No abstract

Ultrafiltration (UF) and coagulation–ultrafiltration (CUF) of raw water from the Des Roches Lake were performed and compared in terms of natural organic matter (NOM) removal and in terms of membrane permeability variation. The study was carried out at a laboratory scale with a membrane having a nominal pore size of 0.035 μμm. The removal of NOM by UF was significant (10% to more than 50% for UV absorbance at 254 nm and 0% to more than 30% for TOC) but was also very dependent on operating conditions. Membrane fouling occurred rapidly and was strongly accelerated as the permeate flux increased. Fouling also induced a decrease of NOM removal. Coagulation before UF brought a strong increase of NOM removal (TOC removal higher than 60% and reduction of UV absorbance at 254 nm higher than 80%). It also largely reduced membrane permeability loss even for permeation flux higher than 150 L·h–1·m–2. In the absence of coagulation, and operating with full recirculation of the concentrate, there was a reduction of the SUVA ratio (UV254/TOC) and iron concentration in the feed water. This indicates a preferential deposit of humic substances onto the membranes and shows that iron either plays a role in membrane fouling or could be an indicator of the presence of fouling substances. Key words: ultrafiltration, coagulation, surface water, fouling, organic matter.

Le procédé conventionnel d'oxydation et de filtration au sable vert est difficilement applicable aux eaux souterraines très chargées en fer et en manganèse. Lors de cette étude pilote, plusieurs modifications à ce procédé ont été testées en vue d'améliorer les performances de traitement d'une eau très concentrée en fer et en manganèse : ajout d'une étape de décantation, usage de sable et d'anthracite recouverts d'oxydes de manganèse, aération forcée. Plusieurs oxydants et modes d'oxydation ont également été testés. Les résultats indiquent que la présence d'une grande quantité de fer a un effet significatif sur l'enlèvement du manganèse et confirment que l'oxydation du manganèse peut être complétée par le chlore libre dans un milieu filtrant autre que le sable vert à condition qu'il y ait un excédent suffisant de chlore libre. Dans les conditions de l'étude, il ressort que la zone cible de performances de filtration peut être atteinte avec tous les modes de fonctionnement testés et avec un dosage en aide filtrant variant de 0,25 à 0,5 mg/L. L'aération entraîne une meilleure sédimentation mais provoque aussi une augmentation plus rapide de la perte de charge dans le milieu filtrant en présence d'aide filtrant.Mots clés : eau souterraine, fer, manganèse, physico-chimique, traitement, séparation, étude pilote.Abstract: Conventional greensand treatment for the removal of iron and manganese from groundwater is not quite appropriate for waters with a high content of iron and manganese. In this pilot study, different modifications to this process were tested to improve treatment performances for water with a high concentration of iron and manganese: addition of a settling tank, use of sand and anthracite covered with manganese oxides, and aeration. Different oxidants and oxidation sequences were also tested. Results show that the presence of a high quantity of iron significantly improves removal of manganese. These results also confirm that manganese oxidation can be realized with free chlorine inside a filter bed other than greensand, provided that the excess of free chlorine is sufficient. Within the conditions studied, the target area for filtration performances can be attained for every treatment sequence tested and for a filtering aid dose ranging from 0.25 to 0.5 mg/L. Although groundwater aeration leads to better settling performances, it also accelerates head loss in the filtering bed when a filtering aid is present.

No abstract