Evaluation and quantification of engineered flocs and drinking water treatability

The provision of safe drinking water requires adequate inactivation of pathogenic organisms. Common drinking water disinfection technologies become decreasingly effective as levels of additional particulate matter in waters to be treated increases.;Using protocols and raw water from the Mannheim Water Treatment Plant (MWTP) in Kitchener, Ontario, Canada, twelve (12) jar tests were performed throughout this investigation. The ultraviolet absorption at 254 nm (UV 254) and turbidity of the supernatant were evaluated during each jar test to investigate potential relationships between these parameters and floc settling rates and structure. Six (6) jar tests were conducted to generate aggregates for settling tests. Samples were collected after a period of settling from the bottom of the jars so that it could be determined whether or not the settling rates and/or sizes of the aggregates that had settled would correspond to the UV254 and turbidity of the supernatant. The six (6) jar tests were then repeated to characterize the fractal structure of the floes by digital image analysis with microscopy. Structural characteristics were calculated from samples of aggregates that were collected prior to settling. This was done to predict particle removal performance (i.e., based on turbidity reduction) by using the floc structural information of the aggregates generated during coagulation and flocculation. Samples of aggregates generated at full-scale at the MWTP were then collected and compared to the results of the bench-scale testing. This analysis was conducted to determine the extent to which bench-scale tests truly simulate full-scale coagulation and flocculation processes.;At the conditions investigated, either alum or polyaluminum chloride (PACT) coagulation at a dose of ∼ 30 mg/L in conjunction with 0.2 mg/L of cationic polyelectrolyte can achieve the lowest levels of UV254 and turbidity (i.e., 0.02 to 0.05 AU and 0.3 to 1.0 NTU, respectively) after flocculation and a period of settling. Results from the settling tests indicated that the production of larger and more settleable flocs could not be described by floc settling velocities and floc sizes. Settling velocities were not directly related to either UV254 or turbidity reductions. Results of the floc characterization tests indicated that measured values of UV254 and turbidity of the supernatant were generally inversely proportional to aggregate D90; that is, the residual UV254 and/or turbidity decreased as the value of D90 increased, which may have been indicative of flocculent settling. No direct relationship could be discerned between D1 (i.e., floc shape) and the UV254 and turbidity of the supernatant; however, the turbidity after flocculation and a period of settling appeared to be inversely proportional to D2 (i.e., porosity). Overall, the results of the experiments have demonstrated that grain size distributions and fractal dimensions might be used to assess and/or predict pre-treatment and/or particle removal performance. Specifically, the relationship between D90 values calculated from samples of flocculated water prior to settling and UV254 and turbidity values of that water after a period of settling may be a simple tool that can be utilized to describe and potentially better predict flocculent settling performance. At present, this appears to be the first such tool of its kind that has been reported.;Full-scale sampling at the MWTP, however, indicated that the size and structure of aggregates generated at bench-scale at the MWTP were clearly not indicative of the size and structure of those produced at full-scale. The aggregates that were generated at full-scale were much smaller and denser than those that were produced in any of the tests (i.e., under all conditions considered). At the MWTP at present, the only reliable indicator of full-scale performance is full-scale data because jar tests are not indicative of full-scale performance (i.e., floc formation). Additional experimentation at the MWTP is required that focuses primarily on optimizing the bench-scale tests utilized to improve full-scale performance predictability. (Abstract shortened by UMI.);Jar tests are performed to simulate full-scale pre-treatment and particle removal processes. Operators typically conduct them in an effort to attempt alternative treatment doses and strategies without altering the performance of the full-scale drinking water treatment plant. However, information obtained from these tests must be evaluated judiciously, as they currently focus on reduction of specific water quality parameters (i.e., ultraviolet absorption at 254 nm and turbidity), and measuring and understanding the significance of coagulant dose on floc size. Consideration of aggregate structure has been less explored due mainly to a lack of appropriate theories to describe the complex random floc structure. Improving the predictive capacity of bench-scale protocols commonly used for optimizing conventional chemical pre-treatment in full-scale drinking water treatment plants is required.