Nowadays filtration processes are still monitored with conventional analyses like turbidity measurements and, in case of flocculation–filtration, with phosphorus analyses. Turbidity measurements have the disadvantage that breakthrough of small flocs cannot be displayed, because of the blindness regarding changes in the mass distributions. Additional particle volume distributions calculated from particle size distributions (PSDs) would provide a better assessment of filtration performance.
Lab-scale experiments have been executed on a flocculation–filtration column fed with effluent from WWTP Beverwijk in The Netherlands. Besides particle counting at various sampling points, the effect of sample dilution on the accuracy of PSD measurements has been reflected. It was found that the dilution has a minor effect on PSD of low turbidity samples such as process filtrate. The correlation between total particle counts, total particle volume (TPV) and total particle surface is not high but is at least better for diluted measurements of particles in the range 2–10 μm. Furthermore, possible relations between floc-bound phosphorus and TPV removal had been investigated. A good correlation coefficient is found for TPV removal versus floc-bound phosphorus removal for the experiments with polyaluminiumchloride and the experiments with single denitrifying and blank filtration.