Les travaux présentés dans cette thèse s'inscrivent dans le cadre de la chaire OXYTEC qui regroupe l'entreprise Air Liquide, CentraleSupélec et le CNRS. Ce travail bénéficie également de l'appui de l'ANR. Ce manuscrit fait état des premiers développements et résultats sur le plan expérimental. Un dispositif complet a été développé au laboratoire EM2C pour étudier l'oxy-combustion de flammes stabilisées sur un injecteur tourbillonneur jusqu'à des pressions de 30 bar. Les résultats présentés dans cette thèse ont traits à des prémélanges dont le combustible est du méthane lorsque le foyer opère à pression atmosphérique. Les techniques de LIF-OH, la PIV, la LDV complétées par des mesures de chimiluminescence et de température sont utilisées (1) pour accumuler des données sur la structure de trois flammes de référence CH4/air, CH4/O2/N2 et CH4/O2/CO2 qui sont utilisées pour valider des outils de simulation de la combustion et des transferts thermiques, (2) élucider les mécanismes de stabilisation des oxy-flammes swirlés pré-mélangées à partir d'études paramétriques sur le nombre de swirl, la vitesse débitante, la vitesse laminaire de flamme et l'angle de l'ouvreau de l'injecteur, et (3) comparer la structure des oxy-flammes diluées au CO2 avec des flammes CH4/air en examinant notamment la position de pied de flamme, la topologie générale de la flamme et les températures des parois du foyer. / The work presented in this thesis falls within the framework of the OXYTEC chair, a partnership between Air Liquide, CentraleSupélec and the CNRS. This work also benefits from the support of the ANR. This manuscript reports the first developments and results on the experimental level. A test rig has been developed to study oxy-combustion of swirl-stabilized flames up to pressures of 30 bar. The results presented deal with premixed conditions where the fuel is methaneand the setup is operated at atmospheric pressure. Laser induced fluorescence on the hydroxyl radical, particle imaging velocimetry, Doppler laser velocimetry completed by chemiluminescence imaging and temperature measurements are used to (1) gather data on the structure of three reference flames CH4/air, CH4/O2/N2 and CH4/O2/CO2 which are used to validate simulations of the reacting flow and heat transfer and (2) elucidate the stabilization mechanisms of premixed swirling oxy-flames from parametric studies by varying the swirl number, the bulk injection velocity, the laminar burning velocity and the injector cup angle and (3) compare the structure of CO2 diluted oxy-flames with CH4/air flames by examining the position of the flame leading edge, the general topology of the flame and the temperatures of flow and the combustor walls.

L’objectif de la thèse est de concevoir un design le plus optimisé possible d’une centrale à charbon pulvérisée fonctionnant en oxy-combustion. Une telle centrale intègre un système de production d’oxygène (ASU), une chaudière, un cycle électrogène, des équipements de dépollution ainsi qu’un système de purification et de compression du CO2 (CPU). Ainsi, dans un premier temps, la thèse portera sur la compréhension, l’analyse et la modélisation des différents procédés qui composent la chaîne d’une centrale fonctionnant en oxy-combustion. Ensuite, les performances et la configuration de la centrale seront optimisées de façon à réduire la destruction d’exergie tout s’assurant de la compétitivité économique de la solution ainsi obtenue. A l’issue de cette thèse, les origines des pertes exergétiques du système étudié et les schémas d’intégration permettant de maximiser les gains énergétique à l’échelle de la centrale seront identifiées. De plus, les stratégies de dépollution des fumées les plus adaptées seront définies et les nouveaux procédés intégrés seront évalués à la fois d’un point de vue technico-économiques et flexibilité. / The objective of the thesis is the conception of an optimized oxy-fired pulverized-coal power plant. Such a power plant is constituted of an oxygen production system (ASU), a boiler, power cycle, depollution equipments and a CO2 purification and compression system (CPU). After a first step consists in understanding, analyzing and modeling the different processes composing the oxy-combustion system; the work will focus on the optimization of the performances and the configuration of the power plant by minimizing exergy destructions while ensuring economic competitiveness of the obtained solution. At the end of the thesis, the origins of the exergetic losses in the system as well as the thermal integration scheme allowing the maximization of the energetic gains at power plant level will be identified. Additionally, the most adapted flue gas depollution strategies will be defined and the new integrated process schemes will be evaluated on both a techno-economic and flexibility basis.

