La présente thèse, concernant l’ingénierie de la photosynthèse naturelle et notamment l’optimisation des procédés de production de microalgues, a été menée à l'Institut Pascal. L'approche suivie repose sur la construction de modèles de connaissance des photobioréacteurs, capables d'en prédire les performances quelle que soit leur géométrie, les conditions d'éclairement, ou le type de microalgues cultivées. Ces modèles de connaissance permettent de dégager des stratégies de conception et de conduite du procédé, qui sont utilisées pour développer et réaliser des démonstrateurs performants, à l'échelle pilote. Cette thèse a pour objet d'étude un photobioréacteur pilote de 24 litres utiles à agitation pneumatique dans lequel la lumière est apportée dans le volume réactionnel par 1000 fibres optiques à diffusion latérale. Sa conception repose sur le principe innovant de Dilution Contrôlée du Flux solaire en Volume (DiCoFluV) dont l'objectif est d'atteindre l'efficacité thermodynamique maximale de la photosynthèse. Au cours de cette thèse, le réacteur a été étudié en conditions d’éclairement parfaitement contrôlées, grâce à des lampes à décharges. Il a été rendu fonctionnel (en incluant sa régulation et son amélioration), puis caractérisé en termes d'hydrodynamique et de transfert radiatif. Notamment, le flux radiatif incident à la surface des fibres optiques a été déterminé par des expériences d'actinométrie (sel de Reinecke), grâce à un traitement original permettant l'analyse de situations avec absorption partielle du rayonnement et en géométrie complexe. Ensuite, une année de culture ininterrompue de la cyanobactérie Arthrospira platensis (dont six phases de fonctionnement continu) ont permis de mesurer la vitesse volumétrique moyenne de production de biomasse et l'efficacité thermodynamique du pilote. Pour les faibles densités de flux testées (dilution), les résultats expérimentaux ont montré l’apparition d’un phénomène de photorespiration qui a été intégré dans le modèle de couplage thermocinétique. De plus, ce manuscrit présente également l'utilisation de la méthode de Monte Carlo pour la résolution des modèles radiatifs dans la géométrie réelle du procédé, définie à partir d’une conception assistée par ordinateur, ce qui est une nouvelle avancée pour le traitement des géométries complexes. Après validation de l'adéquation entre les mesures expérimentales et l'estimation prédictive par le modèle, des premières simulations du pilote en fonctionnement solaire ont été menées, à partir de bases de données solaires (DNI). Les résultats obtenus donnent des premières indications quant aux paramètres d'ingénierie (en particulier le facteur de dilution) menant à une productivité surfacique moyenne annuelle maximale, en fonction de l'implantation géographique. / The present PhD dissertation deals with photobioreaction engineering for efficient microalgae production. The approach is based on the construction of knowledge models that permit predicting the performances of the process, whatever its design, the illumination conditions, or the microalgae species cultivated. These models are used to establish optimal design and control strategies that are implemented to construct and operate pilot-scale plants. Here, a 24 liters air-lift photobioreactor is studied, that is based on the principle of incident solar light-flux dilution for approaching the maximum thermodynamic efficiency of natural photosynthesis. For that purpose, the culture volume is internally illuminated by 1000 light-diffusing optical fibers. As a first step toward solar production, this PhD work focuses on perfectly controlled illumination conditions ensured by discharge lamps. First, the reactor hydrodynamics and radiative transfer are characterized. In particular, incident light-flux density at the fiber surfaces is measured by actinometry (with Reinecke salt), thanks to a novel treatment enabling analyses of situations with partial light absorption and complex geometries. Then, the mean volumetric rate of biomass production and the process efficiency are measured based on one year of continuous Arthrospira platensis culture. For the low radiative flux densities tested (dilution), photorespiration by the cyanobacterium is observed and included in the thermokinetic model. Moreover, this dissertation includes a presentation of the Monte Carlo method for solving the radiative transfer equation withinthe complex geometry of the computer aided design used for manufacturing the reactor. After validation against experimental results, the model is predictively used to simulate the pilot operating with natural solar light, based on solar DNI databases. These results indicate engineering parameters (in particular the dilution factor) for optimal yearly-averaged surface productivity, as a function of Earth location.

