Ce mémoire présente une méthodologie pour identifier des stratégies hydrauliques en réseau de distribution dans le but d’assurer des concentrations suffisantes de chlore résiduel en extrémités de réseau. La recherche a été réalisée en deux phases. La première consistait à identifier les zones de distribution de l’eau potable dans le réseau d’aqueduc à l’étude à l’aide d’une étude au traceur. La seconde phase du projet visait à prédire les concentrations de chlore résiduel en réseau de distribution ainsi que d’émettre des recommandations pour améliorer les concentrations de chlore résiduel en extrémités de réseau. Le secteur sélectionné pour l’étude est un quartier résidentiel de la ville de Québec alimenté en eau potable par deux sources : un réservoir situé en réseau et une usine de traitement de l’eau potable (UTE). Le système de distribution du secteur est caractérisé par la présence de nombreuses extrémités de réseau puisqu’il ne possède aucune interconnexion avec les réseaux limites au sud et à l’est. Les consommations étant faibles à ces endroits, les temps de séjour y sont élevés. De plus, le réseau est le dernier secteur à être desservi dans la partie à l’est de l’aire de distribution de l’UTE. Ces conditions mènent à une dégradation de la qualité de l’eau potable dans le système de distribution. La méthode proposée combine une caractérisation hydraulique, une étude au traceur ainsi qu’une étude de la qualité de l’eau pour identifier le patron de distribution de l’eau en réseau. Également, le modèle a été calé au niveau de l’hydraulique et de la qualité de l’eau potable à partir de données inédites sur la distribution de l’eau, les temps de séjour et la variabilité de la qualité de l’eau dans le réseau. Les données nécessaires au calage du modèle ont été obtenues à partir d’une campagne intensive d’échantillonnage. L’étude au traceur a été réalisée par injection d’une saumure de chlorure de calcium (CaCl2). Les résultats obtenus ont permis d’associer au réservoir les points d’échantillonnage dont la concentration en calcium a augmenté au cours de l’essai. Les hypothèses élaborées suite à l’étude au traceur ont été validées à partie des résultats d’une campagne de caractérisation de la qualité de l’eau effectuée simultanément. Cette validation a été rendue possible par les différences entre les valeurs de chlore résiduel aux deux entrées, dues au fait que l’eau provenant du réservoir y a été re-chlorée avant d’être acheminée dans le réseau de distribution. De plus, cette campagne a permis d’identifier la source d’alimentation en eau potable des points dont la concentration en calcium était demeurée constante au cours de l’essai au traceur. Par la suite, le calage du modèle a permis d’en augmenter les capacités prédictives jusqu’à un niveau jugé satisfaisant. Sur la base du modèle de qualité de l’eau, des stratégies hydrauliques ont été évaluées afin d’optimiser les concentrations en chlore résiduel en extrémité de réseau. Cinq groupes d’essais ont été explorés : (1) fermetures de vannes sur des conduites importantes, (2) ajout de pompe re-circulatoire avec injection d’une solution d’hypochlorite, (3) régulation de pression aux entrées du secteur, (4) ouverture d’une interconnexion avec le réseau limitant à l’est, et (5) combinaison des essais 2 et 4. Suite aux différentes simulations, deux stratégies hydrauliques se sont démarquées et des recommandations ont été émises. Celles-ci impliquent l’ajout d’une pompe re-circulatoire et l’injection d’une solution hypochlorite ainsi que l’ouverture d’une interconnexion avec le réseau limitant à l’est. / This Master’s project presents a methodology that can effectively identify distribution zones within a drinking water distribution network in order to ensure sufficient concentrations of residual chlorine at the dead-ends of a distribution system. The project was carried out in two phases. The first phase of the project consisted in identifying water distribution zones within the network with a tracer study. The second phase was designed to predict concentrations of residual chlorine within the distribution network, as well as make recommendations in hydraulic strategies to improve these concentrations in problematic areas of the network. The sector selected for the study is a residential district of Quebec City supplied with drinking water directly from the main supply pipes from the water treatment plant and by re-chlorinated water from a reservoir. The water distribution network is characterized by several dead-ends due to the absence of interconnexions with the bordering networks to the north and