Submitted by NATALIA DA SILVA RIBEIRO null (nsribeiro2@gmail.com) on 2017-03-21T14:14:32Z No. of bitstreams: 1 Defesa Natalia_Final..pdf: 3776400 bytes, checksum: 4147b87ccc01a8c95a9e60c44f70613b (MD5) / Rejected by Luiz Galeffi (luizgaleffi@gmail.com), reason: Solicitamos que realize uma nova submissão seguindo a orientação abaixo: O arquivo submetido está sem a ficha catalográfica. A versão submetida por você é considerada a versão final da dissertação/tese, portanto não poderá ocorrer qualquer alteração em seu conteúdo após a aprovação. Corrija esta informação e realize uma nova submissão com o arquivo correto. Agradecemos a compreensão. on 2017-03-22T14:39:39Z (GMT) / Submitted by NATALIA DA SILVA RIBEIRO null (nsribeiro2@gmail.com) on 2017-03-22T17:58:52Z No. of bitstreams: 1 Defesa Natalia_Final..pdf: 3778151 bytes, checksum: 08fb9f490ae966bed7312cca924c9c9e (MD5) / Approved for entry into archive by Luiz Galeffi (luizgaleffi@gmail.com) on 2017-03-23T16:23:04Z (GMT) No. of bitstreams: 1 ribeiro_ns_me_guara.pdf: 3778151 bytes, checksum: 08fb9f490ae966bed7312cca924c9c9e (MD5) / Made available in DSpace on 2017-03-23T16:23:04Z (GMT). No. of bitstreams: 1 ribeiro_ns_me_guara.pdf: 3778151 bytes, checksum: 08fb9f490ae966bed7312cca924c9c9e (MD5) Previous issue date: 2017-03-03 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Nesta dissertação, investiga-se através da análise termogravimétrica o comportamento da combustão de amostras de carvão mineral, bagaço de cana-de-açúcar, bagaço de sorgo biomassa e das misturas de carvão-biomassa. A biomassa e o carvão possuem propriedades físico-químicas diferentes que proporcionam comportamento térmico diferente durante o processo de co-combustão, desta forma o objetivo desta pesquisa é caracterizar o comportamento térmico de misturas de carvão mineral com bagaço de cana-de-açúcar e bagaço de sorgo em atmosferas simuladas de combustão (O2/N2) e oxicombustão (O2/CO2). Os experimentos foram realizados em duplicata em um analisador termogravimétrico utilizando uma razão de aquecimento de 10 °C/min. Foi considerada uma granulometria uniforme para todos os materiais (63 µm) com a finalidade de garantir uma mistura homogênea. Foram estudadas quatro proporções de biomassa na mistura (10, 25, 50 e 75%). A partir das técnicas de termogravimetria (TG) e termogravimetria derivada (DTG) foram determinados parâmetros tais como Índice de combustão, sinergismo e energia de ativação, bem como avaliada a influência da atmosfera de combustão sobre esses parâmetros. Os resultados indicam que o bagaço de cana-de-açúcar apresenta valor de energia de ativação inferior ao registrado para o bagaço de sorgo e desempenho de combustão superior ao do bagaço de sorgo. Para as misturas, os melhores resultados foram registrados até a proporção de 25% de biomassa na mistura. Avaliando individualmente cada material, quando se substitui o N2 por CO2 pode-se observar um aumento na reatividade da reação, uma maior oxidação dos materiais e uma melhora nos parâmetros avaliados. Para ambas as misturas não foram observadas mudanças significativas no perfil de combustão quando o N2 é substituído por CO2. No entanto, a presença da biomassa na co-combustão com o carvão, além dos benefícios econômicos e ambientais, aumentou o desempenho da combustão do carvão mineral em ambas as atmosferas. / This dissertation investigates by thermogravimetric analysis the behavior of the combustion of coal, sugarcane bagasse, sorghum biomass bagasse and coal-biomass blends. The biomass and coal have different physicochemical properties that provide different thermal behavior during the process of co-combustion, thus the aim of this research is to characterize the thermal behavior of coal mixed with sugarcane bagasse and sorghum bagasse in simulated atmospheres of combustion (O2/N2) and oxycombustion (O2/CO2). The experiments were performed in duplicate in a thermogravimetric analyzer using a heating rate of 10 ° C/min. A uniform particle size for all materials (63 μm) in order to ensure a homogeneous mixture was considered. Four biomass ratios were studied in the blend (10, 25, 50 and 75%). From the techniques of Thermogravimetry (TG) and Derivative Thermogravimetry (DTG) curves were determined parameters such as: Combustion index, synergism and activation energy and evaluated the influence of combustion atmosphere on these parameters. The results indicate that the sugarcane bagasse presents a lower activation energy value than sorghum bagasse and combustion performance higher than sorghum bagasse. For mixtures, best results were recorded up to 25% proportion of biomass in the blend. Individually evaluating each material, when replacing N2 by CO2 can be seen an increase in the reactivity of the reaction, the increased oxidation of the materials and an improvement in the evaluated parameters. For both blends, no significant changes in combustion profile when N2 substituted by CO2. However, the presence of biomass in co-combustion with coal in addition to economic and environmental benefits increased the combustion performance of coal in both atmospheres. / CNPq: 134366/2015-8