Předložená diplomová práce byla zpracována s cílem vytvořit přehled poznatků v oblasti akvaponické potravinové produkce. Informace získané během tvorby tohoto přehledu pak měly vést, v kombinaci s daty získanými z funkčního provozu, k vytvoření matematického modelu akvaponického cyklu. Na akvaponické farmě provozované společností Flenexa plus s.r.o., která byla zdrojem potřebných procesních dat, měla být dále zpracována a vyhodnocena bilance energie a vody. Nakonec měla být v průběhu práce posouzena možnost implementace mikrořasového fotobioreaktoru do akvaponického cyklu. Úvod práce představuje motivaci vedoucí k potřebě inovovat dnešní potravinovou produkci. Kriticky jsou zhodnoceny predikce vývoje lidské populace, a to pak hlavně z pohledu dopadu, který by tento růst měl na zemědělskou produkci. Současná situace se na základě získaných poznatků ukazuje jako neudržitelná, primárně pak v oblastech vodohospodářství a energetické spotřeby. Následně je jako možné řešení vedoucí ke zlepšení udržitelnosti potravinové produkce zkoumána akvaponie. Akvaponie je definována a její jednotlivé komponenty jsou představeny z hlediska mechanismu jejich fungování a z pohledu jejich návrhu. Mezi popsané oblasti patří například principy tzv. coupled a decoupled akvaponie a popis možných typů hydroponického komponentu. V této části práce je pozornost věnována také představení cyklů jednotlivých živin v rámci akvaponie. Následující a poslední teoretická část práce je pak věnována mikrořasovému fotobioreaktoru. Jsou zde popsány mechanismy, jak motivující, tak odrazující od zakomponování bioreaktoru do akvaponie. V oblasti výhod se jedná hlavně o jeho roli ve stabilizaci pH a spotřebě toxikého amoniaku. Na druhou stranu jeho ekonomické dopady na profitabilitu akvaponie jsou velmi proměnlivé v závislosti na způsobu implementace. Samotný mikrořasový fotobioreaktor je pak v práci detailněji představen. Jednotlivé procesní ukazatele ovlivňující růst řas jsou rozebrány, a to společně s jednotlivými typy fotobioreaktoru, metodami sklizně a využitími pro vyprodukované mikrořasy. Na základě poznatků schromážděných v této práci pak lze jako nejvhodnější k implementaci do akvaponie doporučit hybridní fotobioreaktory, u kterých je většina osvětlení zajištěna v podobě slunečního svitu. Samotná experimentální část práce pak začíná popisem zkoumaného provozu společnosti Flenexa plus s.r.o. z pohledu aplikovaného akvaponického procesu. Jednotka podrobená měření byla provozně stabilní a využívala implementace hydroponického komponentu typu Deep Water Culture (DWC). Spolu s detailním popisem celého provozu jsou poskytnuty a vyhodnoceny vypracované bilance vody a energií. Pozornost je pak přesunuta k matematickým modelům vypracovaným a ověřeným na základě dat a poznatků shromážděných z provozu společnosti Felenexa plus s.r.o. Logika a algoritmy, na kterých jsou oba modely postaveny, jsou v této části vysvětleny a diskutovány společně s hlavními funkcemi a schopnostmi obou modelů. První, primárně statistický model je představen jako nástroj pro použití při uvádění akvaponie do provozu. Druhý, fyzikální model pak v uživatelsky přívětivém formátu představuje základ pro model řízení akvaponické farmy s mikrořasovým fotobioreaktorem. V neposlední řadě jsou nastíněny také cesty možného budoucího vývoje pro oba vytvořené modely. Práce je následně završena shrnutím a diskusí nad poznatky a výstupy získanými během celého tvůrčího procesu.

