Comparison of two online flocculation monitoring techniques for predicting turbidity removal by granular media filtration.

Particulate matter removal in drinking water treatment via direct granular filtration requires specific flocculation conditions (a process typically termed 'high energy flocculation'). Predicting filtered water turbidity based on flocculated water characteristics remains difficult. This study has sought to establish a relationship between filtered water turbidity and the flocculated water characteristics. Flocculation oflow-turbidity raw water was evaluated online using a Photometric Dispersion Analyser (PDA) and a Dynamic Particle Analyser in a modified jar test followed by a bench-scale anthracite filter. Coagulants used were alum, PASS100 and ferric sulphate, in addition to a polydiallyldimethylammonium chloride (polyDADMAC) cationic polymer. They were dosed in warm and cold waters, and flocculated with intensities (G) from 0 to 100 s(-1). Of the two instruments selected to analyse flocculation performance, the Dynamic Particle Analyser was shown to be the most sensitive, detecting small changes in floc growth kinetics and even floc growth under low flocculation conditions which remained undetected by the PDA. Floc size was shown to be insufficient in predicting particulate matter removal by direct granular filtration as measured by turbidity, although a threshold d(v) value (50 microm) could be identified for the test conditions evaluated in this project, above which turbidity was systematically lower than 0.2 NTU.