east and is also the last sector to be supplied on the eastern side of the plant’s service area. Unfortunately, water consumption is very low in these areas resulting in long standing time in the water pipes. These conditions lead to the degradation of water quality. The proposed methodology engages three strategies: a hydraulic characterization of the distribution network, a tracer study and a water quality characterization study. The hydraulic and water quality models were calibrated with original data characterizing the spatial and temporal water quality variations within the network. An intensive sampling campaign was conducted to obtain the necessary data for the calibration of the models. The tracer study was conducted by injecting CaCl2 brine at the reservoir outlet in order to increase calcium concentrations in water from the reservoir. It was then possible to associate sampling points at which calcium concentrations showed an increase during the test as being supplied by the reservoir. A water quality characterization study was conducted simultaneously with the tracer study, making it possible to validate the distribution zones identified. This validation was made possible by the difference between chlorine values at the two entry points: water from the reservoir has distinctively higher levels due to re-chlorination. In addition, it was possible to identify distribution zones at points where the calcium concentration had remained constant during the tracer test. Thereafter, the hydraulic and water quality models were calibrated satisfactorily providing a tool for evaluating hydraulic strategies to boost chlorine concentrations in problematic areas. Five types of hydraulic strategies were explored: (1) valve closures on major pipelines, (2) the addition of a re-circulating pump with chlorine injection, (3) the regulation of pressure at the main entry points of the sector, (4) the opening of an interconnexion with the east neighbouring network, and (5) a combination of strategies 2 and 4. Following the various simulations, two hydraulic strategies were more efficient than the others: the addition of a re-circulation pump with the injection of a hypochlorite solution and the opening of the interconnexion with the east neighbouring network.

Dans certaines usines de traitement d’eau potable, l’ozone est utilisé comme agent oxydant et désinfectant puisqu’il permet de diminuer efficacement la présence de matière organique naturelle (MON), les goûts et odeurs et les microorganismes présents dans les sources d’eau. Toutefois, l’usage de ce produit oxydant, en présence de MON, entraîne la formation de sous-produits de la désinfection ozonés (SPDO) tels que les aldéhydes nonhalogénés (NON-HAL), les haloacétaldéhydes (HAL) et les bromates, substances potentiellement dangereuses pour la santé humaine. Très peu de recherches à grande échelle ont permis de documenter la présence de ces SPD dans l’eau potable. Le nombre important d’échantillons générés dans cette étude de cas permet de documenter les niveaux de ces sous-produits de la désinfection dans les usines et les réseaux de distribution à l’étude. Les résultats démontrent des teneurs en NON-HAL variant entre la limite de détection de la méthode (LDM) et 55,7 μg/L, avec le formaldéhyde et l’acétaldéhyde en plus grande proportion. La présence de HAL se limite principalement à l’hydrate de chloral (CH). Les concentrations de CH varient entre < LDM et 32,5 μg/L. La concentration de bromate dans l’eau a varié entre < LDM et 14 μg/L. Il a aussi été démontré que le chlore liquide peut favoriser la formation de SPDO. De plus, l’injection de chlore liquide dans l’eau est à l’origine de deux transformations de l’acétaldéhyde. Des schémas réactionnels ont été proposés pour présenter ces transformations; celles-ci sont plus rapides en eau chaude qu’en eau froide. À cet effet, la variation saisonnière des SPDO a été traitée dans ce mémoire. Les saisons chaudes semblent favoriser la présence de ces SPD. La dose d’ozone et la nature de la matière organique naturelle sont aussi des facteurs expliquant la variation des SPDO dans les réseaux à l’étude. Dans un deuxième temps, les concentrations de chlorite et de chlorate ont été suivies dans les réseaux de distribution à l’étude. L’utilisation de solutions d’hypochlorite de sodium concentrées démontre une hausse des concentrations de chlorite, de chlorate et de bromate dans l’eau. En réseau de distribution, la rechloration au chlore liquide tend à hausser les concentrations de ces contaminants. Toutefois, les niveaux observés sont toujours sous les normes établies.