This diploma thesis summarizes the knowledge about microalgae, their use, cultivation methods and obstacles that prevent their wider use. In the practical part of the work, a device was designed, constructed, and programmed. This device can analyze the bubbles of the tubular photobioreactor and, based on the obtained data, control its aeration. The Python programming language was used to create the program and the OpenCV library was used to analyze the photographs. The bubble detection is based on the edge detection and the subsequent refinement. The data obtained from the analysis are displayed on the device screen and the data are also stored in a csv file. The discussion lists possible improvements and lessons learned during the creation of this device.

Mikroalgen bieten eine hohe Variation unterschiedlichster Wertstoffe. Hochwertige Lipide, wie Omega-3- Fettsäuren, oder Antioxidantien, wie Astaxanthin und Phycocyanin, werden schon heute in großen Mengen aus Mikroalgen gewonnen. Um diese und weitere Stoffe auf dem europäischen Markt anbieten zu können, sind unterschiedliche Normen und Regelungen zu beachten. Die Einhaltung von Normen und Regelungen wird mit Zertifizierungen bestätigt. Dieser Praxisleitfaden bietet einen ersten Einstieg für alle, die in das Feld der zertifizierten Mikroalgenproduktion eintreten wollen. Er gibt einen Überblick über geltende Regelungen verschiedener Ziel-Märkte und damit verbundene Maßnahmen der Qualitätssicherung. Weiterhin werden detaillierte Checklisten sowie Muster für Standardarbeitsanweisungen und Prozessbeschreibungen (via Prozess-Turtle) zur Verfügung gestellt.:Kurzinformationen Einführung Unternehmensorganisation und Qualitätsmanagement Absatzmärkte, Normen und Verordnungen für Herstellung und Vertrieb von Mikroalgen Ausgewählte Vorschriften für zertifizierte Mikroalgenproduktion Handlungsempfehlung für die Erfüllung von GMP-Anforderungen Typische Standardarbeitsanweisungen Planung von Audits Zusammenfassung

No description available.

Microalgae are considered as a potential feedstock for many promising applications. Some active substances in microalgae have plant biostimulation effects potentially useful in agriculture. However, to produce such a microalgal biomass, specific microalgae cultivation and post-treatment processes must be designed to preserve active substances. A particular focus is provided on cultivation (tubular photobioreactor) and different plausible post-treatment scenarios for microalgae separation (flocculation and centrifugation) and preservation (sterilization and drying). For each step, yield and energy consumption are modeled using data taken from literature or lab and pilot scale experiments. Industrial equipment for scale-up process is also studied by comparing existing systems. These models enable to make an economic evaluation of the whole process and to study its profitability for each scenario. The breakeven price is calculated as a function of the production rate. Several parameters are suggested to improve system efficiency and profitability at the end of this study. However, a better microalgae characterization and more experiments on potential post-treatment systems are required to improve the accuracy of the model.

Sécuriser un approvisionnement fiable de micro-algues est récemment devenu un enjeu industriel. Pour assurer la croissance de micro-algues dans des photobioréacteurs clos, un transfert de masse optimum de l'oxygène et du dioxyde de carbone doit être assuré. Dans cette thèse, une étude du transfert de masse gaz-liquide dans les conduites horizontales a été menée. Dans les trois premiers chapitres, un modèle unidimensionnel de transfert de masse dans le photobioréacteur a été développé. Tout d'abord, le transfert de masse entre une bulle de gaz allongée et un écoulement liquide turbulent a été 'étudié expérimentalement. En considérant l'interface comme étant plane, les coefficients de transfert de masse mesurés sont proches d'une corrélation de Lamourelle (1972). Le modèle de Taitel pour les écoulements stratifiés a été comparé à des modèles plus complets pour la prédiction de l'interface des bulles allongées. Une approche analytique basée sur un bilan de masse et utilisant les modèles testés a ensuite été développée et adaptée à un photobioréacteur afin de prédire l'efficacité de la conversion du CO2 en biomasse en fonction des paramètres d'exploitation. Les deux derniers chapitres visent à développer une simulation numérique du transfert de masse gaz-liquide. Une mesure de la concentration en CO2 dans le sillage d'une bulle de gaz ascendante a été effectuée à l'aide d'une méthode améliorée de traitement des données de Fluorescence Induite par Plan Laser (FIPL). Enfin, une simulation numérique a été réalisée sous COMSOL / Securing a reliable supply of microalgae has recently become an industrial stake. To ensure successful growing of microalgae in enclosed, tubular photobioreactor systems as in Microphyt, an optimum mass transfer of oxygen and carbon dioxide should be secured. In this thesis an investigation of the gas-liquid mass transfer in horizontal pipes has been conducted. In the first three chapters, a one dimensional mass transfer model in horizontal gas-liquid flows will be developed and applied to horizontal photobioreactors. Firstly, a study of mass transfer between an elongated gas bubble under a turbulent liquid flow immobilized in a duct has shown that under the hypothesis considering the interface as a flat plane estimated, the measured mass transfer coefficients appear to be well fitted by a correlation from Lamourelle (1972). The interface prediction for stratified flows have been compared to more complete unit-cell models for intermittent flow interface and found to be a good first estimate. The photobioreactor’s conversion efficiency of CO2 into biomass as a function of operating parameters is investigated using an analytical approach to complete the mass balance and classical two-phase flow approach from Taitel (1976). The last two chapters aim at developing a numerical simulation of gas-liquid mass transfer. A measurement of CO2 wake structures behind free rising bubbles have realized using an improved data treatment method for Planar laser-induced fluorescence (PLIF) and pH sensitivity of fluorescein sodium. Finally, an implementation of the experimental measurements under COMSOL has been realised

Arthrospira (Spirulina) platensis, cianobactéria fotoautotrófica, é importante comercialmente devido ao seu valor nutricional (elevado teor de proteínas, lipídios, vitaminas e minerais), além da presença do ácido graxo γ-linolênico e pigmentos, os quais agregam valor a esta biomassa. Este micro-organismo foi cultivado em fotobiorreator tubular, e os seguintes parâmetros foram avaliados: aplicação de diferentes sistemas de circulação de células, intensidades luminosas, adição simultânea de diferentes proporções de sulfato de amônio e nitrato de sódio como fontes de nitrogênio, adição de dióxido de carbono proveniente de fermentação alcoólica e, finalmente, a aplicação desta biomassa para remoção de metais pesados. Para a avaliação do sistema de circulação de células e da intensidade luminosa foram utilizadas as variáveis dependentes: concentração celular máxima (Xm), produtividade em células (Px), fator de conversão de nitrogênio em células (YX/N), eficiência fotossintética (EF) e teores de proteínas e lipídios da biomassa. Os protocolos de alimentação de nitrogênio e o uso de CO2 proveniente de cilindro ou de fermentação alcoólica foram avaliados com uso das mesmas variáveis já citadas, exceto EF; e, para remoção de Ni2+, Zn2+ e Pb2+ em solução aquosa foram estudados o tempo de adsorção e a influência da concentração inicial de metal em solução; com posterior cálculo da eficiência e capacidade de adsorção, e apresentação de modelos cinéticos e isotermas de equilíbrio. A melhor combinação de resultados em valores médios (Xm = 4055 mg L-1, Px = 406 mg L-1 d-1, YX/N = 5,07 mg mg-1, lipídios