Authors: L.Mauclaire, A. Schurmann, M Thullner, S.Gammeter and J.Zeyer

Source: IWA Publishing, Journal of Water Supply: Research & Technology - AQUA|53.2|2004

Summary

This paper outlines the research carried out on the clogging of "slow" sand filters used in drinking water treatment. The researchers examined the causes of clogging of slow sand filters at a drinking water plant that draws its water from Lake Zurich in Switzerland. The treatment regime involves ozone, coagulation / rapid filtration, granular activated carbon filtration followed by "slow" sand filtration.

The research undertaken investigated the likely reasons for clogging of six "slow" sand filters (three with above average clogging and three with below average clogging based on filter flow rates) using measurements taken over a number of months. These measurements tried to characterise the impact of physical, chemical, biological and hydraulic characteristics on slow sand filters. These measurements included ammonium (as N), nitrate (as N), phosphate, sulphate, dissolved oxygen, dissolved organic carbon, pH, total hardness, temperature, bacterial numbers, polysaccharide content and hydraulic conductivity.

The results of the investigation suggested that the majority of clogging occurred in the top 5cm of the sand bed, which was demonstrated by dramatic head loss. Despite this significant clogging also occurred in deeper areas of the sand beds.

The results of the samples showed little change in nitrate, and sulphate levels during infiltration and were significantly higher in the top layer of highly clogged filters during the research period. Dissolved Organic Carbon (DOC) concentrations varied significantly during infiltration across the year and was significantly lower in less clogged filters. Bacterial abundance increased throughout November to May for both filter types and sampling depths, whereas the polysaccharide content depicted small variations with highest values in November. This was the same for both filter types and sampling depths.

The paper concluded that clogging effects were to a large extent attributed to the abundance of extracellular polymeric substances (the result of biological activity), which reduced the pore size by at least 7% in the highly clogged parts of the filters, rather than physical and chemical causes. This biomass was found to occupy about 7% to 10% of the pore space in the top layer of the filters.

Sydney Water Response:

Filtering removes the solid particles from raw water supplies and prepares the water for disinfection, prior to supplying to consumers. Sydney Water's water filtration plants use a combination of deep bed anthracite (a type of coarse, lightweight coal) top layer and a fine, heavier sand layer beneath. As water runs through the deep layer of coarse anthracite, particles are trapped through the deep layer rather than just on the surface, hence not clogging the filters too quickly. If particles are collected only on the surface (as in single layer sand filters) it would quickly clog the filter, which would need frequent flushing to clear it. After the coarse anthracite has removed most of the larger particles, the sand layer removes any finer particles that the anthracite has not captured. The effectiveness of using sand alone in filtration depends on the water and the characteristics of sand used.

"Slow" sand filters as described in the paper have traditionally been used for small water supplies having relatively low cost single treatment, or for final "polishing" of treated water. This type of filter only operates at low filtration rates requiring a large number of filter beds.

Sydney Water has adopted the use of coagulation/deep bed filtration for use at its water filtration plants, as being more effective than the type of slow sand filters discussed in the paper. The filters are regularly backwashed to keep them clean and minimise the risk of "clogging" associated with "slow" sand filters.