Oxy-combustion is a developing technology that enables carbon dioxide (CO2) capture. Flame temperature and emissivity data were taken on a 150 kWth, pulverized-coal, burner flow reactor (BFR) that has been modified to run oxy-combustion with pure CO2 as simulated recycled flue gas. Data were taken at 78 conditions in which three parameters were varied, namely: the swirl angle of the fuel stream, the location of the oxidizer as it exited the burner, and the flow rate of diluent (pure CO2) added to the outer, secondary stream. At each condition, digital color images were obtained using a calibrated RGB camera. The images were used to determine lift-off length, temperature, and emissivity. The mathematical theory of two-color pyrometry and the calibration process used to measure the camera sensitivity is presented. The two most commonly used emissivity models in two-color pyrometry, the Hottel and Broughton and gray models, were investigated to determine which was the most appropriate for use in an oxy-coal flame. A significant difference of 7% in the temperature and 24% in the emissivity results were found when processing an image with the Hottel and Broughton and gray emissivity models. The Hottel and Broughton model was selected for processing, because the Hottel and Broughton model is more appropriate for soot which appeared to dominate flame emissions. Using the two-color data, several trends were documented. Flame temperature was seen to decrease with increasing CO2 flow rate. Within a given flame along the axial direction, temperature was seen to correlate with emissivity. As emissivity increased, flame temperature was seen to decrease. Many flames were lifted from the burner exit. Lift-off length was decreased and the flames became more attached by: 1) Increasing the amount of swirl given to the fuel stream, 2) Adding O2 to the center primary tube or 3) Decreasing the flow of secondary CO2. At higher center oxygen flow rates (above 8.5 kg/hr), the O2 jet velocity was large causing increased entrainment and mixing which degraded burner performance.

This work produced the retrofit of an air-fired, 150 kW reactor for oxy-combustion which was then used in three oxy-combustion studies: strategic oxy-combustion design, oxy-combustion of petroleum coke, and air versus oxy-combustion radiative heat flux measurements. The oxy-combustion retrofit was accomplished using a system of mass flow controllers and automated pressure switches which allowed safe and convenient operation. The system was used successfully in the three studies reported here and was also used in an unrelated study. A study was completed where a novel high oxygen participation burner was investigated for performance while burning coal related to flame stability, NO, and burnout using a burner supplied by Air Liquide. Parameters investigated included oxygen (O2) injection location, burner swirl number and secondary carbon dioxide (CO2) flow rate. The data showed swirl can be used to stabilize the flame while reducing NO and improving burnout. Center O2 injection helped to stabilize the flame but increased NO formation and decreased burnout by reducing particle residence time. Additional CO2 flow lifted the flame and increased NO but was beneficial for burnout. High O2 concentrations up to 100% in the secondary were accomplished without damage to the burner. Petroleum coke was successfully burned using the Air Liquide burner. Swirl of the secondary air and O2 injection into the center tube of the burner were needed to stabilize the flame. Trends in the data similar to those reported for the coal study are apparent. Axial total radiant intensity profiles were obtained for air combustion and three oxy-combustion operating conditions that used hot recycled flue gas in the secondary stream. The oxygen concentration of the oxidizer stream was increased from 25 to 35% O2 by decreasing the flow rate of recycled flue gas. The decrease in secondary flow rate decreased the secondary velocity, overall swirl, and mixing which elongated the flame. Changing from air to neat CO2 as the coal carrier gas also decreased premixing which elongated the flame. Flame elongation caused increased total heat transfer from the flame. The air flame was short and had a higher intensity near the burner, while high O2 concentration conditions produced lower intensities near the burner but higher intensities and temperatures farther downstream. It was shown that oxycombustion can change flame shape, temperature and soot concentration all influencing heat transfer. Differences in gas emission appear negligible in comparison to changes in particle emission.

