MODELING OF PARTICULATE REMOVAL IN MIXED MEDIA FILTERS USING A POWER EQUATION

Stormwater filters currently in use are typically operated in a downflow mode. The problem with downflow operation is that the filters clog easily, which reduces the flow rate through them. Prior laboratory-scale research on the effects of clogging on the flow rate through the media showed that for both sand filters and for mixed media filters, the flow rate appears to be dependent upon the suspended solids loading on the media. In Urbonas (1999), a power equation was proposed to explain the decay in flow rate through the filter due to the suspended solids loading. The purpose of this project was to investigate if a power equation, as proposed by Urbonas, could be used to determine the change in flow rate through a mixed-media filter as the suspended solids loading on the media surface increases. Using laboratory-scale columns with a 4.8-cm inner diameter (ID), the use of a power equation to model downflow filtration was demonstrated. However, the coefficients for the sand filter tested in the laboratory-scale set-up were significantly differently from those of Urbonas.The lack of agreement between the model coefficients from Urbonas (1999) and the experiments with the 4.8-cm inner diameter columns may be due either to the small size of the columns or to differences in the sands used by the two experimenters. Additional experiments have been performed, with replicate experiments ongoing, to determine if there is a scale-up effect occurring and if so, if it is significant. This second set of experiments is being performed using 10-cm inner diameter columns but using the same sand and the same test-clay concentrations as the earlier experiments performed by Clark (2000). The results of these more recent clogging tests indicate that there is a significant scale-up effect even between these two laboratory-scale columns. Research is currently underway to replicate the experiments with the 10-cm inner diameter columns to determine if this effect is quantifiable.