolution of the specific growth rate and biomass of the Chlorella vulgaris culture in a perfectly stirred photobioreactor is described. At last, results of our model are compared to experimental data, which confirms the accuracy of the whole procedure.

International audience; In reactive extrusion, the extruder is used as a solvent-free continuous chemical reactor able to process highly viscous materials. The chemical transformation of biopolymers by reactive extrusion appears as a very promising technology. Although punctual applications in this field have already been achieved on a laboratory or pilot scale, the amount of work to carry out is still considerable. A wide range of reactions and raw materials may be explored, and the reactions achieved on a laboratory scale have to be optimized and transposed to an industrial scale. Process modelling and simulation constitute useful tools for process understanding, development, optimization and scale-up. Although reactive extrusion modelling has interested many authors, it still remains a challenge because of the complex geometry and the strong coupling between operating parameters, flow conditions, material rheological behavior and reaction kinetics. A steady-state mathematical model for a biopolymer oxidation process by reactive extrusion is here proposed. The model is based on a hybrid approach combining chemical engineering methods and simplified continuum mechanics laws. The combination of these two approaches enables to simplify the calculations related to chemical reactions while ensuring a predictive character. The flexible structure of the model enabled its implementation within a global process simulator. A method to minimize the amount of experimental data required for model parameter adjustment is also presented. The model was validated by experiments conducted on a semi-pilot corotating twin-screw extruder. Even if it may be refined, the model proposed already constitutes a useful tool for later research work dealing with the development, modelling and simulation of chemical reactions in corotating twin-screw extruders.

Reactive extrusion involves complex interactions between operating parameters, flow conditions, material rheological behavior and reaction kinetics. Although reactive extrusion modelling has interested many authors, it still remains a challenge. We propose here a steady-state reactive extrusion model combining chemical engineering methods and simplified fluid mechanics laws. This steady-state model was derived from the dynamic model proposed by Choulak (2004). A rheo-kinetic model for a biopolymer oxidation process induced by coupled thermo mechanical and chemical effects was developed and integrated into the twin-screw extrusion model. This modelling approach enables to provide a predictive model involving very rapid calculation. The reactive extrusion model was then integrated into a static process simulator. The simulations reproduce available experimental data with a satisfying accuracy.

International audience; In order to optimise the membrane emulsification process, there is a need for a model that can be used to keep the production time of an emulsion to a minimum, without impairing its quality. This was the aim of our work. For that, we have studied the influence of transmembrane pressure on the fraction of active pores and on the drop diameter and the drop diameter distribution. We have found that an increase in transmembrane pressure makes it possible to increase the dispersed phase flux by a rise in the number of active pores (from 1 to 13% for the 0.2 µm membrane and from 2% to 45% and 49% for the 0.5 and 0.8 µm membranes, respectively). Furthermore, the increase in transmembrane pressure leads to an increase in the diameter of the droplets produced. For the 0.5 and 0.8 µm pore diameter, we have even found an increase in the polydispersity of the emulsion produced. A thorough study allowed us to locate the origin of this polydispersity: we ruled out post-production coalescence both in the bulk and in the laminar sub-layer. We showed the possibility of the expansion of the three-phase contact line over more than one pore, phenomenon we referred to as membrane coalescence. Based on these results, we built a model which allows to evaluate the risk of membrane coalescence under experimental conditions of dispersed phase flux and transmembrane pressure: the number of active pores at a distance dd/2 from the centre of a given active pore must be less than 2. For the three pore diameter membranes studied, our model matched the experimental results and could explain the polydispersity observed.

This paper presents a nonlinear predictive control (NMPC) strategy applied to the continuous microalgae cultivation process in a closed photobioreactor. The photo-bioreactor system is programmed to operate in a constant biomass density mode, in order to maintain the culture at the optimal population density and sustain high biomass production levels. This objective is achieved by regulating the biomass density while tracking a reference culture medium feeding profile determined off-line. The proposed NMPC approach is validated in simulation, and performances and benefits for the cell cultivations are compared to those obtained when applying a generic model control (GMC) law.