The goal of this thesis was to understand what parameters would be most impactful when delivering dry, pulverized coal in a dilute-phase, with a high-pressure feed-system to a pressurized oxy-combustion (POC) reactor. Many studies have conveyed materials in dense-phase plugs at high-pressure or in dilute-phase flows at atmospheric pressure. Very few studies have fluidized and conveyed materials in dilute-phase flows at high pressure, as we needed to. Additionally, studies which might have been applicable based upon system -pressure and -phase delivered findings that were empirically based and therefore not specifically applicable to non-similar systems. 220 different tests were ran using a bench-scale apparatus consisting of a hopper, connecting conveying pipes, and a filter point (representing the future reactor). The system was pressurized to 300 psi using CO2. Dry, pulverized coal with an average diameter of 50 microns and a bulk density of 800.9 kg/m3 was fluidized and conveyed with different combinations of fluidization inlet and fluidization outlet flowrates. Each specific flowrate combination was tested 3 to 5 times. The resulting coal flowrates were recorded and analyzed to see which flowrate combination delivered 13.6 kgs coal/hr and had the least variability between tests. The fluidization inlet and outlet flowrates, coal moisture content, and system geometry were key parameters. In a 2-inch diameter hopper the fluidization inlet flowrate should be kept at 0.119 m/s or below to keep the fluidization regime within the hopper below the transition point to the bubbling fluidization regime. This was beneficial since less CO2 was needed by the system and smaller perturbations within the bed didn't disrupt flow leaving the hopper. The fluidization outlet flowrate could still advance the fluidization regime within the hopper even if the fluidization inlet flowrate is kept at 0.119 m/s. For a ¼ inch diameter the outlet should be kept at 0.005 m/s or above. Additionally, the standard deviation in the measured coal flowrate decreased dramatically when flow of gas was allowed to exit through the top of the coal column (fluidization outlet). The standard deviation was 8.2 kg/hr with the fluidization outlet closed and 3.5 kg/hr with the fluidization outlet flowing to provide 0.005 m/s in the bed above the coal outlet. Coal should have a moisture content between 3% and 6% to ensure that electrostatic interactions between coal particles is kept to a minimum. Finally, these results were found for specific hopper and fluidization inlet and outlet diameters. If these diameters are changed then some calculation must be done for these results to be applicable to systems that are not like the one described later in this thesis.

Soot in coal combustion simulations is often ignored due to its computational complexity, despite significant effects on flame temperature and radiation. In this research, a 40 kW oxy-coal combustion system is modeled using Large Eddy Simulations (LES) and a semi-empirical monodisperse coal soot model. Simulation results are compared to experimental measurements of temperature, species concentrations, and soot concentration. Cases where soot is modeled are compared with cases where soot is neglected to determine the accuracy benefits of modeling soot. The simulations were able to replicate experimental results within an acceptable level of error. Including soot in the simulations did not consistently increase accuracy for the simulation setup and modeling assumptions used in this research.

Coal ash deposits were collected in a 160 kWth, down-fired oxy-coal reactor under staged and unstaged conditions for four different coals (PRB, Gatling, Illinois #6, and Mahoning). Concentration measurements of carbon, oxygen, sodium, magnesium, aluminum, silicon, phosphorus, sulfur, chlorine, potassium, calcium, titanium, chromium, manganese, iron, nickel, strontium, and barium were gathered from each deposit sample using scanning electron microscopy and energy dispersive spectroscopy (SEM-EDS). Backscattered electron micrographs for each deposit sample were analyzed to gather morphological data. Particle size and shape were studied for each deposit collected. The average particle sizes of the particles in upstream deposits were much larger than the average particle sizes of the downstream deposits. The downstream deposits consisted primarily of spherical particles while the upstream deposits consisted of round, irregular polygonal, and porous particles. Deposit particles are believed to have deposited at all stages of burnout; those depositing early during pyrolysis may have continued to react after deposition. Element maps for the aforementioned elements were collected with SEM-EDS and analyzed to quantify both average composition and composition of individual particles. These values were compared to ASTM ash analyses performed for each coal and ash collected from the flue gas stream with a cyclonic particle separator. It was found that sulfur concentrations of deposits do not correlate with corresponding sulfur concentrations of the coal. Comparison of similar experiments performed with air-combustion show that oxy-combustion deposits contain about twice as much sulfur as air-combustion deposits when burning the same coal. Deposition propensity of each coal was also examined, and the PRB and Gatling coals were found to have a moderately high deposition propensity whereas the deposition propensity of the Mahoning and Illinois #6 coals was fairly low.

