The nitritation-anammox process is an efficient and cost-effective approach for biological nitrogen removal, but its application in treating mainstream wastewater remains a great challenge. The key objectives of this dissertation are to develop nitritation-anammox process to treat wastewater with low-nitrogen strength, understand the fundamental microbiology, and optimize its operation through experimental studies and mathematic modeling. Chapter 2 showed that the nitritation-anammox process has been successfully developed in an upflow membrane-aerated biofilm reactor, where pure oxygen was delivered via gas-permeable membrane module. Chapter 3 demonstrated that hybrid anaerobic reactor (HAR) could be an effective pretreatment method to provide a relatively low COD/N ratio for nitritation-anammox reactor. In Chapter 4, a novel mathematical model has been proposed to evaluate the minimum DO requirement for the nitritation-anammox reactor to achieve the maximum TN removal under various COD/N scenarios (controlled by HRTHAR). Chapters 5 and 6 designed an OsAMX system by linking nitritation-anammox to forward osmosis to remove the reverse-fluxed ammonium while using ammonium bicarbonate as a draw solute. The microbial community structures and dynamics, spatial distributions in these bioreactors were characterized by high-throughput sequencing and fluorescent in situ hybridization techniques. The studies in this dissertation have demonstrated that nitritation-anammox process is a promising alternative for sustainable mainstream treatment with the appropriate pretreatment approach and operation optimization. / PHD

Les bioréacteurs à membranes (BRM), combinant l’épuration biologique des eaux usées à une étape de séparation membranaire, ont montré de très bonnes performances en matière d’élimination de la pollution organique et azotée. Cependant, cette technologie présente des surcoûts de fonctionnement liés essentiellement à l’énergie dépensée pour fournir l’aération nécessaire à l’activité biologique mais utilisée aussi pour lutter contre les phénomènes de colmatage membranaire. Afin de réduire ces besoins énergétiques, un BRM dit autotrophe a été mis en place. Il est caractérisé par une alimentation à une charge organique faible, dont plus de 60% est éliminé par un traitement physico-chimique préalable. De ce fait, le BRM autotrophe assure principalement le traitement d’azote, qui peut être affecté par la limitation en carbone organique nécessaire à la réaction de dénitrification. Ce travail a pour objectif d’étudier le comportement d’un tel système et d’identifier les paramètres déterminants. L’étude a été conduite expérimentalement et par modélisation. Deux campagnes expérimentales ont été réalisées en régime permanent à des charges organiques et des âges de boues différents, suivies d’une expérimentation en régime transitoire provoqué par une montée de la charge en azote ammoniacal. Les résultats des campagnes ont montré qu’il est possible de baisser le rapport de DCO/N jusqu’à 4,5 et l’âge des boues à 40j sans pour autant perturber la dénitrification. Ces résultats ont été appuyés par une modélisation dynamique s’appuyant sur un modèle ASM3s-SMP qui a été développé ensuite calibré sur l’ensemble des résultats expérimentaux. L’analyse du phénomène du colmatage de la membrane au cours de l’expérimentation a montré une tendance au colmatage relativement faible par rapport aux BRM opérant dans des conditions usuelles (DCO/N>10). L’effet du nettoyage mécanique induit par l’ajout d’un matériau granulaire dans le module membranaire s’est avéré peu important dans les conditions étudiées. Une modélisation en régime permanent a permis de développer les expressions des grandeurs caractéristiques du fonctionnement du BRM. Après avoir été validées par des simulations conduites en utilisant le logiciel GPS-X®Hydromantis, ces équations ont permis, au moyen d’approches de modélisation différentes (modèle simple, ASM1 et ASM3), de repérer les paramètres opératoires et cinétiques les plus influents et de déterminer par des expressions analytiques le rapport DCO/N minimal requis pour une réaction de dénitrification complète. / Membrane Bioreactors (MBR) as a combination of biological wastewater treatment and a membrane separation step, showed high performances for organic and nitrogen compounds removal. However, this technology has high running costs linked to energy consumption for aeration. The latter has to be provided for the biological activity and for the membrane scouring that reduces membrane fouling phenomena. In order to decrease these expenses, an MBR called autotrophic was set. It is marked by low organic loading rates due to a physicochemical treatment removing more than 60% of the initial organic matter amount. Mainly, the autotrophic MBR is dedicated to nitrogen removal that can be influenced by the shortage of the organic substrate, needed to achieve the denitrification process. The aim of this work is the