totais = 8,94%, proteínas totais = 30,3%, EF = 2,04%) foi obtida nos cultivos realizados com intensidade luminosa = 120 µ mol fótons m-2 s-1. O tipo de sistema de circulação de células utilizado não influenciou estatisticamente quase nenhuma das respostas estudadas e isto indica a possibilidade de substituição do sistema tradicional airlift pelos outros testados neste trabalho. A mistura da menor quantidade de nitrato em relação à de amônio (protocolo I) resultou na melhor combinação de resultados das variáveis dependentes analisadas(Xm = 4,543 g L-1; Px = 0,460 g L-1 d-1; YX/N = 15,6 g g-1; lipídios totais = 8,39% e proteínas totais = 18,7 %). O tipo da fonte de CO2 não exerceu nenhuma influência estatística no crescimento de A. platensis, o que indica a possibilidade do uso do resíduo (CO2) proveniente de fermentação alcoólica nos cultivos desta cianobactéria. A adsorção dos metais pela biomassa seca de A. platensis reduziu à medida que houve aumento da concentração inicial de metal, e a maior eficiência de adsorção foi para o metal Pb2+ (78%). Chlorella quando comparada à Arthrospira foi considerada melhor adsorvente, devido ao maior valor de eficiência de remoção (86,5%). O modelo, cinético de Ho e Mckay e os modelos de equilíbrio de Langmuir e Freundlich se adequaram bem ao processo de adsorção desta célula. / Arthrospira (Spirulina) platensis, photoautotrophic cyanobacterium, is commercially important due its high nutritional value (high content of proteins, lipids, vitamins, and minerals), besides of the presence of γ-linolenic fatty acids and pigments, which add value to this biomass. This microorganism was cultivated in tubular photobioreactor and the following parameters were evaluated: application of different cell circulation systems, light intensities, different protocols of simultaneous ammonium sulfate and sodium nitrate feeding, addition of CO2 from ethanol fermentation and, finally, the application of this biomass to the heavy metal removal. The dependent variables [maximum cell concentration (Xm), cell productivity (Px), nitrogen-to-cell conversion factor (YX/N), photosynthetic efficiency (EF) and proteins and lipids contents] were used to evaluate the influence of cell circulation systems and the light intensities. The nitrogen feeding protocols and the use of CO2 from cylinder or ethanol fermentation were evaluated using the same parameters aforementioned, except EF; and to analyze Ni2+, Zn2+ and Pb2+ removal, the adsorption time and the influence of initial metal concentration were evaluated with subsequent calculation of the efficiency and capacity adsorption. The kinetic models and equilibrium isotherms were also presented. The best combination of responses\' mean values (Xm = 4055 mg L-1, Px = 406 mg L-1 d-1, YX/N = 5.07 mg mg-1, total lipids = 8.94%, total proteins = 30.3%, EF = 2.04%) was obtained at light intensity = 120 µ mol photons m-2 s-1. The cell circulation system did not exert statistical significant influence on almost all the responses, which suggests that the traditional airlift system could successfully be substituted by the others tested in this work. The mixture of the lower amount of nitrate in relation to ammonium (protocol I) (Xm = 4,543 g L-1; Px = 0,460 g L-1 d-1; YX/N = 15,6 g g-1; total lipids = 8,39% and total proteins = 18,7 %) resulted in the best combination of responses\' mean values. The kind of CO2 source did not exerted any statistical influence on A. platensis growth, which suggests the possibility of using the ethanol fermentation residue (CO2) in the cyanobacterium cultivation. The metal adsorption by A. platensis biomass reduced with increasing initial metal concentration and the adsorption efficiency was higher for metal Pb2+ (78%). Chlorella was considered a better biosorbent when compared to Arthrospira because of the higher values of removal efficiency (86.5%). The kinetic (Ho and McKay) and equilibrium (Langmuir and Freundlich) models are well suited to the adsorption process of this cell.