Despite comprehensive studies addressing the chemical kinetics of butanol isomers, relevant uncertainties associated with the emissions of relevant pollutants persists. Also, a lack of chemistry knowledge of processes designed to produce biofuels limits their implementation at industrial scales. Therefore, the first objective of this thesis was to use cutting-edge kinetic theories to calculate rate constants of propen-2-ol, 1-pronenol, and vinyl alcohol keto-enol tautomerizations, which account for the production of the harmful carbonyl species. The second objective was to use the predictive capabilities of dynamic theories to reveal new chemistry of syngas oxy-combustion in supercritical CO2 and complexities of the zeolite dealumination, two processes involved in coal and biomass conversion. Rate constants computations considered transition state theory with variational effects, tunneling correction, and multistructural torsional anharmonicity. The study also included pressure effects by using and improving the system-specific quantum Rice-Ramsperger-Kassel/modified strong collision model. The atomistic simulations used ReaxFF force fields in hydrogen/oxygen/carbon monoxide/ carbon dioxide mixtures to represent the syngas system and an MFI zeolite with different water loading to model the dealumination. The results show that the studied assisted tautomerizations have much lower energy barriers than the unimolecular process. However, the “catalytic” effect is efficient only if the partner molecule is at high concentrations. Pressure effects are pronounced in the chemically activated tautomerizations, and the improved algorithm to compute pressure-dependent rate constants overcomes the initial difficulties associated with its application to C3 or larger molecules at temperatures above 800-1000 K. Reactive molecular dynamics simulations revealed the role of CO2 as an initiator in the syngas oxy-combustion and a new step involving the formation of formic acid. Those simulations for the zeolite dealumination process also showed that proton transfer, framework flexibility, and aluminum dislodging mediated by silicon reactions are complex dynamic phenomena determining the process. These aspects complement the dealumination theory uncovered so far and establish new paths in the study of water-zeolite interactions. Overall, the rate constants computed in this work reduce relevant uncertainties in the chemical kinetic mechanisms of alcohol oxidation, and the molecular dynamics simulations broaden the chemical knowledge of processes aimed at the utilization of alternative energy resources.

La présente étude a pour objectif d'évaluer et de comparer entre eux divers procédés de captage du CO2 appliqués aux centrales thermiques alimentées en gaz naturel (NGCC) et au charbon pulvérisé (CP). Ces procédés consistent en un captage du CO2 des fumées en post-combustion par des solvants chimiques, une décarbonisation du gaz naturel avec captage du CO2 en pré-combustion par un solvant physique et l'oxy-combustion du combustible avec séparation frigorifique du CO2. Ces procédés ont été évalués à l'aide du logiciel de procédés Aspen PlusTM pour aider à choisir la meilleure option pour chaque type de centrale. Pour la post-combustion, une solution aqueuse basée sur un mélange d'amines (N-méthyldiéthanolamine (MDEA) et triéthylène tétramine (TETA)) a été évaluée. Des mesures d'absorption ont été réalisées entre 298 et 333 K dans un réacteur fermé type cellule de Lewis. La pression partielle du CO2 à l'équilibre, caractéristique de la solubilité du CO2 dans le solvant, a été déterminée jusqu'à 393 K. Les performances sont comparées vis à vis de solvants plus conventionnels tels que la MDEA et la monoéthanolamine (MEA). Pour l'oxy-combustion, un procédé de captage, basé sur une séparation des composants des fumées à faible température, a été développé et appliqué aux centrales NGCC et CP. L'étude a montré que la pureté du flux d'O2 avait une influence non négligeable sur la concentration en CO2 dans les fumées et donc sur les performances du procédé de séparation. La dernière option étudiée est le reformage du gaz naturel qui permet un captage du CO2 en amont du système de production de l'électricité. Plusieurs configurations ont été évaluées : reformage à l'air ou à l'oxygène, pression de reformage et dilution du gaz de synthèse. La comparaison de ces différents concepts suggère que, à court et moyen terme, l'absorption chimique soit le procédé le plus intéressant pour la centrale NGCC. Pour la centrale CP, l'oxy-combustion peut être une option très intéressante, au même titre que le captage en post-combustion par absorption chimique.