Modelling head loss in deep bed filtration of hydroxide flocs

Abstract

The modified filter coefficient model has been adapted to predict the head loss during the deep bed filtration of hydroxide flocs. The differences between the drop in pressure calculated by the model and the measured at an industrial pilot plant decrease from more than 150% to less than 5% if the constant bulking factor is replaced by a variable optimal bulking factor. This variable is a function of both filtration time and filter depth. In this investigation, the small dependence of the optimal bulking factor on the depth is neglected. The different variations of the general model are tested with the same experimental data and compared with each other.The relation between optimal bulking factor and time is calculated with the best fitting model for 11 different experiments. The relation consists of two linear parts if none or only a very small amount of polyelectrolytes is used. With higher polyelectrolyte doses, the relation is hyperbolic.The optimal bulking factor is a combination of (1) a constant bulking factor for the flocs in the influent, (2) a slowly decreasing bulking factor for the deposits in the filter bed which are subjected to ageing phenomena and (3) a sharply decreasing bulking factor, translating the volume of voids blocked during the beginning of a filter run.