International audience; This paper addresses a study of the regulation of the biomass density in a closed microalgal photobioreactor by using a linearizing control approach. The photobioreactor system was programmed to operate in a constant biomass density mode, in order to maintain the culture at the optimal population density and sustaining high biomass production levels. The designed control law uses an input-output linearizing control in an inner loop, a proportional integral Derivative (P.I.D.) regulator has been added to cancel steady errors. In addition, this PID-regulator is supplied with an anti-windup compensator. Thus, the proposed strategy takes the internal dynamics into account and stabilizes them simultaneously with the control of the input-output behaviour of the system. Commonly, the output of the system is the biomass concentration. However, in this study, we propose to consider a new output: the ratio between two state variables, which is equivalent to the total inorganic carbon available quantity per cell. This new quantity is shown to lead to best regulation performance. We illustrate our approach with numerical simulations and show its benefits for the cell cultivations, mainly for ensuring achievement of the culture with regards to classical experiments problems.

International audience; BACKROUND: Spirulina platensis (Toliara, Madagascar) provides a food supplement which can fight against malnutrition and food insufficiency in Madagascar. In this country, the current production from three open basins is not sufficient to cover needs, and presents drawbacks such as low productivity and possible contamination. Thus cultivation of S. platensis in a closed photobioreactor opens the possibility of extending this microalga production. In this study, the influence of colour and intensity of the light on S. platensis growth and protein content was investigated in a bubble column.RESULTS: Growth kinetics were obtained for four colours (green, white, red and blue) and four intensities (400, 800, 1000 and 1200 lux) of light. The influence of light colour on Spirulina growth was discussed. The highest productivity (183.6mg L(-1) d(-1)) and concentration (2643mg L(-1)) were obtained for green light at 1200 lux. The protein content was 58%.CONCLUSION: Using green light allows improvement of S. platensis growth. Performance obtained with the closed system was higher than that reported in the literature.

International audience; The first use of microalgae by humans dates back 2000 years to the Chinese, who used Nostoe to survive during famine. However, microalgal biotechnology only really began to develop in the middle of the last century. Nowadays, there are numerous commercial applications of microalgae. For example, (i) microalgae can be used to enhance the nutritional value of food and animal feed owing to their chemical composition, (ii) they play a crucial role in aquaculture and (iii) they can be incorporated into cosmetics. Moreover, they are cultivated as a source of highly valuable molecules. For example, polyunsaturated fatty acid oils are added to infant formulas and nutritional supplements and pigments are important as natural dyes. Stable isotope biochemicals help in structural determination and metabolic studies. Future research should focus on the improvement of production systems and the genetic modification of strains. Microalgal products would in that way become even more diversified and economically competitive.

International audience; Optimization of Nannochloropsis oculata growth was undertaken using the response surface method. A central composite design was defined to study the effects of temperature, pH, incident light intensity and aeration rate on the maximum growth rate of the microalga. Using statistical analysis, the first model calculated to fit results was twice improved. The final model obtained was used to clarify the effects of each factor and their interactions on the growth of Nannochloropsis oculata. The optimum growth conditions of this microorganism were also estimated as 21 degrees C, 52 mu mol photons m(-2) s(-1), pH 8.4 and 14.7 VVH of aeration rate. These conditions were tested and validated experimentally since the maximum growth rate achieved with these parameters, 0.0359 h(-1), is the best reported in this study.