La norma de instalaciones sanitarias I.S. 010 (2006) no impide la instalación de elementos que puedan ahorrar energía y agua para el caso de agua caliente, solo presenta indicaciones con respecto al uso de los sistemas de agua, desagüe y de las conexiones a utilizar dentro de una edificación, así como las dimensiones y las herramientas adecuadas según el consumo para el cual se diseñará el edificio. Con base en estos datos mencionados, la investigación presentada tiene como núcleo la aplicación de elementos que permitan la recirculación del agua potable en una vivienda, de tal manera que se evite el desperdicio de agua, puesto que, en la actualidad, el mundo pasa por un grave problema con respecto a los recursos hidrológicos. Asimismo, según estudios realizados por la ONU y presentados por la BBC News: “el aumento de la población anual es cada vez mayor en el mundo y asimismo es un indicativo del aumento del consumo de agua y se prevé que para el año 2050 la mitad del mundo estaría pasando por una gran escasez de agua”. En la actualidad, el sistema de agua caliente en las edificaciones genera pérdidas de agua potable debido a que cuando se abre el grifo de agua, debe pasar un tiempo pronunciado para que el agua alcance el nivel de temperatura requerido por el consumidor. En consecuencia, se genera un desperdicio de agua fría, es así que se tiene planeado implementar este sistema, el cual permitirá ahorrar en agua potable y evitar futuras pérdidas, que medidos anualmente conforman una cantidad en metros cúbicos que no es despreciable. Para la aplicación de estos elementos se plantea presentar los beneficios que implicaría su instalación en edificaciones peruanas. Finalmente, se presenta la instalación de este sistema de recirculación de agua potable aplicable en agua caliente para una edificación denominada PROYECTO CUSCO, ubicada en la ciudad de Cusco, distrito de Wánchaq. En la que se presenta la distribución de este nuevo sistema en el proyecto y se cumple con los requisitos impuestos por la norma I.S. 010 (2006) correspondiente a instalaciones sanitarias del reglamento nacional de edificaciones. Como conclusión, se demuestra que es aplicable la utilización de sistemas de recirculación y que benefician no solo al consumidor sino también al planeta, teniendo más opciones de proteger nuestro entorno para las generaciones futuras.

Soil aquifer treatment (SAT), the removal of contaminants during percolation through soil, is a strategy employed in managed aquifer recharge (MAR), one method of indirect potable water reuse. As part of Hampton Roads Sanitation District's (HRSD) MAR project, The Sustainable Water Initiative for Tomorrow (SWIFT), a soil column study was performed using four columns filled with sand taken from the Potomac Aquifer System (PAS) as well as water from various stages in SWIFT's 1MGD demonstration facility. Two pairs of two columns were operated in series, simulating 3 days and 1 month of travel time through aerobic to anaerobic conditions. During Phase 1 of testing, each pair of columns was fed from different stages in the SWIFT treatment process. During Phase 2 of testing, one set of columns was spiked with a conservative tracer bromide, and several contaminants of emerging concern (CECs). The contaminants monitored during both phases included total organic carbon (TOC), nitrogen species, and the disinfection byproducts bromate and NDMA. During Phase 2 of testing, CECs, iron, arsenic, bromide, and sulfate were monitored in addition to those monitored during Phase 1. About 50% of the TOC was removed within 3 days of travel time, with no additional removal observed in 1 month. Nitrate was conserved in the 3-day columns, but completely removed after 1 month, indicating denitrification. Bromate and NDMA were reduced significantly in the 3-day columns and mostly non-detect in the 1-month effluent. Many of the spiked CECs were reduced significantly in the 3-day column indicating degradation. Three compounds exhibited some retardation through both columns but were not degraded. A few compounds, notably perfluorooctanoic acid (PFOA), showed no retardation or degradation. / Master of Science / In order to continue to meet the water demands of the future, potable reuse is a necessary and effective solution. HRSD's SWIFT project aims to create a sustainable source of drinking water through advanced treatment of its wastewater effluent and subsequent recharge of the Potomac Aquifer in a process known as managed aquifer recharge (MAR). During MAR, chemical and biological contaminants are attenuated or removed through a process known as soil aquifer treatment (SAT). HRSD installed pilot-scale soil columns at their 1MGD SWIFT demonstration facility to evaluate the potential removal of contaminants. During the study, removal of contaminants, both regulated and unregulated, was observed. This study demonstrated that SAT provides an effective environmental barrier against many contaminants and helped to inform the level of treatment necessary to protect public health during MAR potable reuse projects.