investigation of the behavior of such system and the identification of most determining parameters. Experimental and modeling studies were carried out. Two steady state experimental campaigns were performed at different organic loading rates and solid retention times. They were followed by an experiment at transient state induced by the nitrogen loading rate rise. Obtained results proved that operation at a COD/N ratio as low as 4.5 did not hamper the denitrification efficiency. Those results were reinforced by a dynamic modeling study based on a model called ASM3s-SMP that was developed and calibrated using the experimental results. Membrane fouling analysis during experimental campaigns showed low fouling propensities compared to MBR operating at more common conditions (COD/N>10). The mechanical cleaning effect due to the addition of a granular material to the membrane module was found insignificant in the case of these operating conditions.A steady state modeling study helped determining mathematical expressions of different variables, yields and rates describing the system operation. After a validation step based on simulations with the use of the GPS-X®Hydromantis software, these expressions associated to different modeling approaches (simple model, ASM1 and ASM3) allowed pointing out the critical operating and kinetic criteria in addition to the minimum COD/N ratio required for a complete denitrification reaction.

Este trabalho buscou avaliar a influência da relação C/N e da fonte de carbono no processo de nitrificação e desnitrificação simultâneas em reator de fluxo contínuo e leito fixo estruturado. Foi utilizado um reator vertical de acrílico, com volume total de 11,0 L e volume útil de 5,5 L, com hastes cilíndricas verticais de espuma de poliuretano como suporte para a biomassa. O sistema foi operado com TDH de 11,2±0,6 horas, aeração intermitente (2 horas com aeração e 1 hora sem aeração) e razão de recirculação de efluente igual a 5. A carga carbonácea afluente foi mantida constante (1,07 kg DQO.m-3.dia-1), ao longo de todo o período experimental, sendo as relações C/N testadas (9,7±1; 7,6±1; 2,9±1 e 2,9±0,4) obtidas a partir da variação na carga nitrogenada aplicada. Duas fontes orgânicas foram avaliadas: sacarose e peptona de carne. A eficiência média de remoção de DQO manteve-se acima de 90%. A eficiência máxima de remoção de N-total (84,6±10,1%) foi obtida para relação C/N de 2,9±1, com concentrações efluentes médias de NTK e N-NH4 de 5,9 e 4,3 mg.L-1, respectivamente. A análise estatística da eficiência de remoção de N-total confirmou que a fonte de carbono não exerceu influência sobre os processos de remoção. A obtenção de eficiências de desnitrificação superiores às calculadas estequiometricamente, em função da fonte de carbono, indicou a ocorrência de possíveis vias complementares para remoção de nitrogênio, como o processo anammox. As velocidades de nitrificação e desnitrificação obtidas nos ensaios cinéticos foram similares para as duas fontes de carbono e para as relações C/N estudadas e da mesma ordem de grandeza das apresentadas na literatura, reforçando a ideia de que a configuração de reator utilizada, aliada às adequadas condições operacionais, permitiu a remoção concomitante de matéria carbonácea e nitrogenada. / This study aimed to evaluate the influence of COD/N ratio and organic source on the simultaneous nitrification and denitrification (SND) in an up-flow structured-bed reactor. The reactor was made of acrylic with a total volume of 11 L and a working volume of 5.5 L and filled with cylinders of polyurethane foam as support for biomass growth. The system was continuously operated with an HRT of 11.2±0.6 h, intermittent aeration (2h with aeration and 1h without aerationd) and a liquid recycle ratio equal to 5. The organic load rate was constant (1.07 kg COD.m-3d-1) during the entire experiment. The COD/N ratios (9.7±1; 7.6±1; 2.9±1 and 2.9±0,4) were obtained from the variation of nitrogen load rate. Two organic sources were evaluated: sucrose and meat peptone. The average COD removal efficiency was above 90%. The maximum total nitrogen removal efficiency was 84.6±10.1%, with average TKN and NH4+-N effluent concentrations of 5.9 and 4.3 mg.L-1, respectively. Statistical analysis of total nitrogen removal efficiency confirmed that the carbon source did not influence over the removal processes. The denitrification efficiencies higher than the stoichiometrically calculated in function of the carbon source, indicated the occurrence of possible paths for nitrogen removal of nitrogen as anammox process. The nitrification and denitrification rates obtained in kinetic experiments were similar for the two sources of carbon and all C/N ratio studied at the same order of magnitude in relation to those reported in the literature, enhancing that the reactor configuration tested combined with the proper operational conditions allowed the organic matter and nitrogen removal simultaneously.