Cet article presente l’une des operations unitaires les plus anciennes, la distillation, dont l’utilisation est tres large dans l’industrie : petrole en premier lieu, mais aussi chimie, pharmacie, agroalimentaire... Elle est abordee en tant qu’operation compartimentee, comportant soit un seul etage (distillation simple), soit plusieurs etages ; la circulation des deux phases liquide et vapeur se fait alors a contre-courant (rectification). Le fonctionnement reel des etages est tres complexe. On l’approche cependant avec une precision acceptable par le modele de l’etage theorique ou ideal. Un etage ideal assure deux fonctions essentielles : les phases effluentes sont en equilibre thermodynamique ; la separation physique (desengagement) des deux phases effluentes est parfaite : il n’y a pas d’entrainement mecanique d’une phase dans l’autre. Cette definition depasse le cadre de la distillation et peut s’appliquer aux autres operations unitaires que nous abordons dans les articles suivants (extraction liquide-liquide, gaz-liquide, etc.). Toutefois, dans le cas particulier de la distillation, ou transferts de matiere et de chaleur sont concomitants, on suppose de plus que l’etage est adiabatique et isobare. Bien qu’elles apportent de grandes simplifications, ces deux hypotheses ne sont pas indispensables. L’interet fondamental de ce modele est de donner la possibilite de dimensionner un appareil sans avoir recours ni a des donnees cinetiques (equilibre thermodynamique atteint), ni a des donnees relatives aux ecoulements : l’absence d’entrainement ainsi que de transport de matiere entre les etages permet de donner aux relations de bilan interetage (relations operatoires) une forme particulierement simple. Ses inconvenients sont lies aux hypotheses ; on determine un nombre d’etages ideaux mais, de par la nature du modele, on ne peut donner aucune indication sur les points suivants : relation entre le nombre d’etages theoriques et le nombre d’etages reels ; dimensionnement des etages reels. Ces points ne peuvent etre precises qu’en recourant soit aux regles de l’art (l’experience), soit a des modeles moins simples (cas des operations compartimentees non ideales Transfert de matiere- Autres operations compartimentees ). Toutefois, appuye sur l’extrapolation raisonnee des installations existantes, ce modele permet une estimation souvent precise du dimensionnement des appareils a compartiments separes. Il est universellement employe pour les etudes d’avant-projet. Deux types de problemes peuvent etre traites avec le modele de l’etage ideal : le dimensionnement d’une installation neuve et l’adaptation d’une installation existante. La presentation de la distillation est ici faite en adoptant le premier point de vue. La recherche des solutions repose systematiquement sur des methodes graphiques : la comprehension en est facilitee et la transposition de ces methodes iteratives au calcul automatique est aisee. La methode de Ponchon-Savarit, la plus generale, est presentee systematiquement. La methode de MacCabe et Thiele est introduite dans les cas particuliers ou elle apporte une reelle simplification (cas des flux molaires constants des deux phases). Ces considerations, developpees ici dans le cadre de la distillation, restent valables pour les autres operations unitaires decrites dans les articles suivants.

Simulation and optimization of multi-stage adiabtic catalytic reactors is now well developed in chemical industries, design and companies and universities. In a first part we present briefly a program package developed in our laboratory, first for chemical industry, and secondly for chemical engineering education. In this presentation we show how it is possible to take into account new possibilities in direct optimization procedures and new terms in objective functions. In a second part we examine design problems of optimization, including many sub-problems in the definition of objective functions. Then we try to show how it is possible to sensibilize our students on the possibilities of choice for decision variables by using this package.

International audience

Faced with the growing awareness of environmental issues, the research turns to the development of bioprocesses to develop new solutions to environmental problems. This thesis concerns the study of the feasibility of a process for CO2 capture from the culture of the microalgae Chlorella vulgaris in a continuous photobioreactor. This work has identified the algae C. vulgaris as a promising candidate for this application. Indeed C. vulgaris has a capacity of biomass production and CO2 biofixation very interesting for this application. Studies in this thesis allowed us to update the complex interactions between the algal cells and high CO2 concentrations. Also they have provided a deeper understanding of existing locks for the development of a process for CO2 capture and the need to take into account all the parameters of culture (light, nitrate concentration). A model for the growth of C. vulgaris in continuous photobioreactor has been proposed. This model shows good correlation with experiments; although future studies are still needed to be able to fully simulate the behaviour of algae in photobioreactor cultures. Finally a study of pre-design has been proposed allowing highlighting the many questions that still exist before being able to adapt the laboratory process to an industrial scale., Face à la montée de la prise de conscience des enjeux écologiques actuels, la recherche se tourne vers le développement des bioprocédés pour développer de nouvelles solutions aux problèmes environnementaux. Cette thèse porte sur l’étude de la faisabilité d’un procédé de capture de CO2 à partir de la culture de la microalgue Chlorella vulgaris en photobioréacteur continu. Ce travail a permis d’identifier l’algue C. vulgaris comme une candidate prometteuse pour cette application. En effet C. vulgaris présente une capacité de production de biomasse et de fixation de CO2 très intéressante pour cette application. Les études menées lors de ce travail de thèse ont également permis de mettre à jour les interactions complexes entre les cellules algales et le CO2 présent à de fortes concentrations. De même, elles ont apporté un approfondissement à la compréhension des verrous existants pour le développement d’un procédé de captage du CO2 et de la nécessité de prendre en compte tous les paramètres de culture (lumière, concentration en nitrate). A partir des études menées, il a été possible de proposer un modèle pour la croissance de C. vulgaris en photobioréacteur continu. Bien que de futures études soient encore nécessaires pour être en mesure de parfaitement modéliser le comportement de l’algue lors de cultures en photobioréacteur, ce modèle présente une bonne corrélation avec les expérimentations. Enfin une étude de pré-dimensionnement a été proposée qui a permis de mettre en lumière les nombreux points d’interrogations encore existants avant d’être en mesure d’adapter le procédé de laboratoire à une échelle industrielle