Attached growth biological processes offer advantages over traditional water purification technologies through high biomass retention, easy sludge-water separation, multiple multispecies synergies in proximity, resilience to shock loading, low space requirements, and reactor operational flexibility. Traditionally, attached growth refers to biofilms that require abiotic carrying media for bacteria to attach and grow on. While biofilms have been broadly applied in wastewater treatment, its potential for potable reuse or stormwater treatment has not been well studied. The treatment trains of pre-ozonation followed by biologically active filtration (ozone- BAF) is an advanced biofilm technology for potable reuse that can generate high-quality potable water at reduced energy and chemical demands by removing pollutant through three different pathways: oxidation, adsorption, and biodegradation. However, these pathways can result in both desirable and undesirable effects, and the mechanism behind it is still unclear. To understand the mechanisms of various pollutant removal, parallel performance comparisons of ozone-BAF treatment trains with spent and regenerated granular activated carbon (GAC), along with a range of pre-oxidant ozone doses were performed. Another common issue of BAF is the headloss buildup during its operation, which has become a significant energy and maintenance burden at many utilities. Thus, a mathematical model was developed to predict BAF headloss buildup in response to organic removal and nitrification. For stormwater treatment, the feasibility of using biofilms for stormwater biological nitrogen removal (BNR) is still largely unknown, as very limited research effort has been dedicated to this aspect. Thus, a mathematical model was developed to evaluate the potential of using BNR techniques for stormwater nitrogen removal. Aerobic granules are an even more advanced attached growth process, which eliminates the need for abiotic carrying media. So far, aerobic granular sludge is only formed in sequential batch reactors but not in a continuous flow system. Therefore, continuous flow aerobic granulation from traditional activated sludge was investigated and, for the first time, successfully achieved in continuous flow plug-flow bioreactors fed with real municipal wastewater. Besides, the role and critical value of an essential operational parameter, feast/famine ratio, for continuous flow aerobic granulation were determined. / Doctor of Philosophy / Water scarcity and increasing water demand caused by urban population growth and climate change is a reality throughout the world. Thus, process intensification of the current water and wastewater technologies is gaining increasing attention globally. Comparing to traditional water purification technologies, attached growth biological processes offers advantages such as high biomass retention, easy sludge-water separation, multiple multispecies synergies in proximity, resilience to shock loading, small footprint requirement, and reactor operational flexibility. Traditionally, attached growth refers to biofilms that require abiotic carrying media for bacteria to attach and grow on. While biofilms have been broadly applied in wastewater treatment, its potential for potable reuse or stormwater treatment has not been well studied. For potable reuse, the treatment trains of pre-ozonation followed by biologically active filtration (ozone-BAF) is an advanced biofilm technology that can generate high-quality potable water at reduced energy and chemical demands by removing pollutant through different pathways. However, the mechanism behind it is still unclear. To understand the mechanisms of various pollutant removal, parallel performance comparisons of ozone-BAF treatment trains operated with different operational conditions were performed in this dissertation. Another common issue of BAF is the headloss buildup during its operation, which has become a significant energy and maintenance burden at many utilities. Thus, a mathematical model was developed to predict the headloss buildup during BAF operation. For stormwater treatment, the feasibility of using biofilms for stormwater biological nitrogen removal (BNR) is still largely unknown, as very limited research effort has been dedicated to this aspect. Thus, a mathematical model was developed to evaluate the potential of using BNR technique for stormwater. Aerobic granules are an even more advanced attached growth process. However, aerobic granular sludge is so far only formed in sequential batch reactors which are incompatible with the continuous flow nature of most wastewater treatment plants. Therefore, aerobic granulation from traditional activated sludge was investigated and, for the first time, successfully achieved in continuous flow plug-flow bioreactors fed with real municipal wastewater. Besides, the role of an essential operational parameter, feast/famine ratio, for continuous flow aerobic granulation was determined.

