Evaluating the effects of the flow direction on the performance of the rapid sand filter

Abstract

The rapid sand downflow filter is widely used in water treatment plants. On the other hand, this filter has some drawbacks included the significant development of the head loss via the filter media because most of the rejected particles are removed by the upper layers. As well, the filter particles are redistributed during the backwash process causing the settling of fine particles on the upper part of the filter media, and this needs to increase the number of backwash processes. For these reasons, the cost of the produced water increases. The aim of the present study is to explore the possibility of using the upflow sand filter (UF-Filter) as a good alternative to the downflow sand filter (DF-Filter). To achieve the aims of the present study, a comparison was made between the performance of both filters through simultaneous experiments under different operating conditions. These conditions included changing of the filtering velocity from 5 m/h to 10 m/h and the initial water turbidity with a range of (10 - 200) NTU. The sand media with sizes of (0.6 - 1mm) and with 63 cm of depth was used. Experimental results show that the turbidity removal efficiency of the DF-Filter is of about 1.1 times that of the UF-Filter. On the other hand, the UF-Filter has higher turbidity removal efficiency than the DF-Filter by about 1.1 times when the initial turbidity of the influent water is greater than 150 NTU and the filtration velocity is equal to 10 m/hr. These differences in the removal efficiency between both filters can be considered as few values. The average filtration efficiency of the UF-Filter operated with the filtration velocity of 5 and 7.5 m/h is higher than that of the DF-Filter operated with the filtration velocity of 7.5 and 10 m/h, respectively under the same operating conditions. The filtration efficiency of both filters increases when the backwash was carried out before each experimental process instead of replacing the filter media. Also, the head loss of the DF-Filter is significantly increased due to redistribution of the sand media taking place during the backwash cycle, while the head loss of the UF-Filter is not affected. The head loss of the UF-filter at the end of each experimental run is less than that of the DF-Filter by about (18.18 % - 45.31 %) when the filter medium is replaced and this range is increased to about (53.31 % - 62.34 %) when the backwash is performed prior to the start of the experimental work. Thus, the decrease in head loss leads to an increase in the filter running time and decrease the number of backwash process.