Submitted by Marilene Donadel (marilene.donadel@unioeste.br) on 2019-03-11T22:27:03Z No. of bitstreams: 1 Camila_Hinterholz_2018.pdf: 6217665 bytes, checksum: a00620aa466c15bb7a0ac5126e8c3614 (MD5) / Made available in DSpace on 2019-03-11T22:27:03Z (GMT). No. of bitstreams: 1 Camila_Hinterholz_2018.pdf: 6217665 bytes, checksum: a00620aa466c15bb7a0ac5126e8c3614 (MD5) Previous issue date: 2018-11-13 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The objective of this study was to develop a laboratory-scale photobioreactor (PBR) capable of obtaining microalgae cultures with high cell density. For this, the cultivation of Poterioochromonas malhamensis was evaluated in a flat-plate PBR, built with a useful volume of 10 L. The microalga strain was isolated from an artificial lake from Toledo – PR. The whole study was based on the System Analysis Theory, in which the research was divided into four hierarchical levels. In the first level, the mass transfer (MT) parameters were evaluated by means of a fractionated factorial design (FFD) considering four factors: gas inlet flow (Qgas), CO2 in the inlet gas (φCO2), antifoam concentration (CAE) and salinity of the medium (φsal), whose adjustable values may represent the MT for the responses K_L a_(O_2 ),〖 K〗_L a_(CO_2 ),O_2^eq e CO_2^eq. Thus, the parameters K_L a_(O_2 ) (14.88 to 79.32 h-1) and K_L a_(CO_2 ) (0 to 125.40 h-1) were verified experimentally, as well as the equilibrium concentration variations: O_2^eq (37.33 to 99.66%) and CO_2^eq (0 to 98.33%) respectively, relative to the equilibrium values of the species dissolved in pure water. Still in the first level, the Euler-Euler model was used for the study of the fluid dynamics inside the PBR, with the software Comsol®, being evaluated three configurations: PBRA – without disperser; PBRB – with simple bubble disperser; and PBRC – with perforated disperser, whose bubble dispersion proposal ensured a higher frequency of the light-dark cycles, according to the simulations of water and bubbles flows performed in mono and multiphase systems, respectively. The influence of the bubble dispersion module on cell growth was verified experimentally with PBRA and PBRC runs, which showed a 175% increase in biomass production (15.7 g L-1) with use of PBRC. In the second hierarchical level, the tolerance of the strain to the adjustable conditions of the system was studied: Qgas, φCO2, in addition to temperature, lighting conditions, nutrient concentrations and organic carbon, using the pulse technique. Within the evaluated ranges, the results did not show inhibition of growth, but was accentuated in values above 30 °C. Also, at this level, assuming the production of 1 g L-1 biomass, M-8 and BG-11 media were optimized by linear mathematical programming, whose objective function subjected to constraints based elemental composition of the biomass. Therefore, the original and optimized media were evaluated regarding biomass production, as well as protein, carbohydrate, lipid and pigment chlorophyll-a, chlorophyll-b and carotenoids content of cells. ANOVA indicated the best medium for the production of biomass (M-8), whose optimized values were prepared in the nitrogen source evaluation cultures. In these, three inorganic sources – NH4NO3, NaNO3 and KNO3; and an organic – urea were compared, and it was verified by ANOVA that the most appropriate source for the production of biomass was urea. In the third hierarchical level, a growth model was proposed based on mass balances and the definitions of specific transformation velocity, conversion definition, CO2 and O2 dissociation reactions in water and Henry's Law. After applying the phenomenological modeling of unstructured and non-segregated models, a set of stoichiometric equations was built, which were used to evaluate kinetic models obtained from the literature, by which the influence of light, dissolved CO2 and O2, pH and temperature on cell growth. Using the results obtained at previous levels, the fourth-level study indicated that the simplified model proposed to describe the experimental data of Poterioochromonas malhamensis culture in closed PBR (cell growth, CO2 consumption in the gas phase, and variation in the pH of the medium) was adequate, whose simulations were also satisfactory for the consumption of the inorganic dissolved carbon species, as well as for the mass transfer between the phases. In this step, 32 kinetic and stoichiometric parameters were estimated in the fit of the proposed model to the