Soil aquifer treatment (SAT) during managed aquifer recharge has been studied as a method of providing additional environmental barriers to pathogens and contaminants in indirect potable reuse (IPR) applications. A soil column study was conducted by Hampton Roads Sanitation District in order to evaluate the effectiveness of SAT, as a component of its IPR project involving the replenishment of the Potomac Aquifer System (PAS), in providing a sustainable source of drinking water. Four packed soil columns were constructed with sand from the PAS and were designed to simulate the travel time of 3 days and 30 days. The tests conducted aimed at evaluating pathogen removal (MS2, E. coli and Cryptosporidium oocysts); evaluating attenuation of regulated (nitrate, nitrite, bromate, trihalomethane (THM), haloacetic acids (HAA), organic carbon) and unregulated contaminants of concern that affect drinking water quality. Effective pathogen removal was observed with 6 to 7-log removals of MS2 and E. coli and 3 to 5-log removals of microbeads, used as a surrogate for Cryptosporidium. Removal across 3 day columns was comparable to 30-day columns but the potential to achieve higher removal with longer retention time was acknowledged. Nitrate, bromate, THMs and HAAs were completely reduced in 30-day columns. Total organic carbon was removed at 25 – 35% in all four columns. Seven out of the 106 contaminants of emerging concern (CEC) tested were consistently detected in the column feed and effluent at concentrations greater than 100 ng/L; some compounds showed potential for removal while no conclusive results were drawn for the remaining compounds. / MS / Potable reuse is a sustainable solution to the increasing water demands of the present and more so the future. Hampton Road Sanitation District (HRSD) aims to treat effluent from its wastewater treatment plants using advanced treatment process for direct recharge of the Potomac aquifer system. This is a method of indirect potable reuse termed as managed aquifer recharge (MAR). MAR can provide additional environmental barriers to contaminants present in water through a process of natural attenuation called soil aquifer treatment (SAT). A soil column study was conducted at HRSD’s pilot scale facility in order to asses SAT under controlled conditions at a meaningful scale. Attenuation of pathogens; regulated contaminants (nitrate, nitrite, bromate, etc.) and contaminants of emerging concern was evaluated through the soil columns. The results showed effective removal of most contaminants of interest which demonstrated that SAT potentially improves water quality and meets public health standards in potable reuse applications.

The Potomac River Basin is significant for both public and ecological health as it flows directly into the ecologically-sensitive Chesapeake Bay. It is a drinking water source for about 5 million people living in Maryland, Virginia, and Washington D.C. The discovery of intersex fish, an indicator of poor ecological health, in the Chesapeake Bay occurred in the 2000s, and has led to a series of studies in the watershed to determine the sources and magnitude of endocrine disruption. Endocrine disrupting compounds (EDCs) are exogenous chemicals that interfere with the endocrine system and can cause detrimental health effects at low concentrations. This study aims to understand a best management practice referred to as planned indirect potable reuse (IPR) and its impacts on EDCs. The Occoquan Watershed is a planned IPR subwatershed of the Potomac River Basin. Water samples were collected at the water reclamation plant discharge (Upper Occoquan Service Authority), up- and downstream of that location along Bull Run, and at the water treatment plant intake (Frederick P. Griffith WTP) in the Occoquan Watershed to assess planned IPR. Samples were also collected at a water treatment plant (James J. Corbalis WTP) along the Potomac River for comparison as an unplanned IPR location. These two groups of samples were analyzed for EDCs (categorized into two groups: estrogen hormones and other synthetic organic compounds (SOCs)), nutrients, and other water quality parameters. The infrequency of estrogen hormones and SOC patterns indicate planned and unplanned IPR are both viable approaches to provide safe drinking water / Master of Science / Our river systems are important to maintain both for human and environmental health. The Potomac River Basin is the area of land drained by the Potomac River and its tributaries. The Potomac River Basin is significant for both public and ecological health as it flows directly into the ecologically-sensitive Chesapeake Bay. It is a drinking water source for about 5 million people living in Maryland, Virginia, and Washington D.C. The discovery of intersex fish, or when a single fish has both male and female characteristics, occurred in the Chesapeake Bay in the 2000s. Fish health is often an indicator of poor environmental health, and in this case endocrine disruption. This discovery led to a series of studies in the watershed to determine the sources and magnitude of endocrine disruption. Endocrine disrupting compounds (EDCs) are external chemicals that interfere with the endocrine system once they enter the body of a human or another organism, and can cause detrimental health effects even at low concentrations. This study aims to understand a best management practice, or a type of water pollution control, referred to as planned indirect potable reuse (IPR) and its impacts on EDCs. IPR occurs when wastewater from a community is discharged into to a river or a reservoir, and then downstream it is withdrawn from that same source for drinking water purposes. This can be either planned or unplanned. Planned IPR is becoming more common as population, especially in urban areas, increases. The Occoquan Watershed is a planned IPR subwatershed of the Potomac River Basin. Water samples were collected at the water reclamation plant discharge (Upper Occoquan Service Authority), up- and downstream of that location along Bull Run, and at the water treatment plant intake (Frederick P. Griffith WTP) in the Occoquan Watershed to assess planned IPR. Samples were also collected at a water treatment plant (James J. Corbalis WTP) along the Potomac River for comparison as an unplanned IPR location. These two groups of samples were analyzed for EDCs (categorized into two groups: estrogen hormones and other synthetic organic compounds (SOCs)), nutrients, and other water quality parameters. The infrequency of estrogen hormones and SOC patterns indicate planned and unplanned IPR are both viable approaches to provide safe drinking water.

