Evaluation of a modified model for deep bed filtration

Abstract

The status of fundamental models of the performance of deep bed filters is reviewed and a modified model for monodisperse suspensions is presented. Particle removal is described by a modified version of the Vigneswaran and Tulachan model for the complete cycle of transient filter performance. A modification of the O'Melia and Ali approach and two linear approaches for headloss modeling are also described. The sensitivity of model predictions to variations in parameter values is presented. A comparison with experimental results for ripening and steady state conditions shows that particle removal can be described well while the observed linear (with time) headloss results cannot be described by modifications of the O'Melia and Ali approach. A surface area based linear headloss model appears to capture the effects of suspended particle size on headloss development while effects of coagulant type must be described by changes in model parameter values.