An investigation into the influence of sand size, bed depth, rates of filtration and temperature on the quality of filtrate from a slow sand filter

• Main Author: Aydin, Mehmet Emin
• Published: Loughborough University 1993
• Subjects: 621.69; Water treatment by sand filtration
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.334368

The objective of this research project was to investigate the influence of the principal design and operational parameters on the quality of the filtrate from a slow sand filter. These parameters are sand effectivc size, depth of sand bed, rate of filtration, water temperature and influent quality. Attempts were then made to establish relationships between the variables employed by means of regression analyses of the experimental data. To carry out the investigation three laboratory scale filter columns were constructed each measuring 150 mm in diameter and 3010 mm height. The filtration medium in each column consisted of a 1.2 m depth of sand on a 0.3 m depth of gravel. For the three individual columns three different sand sizes were employed. These were 0.17 mm effective size (ES), 0.35 mm ES and 0.45 mm ES. Water to be filtered was abstracted from the Burleigh Brook, adjacent to the Civil Engineering laboratories. As required small amounts of settled sewage, obtained from the Loughborough Water Reclamation Works, were added to increase both the turbidity and the count of indicator bacteria. The filters were operated at five different filtration rates at three different temperatures. These temperatures were 25 degrees C, 15 degrees C, 5 degrees C and the filtration rates repeated for each temperature were 0.1, 0.2, 0.3, 0.4, 0.5 m/h. The filters were operated for not less than one month at each filtration rate. In order to assess the efficiency of the filters water samples were taken and analysed from the influent and from the filtrate and also from a series of sampling taps positioned at various depth down the sand media. These water samples were tested for total coliform bacteria, for faecal coliform bacteria, suspended solids and turbidity. Regression analyses were then carried out on the data obtained and regression models were developped for 100 mm and 1200 mm sand depths of each filter for each period. Occasional tests (i. e. at least once for each flow rate) were also carried out for nitrate and ammoniacal nitrogen, colour, pH, conductivity, total organic carbon (TOC) and dissolved oxygen. Following the completion of the filtration operation at three temperatures and five filtration rates the sands of the filtration media were also examined. Sand samples were taken for examination from 0, 50, 100, 150, 200, 300, 400, 600, 800, 1000 and 1200 mm below the sand surface of each filter. Then in order to assess both the penetration of the removed solids into the filter bed and to investigate the concentration of the biological film developed within the sand bed, sand samples were analysed for suspended solids, turbidity, standard plate count bacteria and particulate organic carbon. Regression analyses were also carried out on these data and variation of each parameter against each filter bed was modelled. The solid and biological deposits on the sand samples were also viewed with a Scanning Electron Microscope.