The Occoquan Reservoir is a public water supply in northern Virginia. The Occoquan Watershed has developed over the years from rural land uses to metropolitan suburbs within easy commuting distance from Washington, DC. Due to this urbanization, the Occoquan Reservoir is especially vulnerable to hypereutrophication, which results in problems such as algal blooms (including cyanobacteria), periodic fish kills, and taste and odor problems. In the 1970's, a new management plan for the Occoquan Reservoir called for the construction of the Upper Occoquan Sewage Authority (UOSA), an advanced wastewater treatment plant that would take extraordinary measures for highly reliable and highly efficient removal of particulates, organics, nutrients, and pathogens. Eliminating most of the water quality problems associated with point source discharges in the watershed, this state-of-the-art treatment is the foundation for the successful indirect surface water reuse system in the Occoquan Reservoir today. A limnological analysis of thirty years of water quality monitoring data from the reservoir and its two primary tributaries shows that the majority of the nutrient and sediment load to the reservoir comes from nonpoint sources, which are closely tied to hydrometeorologic conditions. Reservoir water quality trends are very similar to trends in stream water quality, and the tributary in the most urbanized part of the watershed, Bull Run, has been identified as the main contributor of sediment and nutrients to the reservoir. Despite significant achievements in maintaining the reservoir as a source of high quality drinking water, the reservoir remains a phosphorus-limited eutrophic waterbody. / Master of Science

Cette thèse a pour objectif d'améliorer la compréhension et les connaissances des problèmes liés à l'occurrence de molécules odorantes présentes dans les eaux brutes et les eaux traitées. Les cas à l'étude sont trois réseaux d'eaux potables dans les villes de Québec et Lévis (Canada). Cette thèse est divisée en quatre chapitres: Le premier chapitre porte sur la mise au point d'une méthode d'analyse pour détecter certaines molécules odorantes. Il introduit une nouvelle méthode analytique pour quantifier cinq composés odorants: géosmine, 2-méthylisobornéol, 2-isopropyl-3-méthoxypyrazine, 2-isobutyl-3-méthoxypyrazine et 2, 4, 6-trichloroanisole. Cette méthode utilise une micro-extraction sur phase solide en espace de tête, combinée à une analyse chromatographique en phase gazeuse (CG) et un spectromètre de masse (SM), de type trappe ionique. La double fragmentation (tandem SM/SM) a été optimisée pour les composés ciblés. Cette méthode a permis d'obtenir des limites de détection inférieures à 1 ngL"1. Elle a été appliquée avec succès pour l'analyse des composés conférant des odeurs de "terre et moisi" aux sources d'eaux municipales, riches en matières organiques, ainsi qu'aux eaux traitées correspondantes. Le second chapitre concerne les relations entre deux composés odorants (géosmine et 2-méthylisobornéol) et 39 paramètres de qualité d'eau (chimiques, physiques et microbiologiques) dans trois sources d'alimentations très différentes en termes de qualité, lesquelles alimentent en eau potable les villes de Québec et Lévis (Canada). Cette étude a été menée par analyse en composantes principales (ACP). L'objectif de cette étude a été de montrer qu'une approche multidmensionnelle par analyse en composantes principales utilisant la matrice des composantes permettait de différencier les processus à l'origine de l'occurrence de ces composés odorants, de ceux qui n'en étaient pas ou peu la cause et ainsi d'extraire les variables les plus représentatives de ces processus. Le troisième chapitre se focalise sur la modélisation des concentrations de géosmine dans les eaux brutes par des méthodes linéaires et non-linéaires telles que : la régression linéaire multiple, la régression par composantes principales et les réseaux de neurones artificiels. Les méthodes de sélection des variables ont également fait l'objet d'une étude: "pas à pas", d'analyse de sensibilité, d'analyse en composantes principales. Le développement d'outils d'aide à la décision pour de telles problématiques semble une solution rapide et économique pour le remplacement des méthodes analytiques. Le quatrième et dernier chapitre porte sur l'évolution spatiale et temporelle des concentrations de géosmine et 2-méthylisobornéol dans les eaux brutes, les eaux en cours de traitement en usine et les eaux potables des villes de Lévis et Québec. Une attention particulière a été donnée à l'enlèvement des ces deux molécules par les procédés de clarification et d'ozonation ainsi qu'au devenir de ces molécules dans les réseaux d'eaux potables.