experimental data, through an algorithm based on the Genetic Algorithms implemented in the software Maple®. Therefore, this study presented the use of the perforated bubble disperser module coupled to the proposed PBR, which allowed the achievement of what is characterized by ultra-high density culture, since ≈ 15 g L-1 of biomass were produced by the culture of P. malhamensis. It was also highlighted the possibility of optimizing important conditions for biomass cultivation, as well as for the design of PBRs, through the collection of experimental data, associated to appropriate statistical methodologies of evaluation and to kinetic and fluid dynamics modeling, according to the General Theory of Systems applied. / Este estudo teve por objetivo o desenvolvimento de um fotobiorreator (PBR) em escala laboratorial capaz de obter culturas de microalgas com alta densidade celular. Para isto, avaliou-se o cultivo de Poterioochromonas malhamensis em um PBR de placas planas, construído com volume útil de 10 L. A cepa desta microalga foi isolada de um lago artificial de Toledo – PR. Todo o estudo foi baseado na Teoria Geral de Sistemas em que se dividiu a pesquisa em quatro níveis hierárquicos. No primeiro nível avaliaram-se os parâmetros da transferência de massa, por meio de um planejamento fatorial fracionado (PFF) considerando-se quatro fatores: vazão de entrada de gás (Qgás), CO2 no gás de entrada (ϕCO2), concentração de antiespuma (CAE) e salinidade do meio (ϕsal), cujos valores ajustáveis podem representar a TM para as respostas K_L a_(O_2 ),〖 K〗_L a_(CO_2 ),O_2^eq e CO_2^eq. Com isso, foram verificados experimentalmente os parâmetros: K_L a_(O_2 ) (14,88 a 79,32 h-1) e K_L a_(CO_2 ) (0 a 125,40 h-1), bem como as variações de concentração das espécies dissolvidas no equilíbrio: O_2^eq (37,33 a 99,66%) e CO_2^eq (0 a 98,33%), respectivamente relativo aos valores de equilíbrio das espécies dissolvidas em água pura. Ainda no primeiro nível, o modelo de Euler-Euler foi utilizado para o estudo da fluidodinâmica dentro do PBR, por meio do software Comsol®, sendo avaliadas três configurações: PBRA – sem dispersor; PBRB – com dispersor de bolhas simples; e PBRC – com dispersor perfurado, cuja proposta de dispersão de bolhas garantiu maior frequência dos ciclos claro-escuro, de acordo com as simulações de fluxos de água e de bolhas realizadas em sistema mono e multifásico, respectivamente. A influência do módulo dispersor de bolhas sobre o crescimento celular foi verificada experimentalmente por ensaios feitos com o PBRA e o PBRC, a partir dos quais se verificou o aumento de 175% na produção de biomassa (15,7 g L-1) com a utilização do PBRC. No segundo nível hierárquico, foi estudada a tolerância da cepa frente às condições ajustáveis do sistema: Qgás, ϕCO2, CAE, além de temperatura, condições de iluminação, e concentrações de nutrientes e carbono orgânico, utilizando-se a técnica de pulsos. Dentro das faixas avaliadas, os resultados não mostraram inibição do crescimento, porém este foi acentuado em valores acima de 30°C. Também, neste nível, admitindo-se a produção de 1 g L-1 de biomassa, os meios M-8 e BG-11 foram otimizados por programação matemática linear cuja função objetivo sujeita a restrições baseadas na composição elementar da biomassa. Com isso, os meios originais e otimizados foram avaliados quanto à produção de biomassa, e ao teor de proteínas, carboidratos, lipídeos e pigmentos clorofila-a, clorofila-b e carotenoides. A ANOVA indicou o melhor meio para a produção de biomassa (M-8), cujos valores otimizados foram preparados nos cultivos de avaliação da fonte de nitrogênio. Nestes, três fontes inorgânicas – NH4NO3, NaNO3 e KNO3; e uma orgânica – ureia foram comparadas, verificando-se pela ANOVA que a fonte mais indicada para a produção de biomassa foi a ureia. No terceiro nível hierárquico, foi proposto um modelo de crescimento com base em balanços de massa e nas definições de velocidade específica de transformação, definição de conversão, reações de dissociação do CO2 e O2 em água e a Lei de Henry. Após a aplicação da modelagem fenomenológica de modelos não estruturados e não segregados, obteve-se um conjunto de equações estequiométricas, que foram utilizadas para avaliar modelos cinéticos obtidos da literatura, pelos quais se avaliaram a influência da luz, CO2 e O2 dissolvidos, pH e temperatura sobre o crescimento celular. Utilizando-se os resultados alcançados nos níveis anteriores, o estudo do quarto nível indicou que o modelo simplificado proposto