The development of biorefineries requires integrating and optimizing plants and handling a large number of material flows and unit operations. The development of a process simulator dedicated to this field would thus be of great interest. This is what we intended to initiate by relying on the example of the oxidation of biopolymers by reactive extrusion. Reactive extrusion is characterized by a strong coupling between flow, heat transfer and reaction kinetics. This coupling depends on the desired reactions. We here intended to elaborate aflexible model, being easily integrated into a static process simulator, and enabling to reach agood compromise between the predictive character of the model and the amount of experiments required to adjust model parameters. Therefore, we adopted a hybrid modelling approach combining a flow description based on ideal reactors and continuum mechanics laws. Flow is modeled as a cascade of continuous stirred tank reactors (CSTR) with possible backflow. Flow rates between CSTRs are calculated using physical laws taking into account the operating conditions and geometric parameters of the equipment. Each CSTR is characterized by a filling ratio, which depends on the operating conditions. The calculation of steady-state filling ratio, pressure and flow rates between the CSTRs is achieved by performing a material balance in each CSTR. Material temperature in each CSTR is calculated through a thermal balance. The chemical modification of the material is described using three reactions: the oxidative depolymerization, the formation of functional groups(carbonyl and carboxyl) and the thermomechanical degradation of the biopolymer induced by heating and shearing. The number-averaged and weight-averaged molecular weight of the biopolymer and the oxidant content in each CSTR are computed simultaneously by applying the moment operation to population balance equations. Viscosity is linked to the mean molecular weight. An iterative algorithm enables to couple material balance, thermal balance and reaction kinetics. The experimental data required for model validation were provided by the experimental platform developed at the CVG (Centre de Valorisation des Glucides,Amiens, France) in the frame of the Synthons program. A method was proposed in order to adjust model parameters with a minimal number of experimental data, enabling to assess the predictive character of the model. Once the parameters were adjusted, the reactive extrusion model enabled to reproduce the experimental results obtained with different raw materials,flow rates, screw rotation speeds, and using two extruders with different size and screw configuration. The integration of the reactive extrusion model into a process simulator - the USIM PAC software - enabled to simplify its implementation. This constitutes a promising step in a perspective of process optimization and scale-up, and enables to simulate a reactive extrusion operation within a global plant simulator.; Le développement des bioraffineries repose sur une conception optimisée d’installations industrielles en synergie comportant un grand nombre de flux de matière et d’opérations unitaires. Le recours à des simulateurs de procédés présente un intérêt certain dans la conception, l’analyse et l’optimisation de tels procédés. Souhaitant initier le développement d’un outil de simulation adapté à ce secteur, nous nous sommes appuyés sur l’exemple d’un procédé d’oxydation de biopolymères par extrusion réactive. Les procédés d’extrusion réactive sont caractérisés par un couplage intime entre écoulement, thermique et cinétiques réactionnelles. Les modalités de ce couplage dépendent des réactions visées. Souhaitant proposer un modèle flexible, intégrable dans un simulateur statique de procédés, et permettant d’atteindre un bon compromis entre la prédictivité et la quantité d’essais nécessaires pour ajuster les paramètres du modèle, nous avons opté pour une approche de modélisation mixte reposant à la fois sur une représentation de l’écoulement à l’aide de réacteurs idéaux et sur des lois de la mécanique des fluides. L’écoulement est modélisé par une cascade de réacteurs continus parfaitement agités (RCPA) avec reflux. Chaque RCPA est caractérisé par un taux de remplissage qui dépend des conditions opératoires. Le calcul du taux de remplissage des RCPA, de la pression matière et des débits circulant entre les RCPA en régime permanent est effectué en réalisant un bilan matière sur chaque RCPA. La température matière dans chacun des RCPA est calculée grâce à un bilan thermique. La modification chimique du matériau est décrite à l’aide de trois réactions : l’oxydation dépolymérisante, la formation de groupements fonctionnels (carbonyles et carboxyles) et la dégradation thermomécanique du biopolymère sous l’effet de la chaleur et des contraintes de cisaillement. L’établissement des équations de bilan de population auxquelles on applique la méthode des moments, permet de calculer simultanément les masses molaires moyennes en nombre et en poids du polymère ainsi que la teneur en agent oxydant dans chacun des RCPA. La viscosité est reliée à masse molaire moyenne. Un algorithme de calcul itératif permet de coupler le bilan matière, le bilan thermique et le calcul réactionnel. Les données expérimentales nécessaires à la validation du modèle ont été fournies par la plate-forme expérimentale mise au point au CVG (Centre de Valorisation des Glucides, Amiens) dans le cadre du programme Synthons. Une méthode d’ajustement des paramètres du modèle à partir d’un nombre minimal de données expérimentales a été proposée, permettant d’évaluer le caractère prédictif du modèle. Le modèle d’extrusion réactive ainsi ajusté a permis de reproduire les résultats expérimentaux obtenus pour différents matériaux, débits, vitesses de rotation, et sur deux extrudeuses detaille et de configuration différentes. L’intégration du modèle d’extrusion réactive dans un simulateur de procédés - le logiciel USIM PAC - a permis de simplifier sa mise en œuvre,offre des perspectives en optimisation et dimensionnement d’équipement et rend possible la simulation de l’opération d’extrusion réactive au sein d’une chaîne de transformation complète.

