Evaluation of a Filter Backwash Recovery Plant to Establish Guidelines for Design and Future Operation

Feb 2004

Executive Summary

Spent filter backwash water is characterised by a high concentration of suspended solids. Algal cells, invertebrates, bacteria, viruses and protozoa trapped by the sand filters will be present in higher concentrations in spent filter backwash water compared to the clarified water. Recycling spent filter backwash to the head of works without treatment may upset the treatment process and/ or increase the risk of affecting the final water quality.

A filter backwash recovery plant (FWWRP) was constructed at Rand Water's Zuikerbosch Pumping Station to treat spent filter backwash to potable standard. The plant was commissioned in early 1998 and a monitoring and evaluation program was initiated in August 1998. The technology included a conventional treatment plant, consisting of pretreatment, coagulation, flocculation, high rate clarification, sand filtration and tertiary experimental treatment units such as ozone, granular activated carbon (GAC) and membrane filtration treatment. The aim of the project was to establish guidelines for the design and future operation of filter backwash recovery plants. Emphasis was placed on the removal of suspended solids, protozoa, algae and taste and odour causing compounds.

The suspended solids in the spent filter backwash were effectively reduced from 352 NTU to 0,18 NTU in the filtered water. Water with turbidity of between 5 and 10 NTU were fed onto filters and a filtered water turbidity of less than 0,3 NTU was consistently produced. The dosing of a filter aid ahead of the sand filters was therefore not required to achieve the target filtrate turbidity of 0,5 NTU. More particles were observed in the filtered water compared to the filtered water of the main process, although very similar turbidities were measured in the samples. The standard plate count of the filtered water was slightly higher than counts observed in the filtered water of the main treatment process at Rand Water. No protozoa cysts or oocysts were detected in the filtered water at the filter backwash recovery plant during the investigative period. Very high invertebrate numbers were detected in the spent filter backwash at times. Invertebrate numbers in the filter effluent exceeded the Rand Water recommended limit but complied with the maximum permissible limit. The chlorine demand of the filtered water was the same as in filtered water from the main treatment process. The average total organic carbon concentration of the filtered water was determined to be 3,7 mg/l. The levels of the inorganic species present in the filtered water were all within the SANS 0241 (2001) limits for Class O water. Algal enumeration and speciation yielded a variety of species present in the spent filter backwash at low levels. Algae were absent from nearly all of the samples, which were also reflected in the low concentrations of chlorophyll a. No taste and odour problems were experienced over the duration of the project. Spent filter backwash could therefore be treated with success to potable standard in a conventional treatment plant. The optimum dosages of FeCl3 and cationic polyelectrolyte were determined as 3,5 mg/l as Fe and 6 mg/l respectively. Anionic polyelectrolytes may perform better than the cationic polyelectrolytes but care must be exercised not to exceed the maximum permissible dosage. The recycling of sludge is advantageous to the efficiency of suspend solids removal in the clarifier. The best results were achieved when the highest density sludge was recycled. Turbidities of 5 to 10 NTU were fed onto the sand filter. Decreasing the upflow rate through the laminar plate settler could reduce the overflow turbidity to less than five NTU.

An ozone dosage of 0,36-mg O3/mg TOC increased the assimilable organic carbon (AOC) by 39%. A 5-minute empty bed contact time (EBCT) was not enough to reduce the AOC levels to that of the feed water. The TOC adsorption capacity of the granular activated carbon (GAC) was reduced within a short period (10 000 to 15 000 bed volumes) to reach a constant TOCeffluent/TOCfeed. The same pattern was observed for all the GAC steams. The trend was also observed for the UV absorbing substances, except that steady state conditions were reached at different UVeffluent/UVfeed ratios. More UV absorbing substances would be removed when ozonation precedes GAC. Chlorine demand was reduced quite significantly during GAC filtration.

Superior permeate quality was delivered by the microfiltration plant. The transmembrane pressure (TMP) developed quickly at a flux of 100 l/m2.h (LMH) and a backwash interval of 30 minutes. Both acidic and caustic "cleaning in place" (CIP) procedures were required to remove the compounds responsible for fouling the membrane surface. Membrane treatment is still quite an expensive process in South Africa and high flux rates would be required for it to be competitive. The dosing of chemicals upstream of the membrane might have contributed to the blocking of the membrane. Poor control over the feedwater quality to the membrane plant would reduce the life of the membranes and make it an even more expensive technology to use.

The increased risk of protozoa in the potable water may necessitate the use of technology capable of removing or inactivating protozoan oocysts. Recently published literature showed that protozoa could be inactivated at low dosages by both low and medium pressure ultraviolet (UV) radiation. An UV unit was tested to determine the performance limits of the technology. Standard plate counts of approximately 60 and 25 colony forming units per ml (cfu/ml) were obtained at dosages of 25 and 31 mJ.cm-2 respectively. A minimum dosage of 41 mJ.cm-2 would be required to produce a water quality that would comply with the production limit of 10 cfu/ml. Disinfected filtered water from the filter backwash recovery plant would be blended with filtered water from the main treatment process. Maintaining a high enough chlorine residual during the blending process would make UV disinfection a process option to minimise the risk of viable protozoa in the final water from filter backwash recovery plants.