Modélisation statistique d'un pilote membranaire semi-industriel à l'aide de la logique floue

International audience; Membrane bioreactors (MBR) have been recently used in large scale wastewater treatment plants. Modelling the functioning of such processes could help the development of adequate control strategies. In this study, a fuzzy inference method is used to i) simulate a 4,5 months campaign on a semi-industrial MBR treating rejection water from sludge treatment, ii) discriminate the influence of operating parameters on membrane fouling. The results show not only a good representation of the permeability evolution, but also coupled influences of sludge properties and operating parameters on fouling.