L’utilisation d’oxydants, comme l’hypochlorite de sodium, est une stratégie très efficace en traitement de l’eau pour inactiver les microorganismes responsables de plusieurs maladies. Toutefois leur utilisation favorise également la formation de sous-produits de la désinfection (SPD), tels que les trihalométhanes et les acides haloacétiques. La présence d’iodure et de bromure dans l’eau brute peut également conduire à la formation d’autres familles de SPD, les trihalométhanes iodés (THMi) et les acides haloacétiques iodés (AHAi). Cette étude a comme objectif d’évaluer la variabilité des THMi et AHAi dans deux petits réseaux de distribution de la région de Québec (notés R1 et R2) ainsi que de mesurer l’enlèvement des THMi lors de l’utilisation de filtres domestiques. L’échantillonnage de l’eau s’est effectué entre juin et décembre 2017 sur différents points dans l’usine de traitement des eaux et dans le réseau de distribution. Pour le réseau R1, des concentrations moyennes de 2,99 μg/L en THM iodés et de 0,74 μg/L en AHA iodés sont observées, alors que la concentration maximale obtenue est de 4,02 μg/L pour les THM iodés et de 2,46 μg/L pour les AHA iodés. Des concentrations moyennes de 2,98 μg/L en THM iodés et de 0,51 μg/L en AHA iodés sont observées dans le Réseau R2, alors que la concentration maximale obtenue est de 4,05 μg/L pour les THM iodés et de 2,33 μg/L pour les AHA iodés. Les concentrations en THM iodés et en AHA iodés augmentent dans l’usine, particulièrement à la suite de la post chloration et du passage de l’eau par le bassin de contact. Les concentrations en THM iodés restent ensuite stables dans le réseau de distribution alors que celles des AHA iodés diminuent. De plus, l’enlèvement des THMi s’avère très efficace après la filtration de l’eau sur un filtre domestique de type BritaMD, avec un pourcentage d’enlèvement moyen de 100%. / The use of oxidants, such as sodium hypochlorite, is a very effective strategy in treating water to inactivate microorganisms responsible for several diseases. However, their use also favors the formation of disinfection by-products (SPD), such as trihalomethanes and haloacetic acids. The presence of iodide and bromide also allows the formation of other families of SPD, iodinated trihalomethanes (THMi) and iodinated haloacetic acids (AHAi). The purpose of this study is to evaluate the variability of THMi and AHAi in two small distribution networks in the Quebec City region (R1 and R2) and to measure the removal of THMi when using household filters. Water sampling took place between June and December 2017 at various points in the water treatment plant and in the distribution network. For the R1 network, mean concentrations of 2.99 μg / L in iodinated THM and 0.74 μg / L in iodinated AHA were observed, while the maximum concentration achieved was 4.02 μg / L for THMs. iodized and 2.46 μg / L for iodinated AHAs. Mean concentrations of 2.98 μg / L in iodinated THM and 0.51 μg / L in iodinated AHA were observed in Network R2, while the maximum concentration obtained was 4.05 μg / L for iodinated THMs. and 2.33 μg / L for iodinated AHAs. Iodine THM and iodinated AHA concentrations increase in the plant, particularly because of post-chlorination and passage of water through the contact basin. The iodinated THM concentrations then remain stable in the distribution network while those of the iodinated AHAs decrease. On the other hand, the removal of THMi is very effective after filtration of water on a BritaMD type household filter, with an average removal percentage of 100%.