Untitled

A semi-empirical crossflow ultrafiltration (UF) model that can predict permeate flux as a function of time for different operating conditions was considered. Model predictions were analysed and compared with the experimental results obtained in a crossflow UF plant under different operating conditions. The crossflow UF experiments were performed under different transmembrane pressures and crossflow velocities. Tubular ceramic membranes supplied by Orelis (France) were used. The molecular weight cut-off of the membranes was 15 kg/mol. They consisted of a TiO2–ZrO2 active layer on a carbon support. Polyethylene glycol with a 35,000 g/mol molecular weight was used as a standard macromolecule in the UF experiments to test the model. A solute concentration of 5 g/L was considered. Model predictions were better for low crossflow velocities.

One of the disadvantages of this model is the fact that it is not possible to estimate the parameters χ and δ theoretically. However, the model is interesting because it predicts permeate flux decline with time only from steady-state experimental data. Although the model is semiempirical, it does not require experimental data from flux decline with time to perform the empirical estimation. The model predicts lower permeate flux than that experimentally observed; consequently, it predicts more fouling than that experimentally observed. Model predictions are better for a crossflow velocity of 1 m/s than for 2 m/s.