Résumés
Résumé
Lors de la filtration sur fibres creuses de suspensions de bentonite, il a été mis en évidence que le facteur limitant est la formation d'un dépôt à la surface de la fibre.
Parlant de ces considérations, un dispositif optique a été mis au point pour suivre l'évolution spatio-temporelle du dépôt en cours de filtration sur fibre creuse à peau externe.
Ainsi, il a pu être mis en évidence que l'influence de la vitesse d'écoulement n'est sensible qu'après un certain temps de filtration ou un certain volume filtré.
Dans le domaine testé des faibles vitesses d'écoulement, il existe une période initiale où la loi de filtration sur gâteau est applicable.Dans le cas de fibres creuses à peau externe, l'optimum technique (productivité maximale) ou économique (coût minimal du m3 traité) serait donc plus à rechercher dans une optimisation de la séquence filtration-lavage que dans une augmentation de la vitesse d'écoulement.
L'efficacité du réentraînement du dépôt lors du lavage peut être suivie grâce à l'équipement et au capteur développés.
Mots-clés:
- Microfiltration,
- contre lavage,
- formation du dépôt
Abstract
Use of membranes (MF or UF) for water treatment is now developed at an industrial scale. This approach is considered as a good solution to deal with the periodic strong increases in suspended solids contents which occur in karstic aquifers.
Under these conditions, the main contribution to the growth of hydraulic resistance is related to the formation of a deposit on the membrane surface.
Thus, an experimental study was performed at the laboratory scale in order to perform on line characterization of deposit formation during the filtration of a bentonite suspension through a hollow fiber.
The filtration module is transparent and contains only one hollow fiber with its external skin.
An optical device formerly developed was used as a sensor : a horizontal laser beam is focussed at the surface of the hollow fiber which is held perpendicular to the beam and can be moved step by step in the micrometric range.
Experiments were performed with a bentonite suspension (concentration 0.375 kg/m3) and under different operating conditions :
- dead end filtration and crossflow filtration with velocities up to 0,30 m/s in the annular space (which means Re ≈ 2000, velocity gradient ≈ 1000 S-1, shear stress ≈ 1 Pa) ;
-transmembrane pressure between 85 and 185 Kpa.
One of the main conclusions of this study is that influence of the velocity appears only after an initial period during which the growth of the deposit and the filtrate flowrate variation with time are in agreement with the equations of dead end filtration at constant pressure.
Thus a combination of dead end filtration and periodic backwash for removing the deposit might be the best solution for getting the highest productivity with minimal energy consumption.
These optimal conditions may be simulated using the equations of dead end filtration at constant pressure and making the assumption that the backwashing is fully efficicent.
The experimental study of backwashing, using the optical sensor for measuring the diameter of the fiber after each backwashing, proves that this assumption is correct and that it is possible to get a steady mean filtrate flux by periodic backwashing.
From the theoretical study, the optimal duration of the filtration period is rather short (about 10 minutes) but the decline after the maximum is very malt, which means that the fiitration period may be longer without significative variation of the productivity.
Thus the method presented here is able, if applied to actual raw water, to give technical and economical elements of comparison between dead end filtration with periodic backwashing and crossflow filtration.
Keywords:
- Membrane filtration,
- backwash,
- deposit formation