para descrever os dados experimentais de cultivo de Poterioochromonas malhamensis em PBR fechado (crescimento celular, consumo de CO2 na fase gasosa, e variação no pH do meio) foi adequado, cujas simulações mostraram-se satisfatórias também para o consumo das espécies de carbono inorgânico dissolvido, bem como para a transferência de massa entre as fases. Nesta etapa, 32 parâmetros cinéticos e estequiométricos foram estimados no ajuste do modelo proposto aos dados experimentais, por meio de um algoritmo baseado nos Algoritmos Genéticos implementado no software Maple®. Portanto, o presente estudo apresentou a utilização do módulo dissipador de bolhas perfurado acoplado ao PBR proposto, o que permitiu a obtenção do que se caracteriza por cultura de densidade ultra alta, uma vez que ≈15 g L-1 de biomassa foram fornecidos pelo cultivo de P. malhamensis. Destacou-se, também, a possibilidade da otimização de condições importantes para o cultivo de biomassa, bem como para o projeto de PBRs, por meio do levantamento de dados experimentais, associado às adequadas metodologias estatísticas de avaliação e à modelagem cinética e fluidodinâmica, conforme se verificou pela Teoria Geral de Sistemas aplicada.

Cette thèse s’attache à la maximisation de la consommation du dioxyde de carbone par les microalgues. En effet, suite aux différentes problématiques environnementales actuelles liées principalement aux émissions importantes de gaz à effet de serre et notamment le CO2, il a été démontré que les microalgues jouent un rôle très prometteur pour la bio-fixation du CO2. Dans cette optique, nous nous intéressons à la mise en place d’une loi de commande robuste permettant de garantir des conditions opératoires optimales pour une culture de la microalgue Chlorella vulgaris dans un photobioréacteur instrumenté. Cette thèse repose sur trois axes principaux. Le premier porte sur la modélisation de la croissance de l’espèce algale choisie à partir d’un modèle mathématique traduisant l’influence de la lumière et de la concentration en carbone inorganique total. En vue de la commande, le deuxième axe est consacré à l’estimation de la concentration cellulaire à partir des mesures disponibles en temps réel du dioxyde de carbone dissous. Trois types d’observateurs ont été étudiés et comparés : filtre de Kalman étendu, observateur asymptotique et observateur par intervalles. Le dernier axe concerne l’implantation d’une loi de commande prédictive non-linéaire couplée à une stratégie d’estimation pour la régulation de la concentration cellulaire autour d’une valeur maximisant la consommation du CO2. Les performances et la robustesse de cette commande ont été validées en simulation et expérimentalement sur un photobioréacteur instrumenté à l’échelle de laboratoire. Cette thèse est une étude préliminaire pour la mise en œuvre de la maximisation de la fixation du dioxyde de carbone par les microalgues. / This thesis deals with the optimization of carbon dioxide consumption by microalgae. Indeed, following several current environmental issues primarily related to large emissions of CO2, it is shown that microalgae represent a very promising solution for CO2 mitigation. From this perspective, we are interested in the optimization strategy of CO2 consumption through the development of a robust control law. The main aim is to ensure optimal operating conditions for a Chlorella vulgaris culture in an instrumented photobioreactor. The thesis is based on three major axes. The first one concerns growth modeling of the selected species based on a mathematical model reflecting the influence of light and total inorganic carbon concentration. For the control context, the second axis is related to biomass estimation from the real-time measurement of dissolved carbon dioxide. This step is necessary for the control part due to the lack of affordable real-time sensors for this kind of measurement. Three observers structures have been studied and compared: an extended Kalman filter, an asymptotic observer and an interval observer. The last axis deals with the implementation of a non-linear predictive control law coupled to the estimation strategy for the regulation of the cellular concentration around a value which maximizes the CO2 consumption. Performance and robustness of this control law have been validated in simulation and experimentally on a laboratory-scale instrumented photobioreactor. This thesis represents a preliminary study for the optimization of CO2 mitigation strategy by microalgae.