This thesis deals with the control of closed microalgal photobioreactor. Mathematical modeling and model parameters identification are particularly approached. The kinetics and light transfer model parameters were obtained by fitting the experimental data acquired along the exponential phase of batch cultivation of Porphyridium purpureum under different conditions in the photobioreactor (2.5 liters). Biomass is one of the most valuable variables to control. Because biomass concentration is very difficult to measure on-line, a software sensor for its estimation is proposed. An extended Kalman filter (EKF) is developed, based on a dynamic process model in combination with total inorganic carbon measurement within an algorithm of state observation. This software sensor leads to good results and was experimentally validated in a new bubble column photobioreactor (9.6 liters) which carried different probes and sensors necessary to implement a control system. Several control strategies for microalgae culture were also proposed: non-linear and linear controllers, PID controller, generic model controller (GMC), input-output linerarizing controller and non-linear predictive controller. We illustrated our approaches with numerical simulations which allowed us to choose the most successful strategies and to validate them on the instrumented photobioreactor., Cette thèse porte sur la commande d’un photobioréacteur continu pour la culture des microalgues. La modélisation du système et l’identification des paramètres du modèle ont été particulièrement abordées. Les paramètres cinétiques et de transfert de l’énergie lumineuse ont été obtenus à partir des données acquises le long de la phase exponentielle de la culture de Porphyridium purpureum en mode batch sous différentes conditions expérimentales dans un photobioréacteur de 2,5 litres. La concentration en biomasse est la quantité physique la plus importante à réguler. Puisque elle n’est pas toujours facile à mesurer en ligne, un capteur logiciel permettant d’estimer la quantité de biomasse produite à partir des mesures en ligne est proposé. Il s’agit d’un filtre de Kalman étendu (EKF) qui combine le modèle du système avec la mesure de la concentration du carbone inorganique total et d’autres variables physiques (pH, intensité lumineuse,…) afin d’estimer en ligne la concentration de la biomasse. Ce capteur logiciel a été validé expérimentalement en utilisant un nouveau photobioréacteur de 9,6 litres de type colonne à bulles, équipé de plusieurs dispositifs destinés à mettre en place un système de commande et de régulation, et de techniques de mesure permettant d’accéder à l’évolution des principales variables. Plusieurs démarches pour la commande de la culture de microalgues ont été également proposées : commande non-linéaire et linéaire, régulation par PID, commande par modèle générique GMC, commande linéarisante par retour d’état et commande prédictive non-linéaire. Ces différentes commandes ont été mises oeuvre en simulation ce qui nous a permis de choisir les stratégies les plus performantes et de les valider sur le photobioréacteur instrumenté.