Two sets of triplicate pilot algal raceway ponds (1000-L, 0.30-m deep, paddle wheel mixed) were operated for 14 months at a California wastewater treatment plant to treat wastewater and generate algal-bacterial biomass as biofuel feedstock. Two experiments were run to determine the effect on biomass productivity of (1) hydraulic residence time (HRT: 2, 3, 4, or 4.5 days) and (2) feeding schedule (18 small pulses during 8 AM-4 PM [diurnal] versus 20 large pulses during 4 AM-12 AM [diel]). The target productivity was at least 20 g volatile suspended solids per m2 of pond per day. Additional output variables were followed during the experiments: treatment performance and the effectiveness of biomass harvesting though bioflocculation. Productivity was consistently higher in ponds with a 2-d HRT versus longer HRTs. Average productivity for the 2-d HRT ponds and the variable-HRT ponds (3.6-d average HRT) were 30.1 and 23.4 g/m2-d, respectively. Productivity data collected during the feed regime experiment were highly variable, and average productivities were the same at 26 g/m2-d. During both experiments, both pond sets exceeded the target of 20 g/m2-d on an annual basis. During the hydraulic residence time experiment, the average pond productivity throughout the HRT experiment for the 2-d HRT and 3-d HRT ponds were 30.1 and 23.4 g/m2-d, respectively. Settling efficiency was high for both 2- and 3-d HRT ponds with average turbidity removal of 87-89%. However, total ammonia nitrogen (TAN) concentrations in the 2-day HRT pond effluent were 50-94% higher than in the 3-d HRT pond effluents, although effluent TAN concentrations in both ponds were approximately the same during mid-summer. Furthermore, effluent biochemical oxygen demand (BOD5) concentrations were similar, with the supernatant of Imhoff cones settled for 24 hours containing 24-27 mg/L BOD5 (81-92% removal). In general, the 3-d HRT ponds provided better treatment than the 2-d HRT ponds. During the feeding regime experiment, no productivity or BOD5 removal differences were evident. However, the 3-d HRT ponds had consistently 8 mg/L more effluent TAN than the 2-d HRT ponds.

L’objectif de cette thèse est le développement et la validation d’un outil numérique permettant d’évaluer la performance d’un ouvrage de décantation. L’approche Euler-Lagrange est retenue pour la modélisation de l’écoulement et du transport solide. De nouvelles conditions d’interaction particule/paroi permettent de restreindre le dépôt aux zones présentant des caractéristiques géométriques et hydrodynamiques favorables. Le modèle numérique est confronté à quatre jeux de données expérimentales. L’ensemble de ces expérimentations permettent d’investiguer une large gamme des paramètres représentatifs de l’écoulement et du transport solide. La comparaison des résultats numériques et expérimentaux permet de conclure sur la capacité du modèle à prévoir l’abattement des particules dans un ouvrage de décantation avec une précision de l’ordre de 5 pourcents lorsque les dépôts ont lieu sur l’ensemble du fond et de 10 pourcents dans le cas de dépôts localisés dans des zones préférentielles. / The objective of this thesis is the development and the validation of a numerical tool to estimate the trapping efficiency of settlers. The Euler-Lagrange approach is chosen to simulate the flow and the sediment transport. New particle/wall interaction conditions are developed to take into account the geometric and hydrodynamic characteristics to enable the deposit. The numerical model is compared to four sets of experimental data. All these experiments enable to study a wide range of parameters that describe the flow and the solid transport. The comparison between the numerical and experimental results shows that the model is able to estimate the trapping efficiency of particles in a settler with an accuracy of about 5 percent when the particles settle all over the bed of the basin and about 10 percent in case of located deposits.