Este trabalho buscou avaliar a influência da relação C/N e da fonte de carbono no processo de nitrificação e desnitrificação simultâneas em reator de fluxo contínuo e leito fixo estruturado. Foi utilizado um reator vertical de acrílico, com volume total de 11,0 L e volume útil de 5,5 L, com hastes cilíndricas verticais de espuma de poliuretano como suporte para a biomassa. O sistema foi operado com TDH de 11,2±0,6 horas, aeração intermitente (2 horas com aeração e 1 hora sem aeração) e razão de recirculação de efluente igual a 5. A carga carbonácea afluente foi mantida constante (1,07 kg DQO.m-3.dia-1), ao longo de todo o período experimental, sendo as relações C/N testadas (9,7±1; 7,6±1; 2,9±1 e 2,9±0,4) obtidas a partir da variação na carga nitrogenada aplicada. Duas fontes orgânicas foram avaliadas: sacarose e peptona de carne. A eficiência média de remoção de DQO manteve-se acima de 90%. A eficiência máxima de remoção de N-total (84,6±10,1%) foi obtida para relação C/N de 2,9±1, com concentrações efluentes médias de NTK e N-NH4 de 5,9 e 4,3 mg.L-1, respectivamente. A análise estatística da eficiência de remoção de N-total confirmou que a fonte de carbono não exerceu influência sobre os processos de remoção. A obtenção de eficiências de desnitrificação superiores às calculadas estequiometricamente, em função da fonte de carbono, indicou a ocorrência de possíveis vias complementares para remoção de nitrogênio, como o processo anammox. As velocidades de nitrificação e desnitrificação obtidas nos ensaios cinéticos foram similares para as duas fontes de carbono e para as relações C/N estudadas e da mesma ordem de grandeza das apresentadas na literatura, reforçando a ideia de que a configuração de reator utilizada, aliada às adequadas condições operacionais, permitiu a remoção concomitante de matéria carbonácea e nitrogenada. / This study aimed to evaluate the influence of COD/N ratio and organic source on the simultaneous nitrification and denitrification (SND) in an up-flow structured-bed reactor. The reactor was made of acrylic with a total volume of 11 L and a working volume of 5.5 L and filled with cylinders of polyurethane foam as support for biomass growth. The system was continuously operated with an HRT of 11.2±0.6 h, intermittent aeration (2h with aeration and 1h without aerationd) and a liquid recycle ratio equal to 5. The organic load rate was constant (1.07 kg COD.m-3d-1) during the entire experiment. The COD/N ratios (9.7±1; 7.6±1; 2.9±1 and 2.9±0,4) were obtained from the variation of nitrogen load rate. Two organic sources were evaluated: sucrose and meat peptone. The average COD removal efficiency was above 90%. The maximum total nitrogen removal efficiency was 84.6±10.1%, with average TKN and NH4+-N effluent concentrations of 5.9 and 4.3 mg.L-1, respectively. Statistical analysis of total nitrogen removal efficiency confirmed that the carbon source did not influence over the removal processes. The denitrification efficiencies higher than the stoichiometrically calculated in function of the carbon source, indicated the occurrence of possible paths for nitrogen removal of nitrogen as anammox process. The nitrification and denitrification rates obtained in kinetic experiments were similar for the two sources of carbon and all C/N ratio studied at the same order of magnitude in relation to those reported in the literature, enhancing that the reactor configuration tested combined with the proper operational conditions allowed the organic matter and nitrogen removal simultaneously.