Advisor
John E. Tobiason
Publication Date
2-2005
Abstract
The Aquarion Water Company (AWe) has funded several research studies. Performed by graduate students at the University of Massachusetts at Amherst, including this research on the Stamford, Connecticut water distribution system (WDS). Some of the recent research has been performed in response to the proposed Stage II Disinfectants/Disinfection By-Product (DBP) Rule (S2DDBPR) (US EPA, 2003). Due to the new requirements that fall under the pmposed S2DDBPR, an initial distribution system evaluation (IDSE) on the Stamford WDS is required to locate potential monitoring sites for compliance with DSP regulations. Between July 2002 and August 2003, three field study events (summer 2002, winter 2003, and sunnner 2003) were conducted on the Stamford WDS to calculate water age at different locations, as well as pmvide a better understanding of water quality (WQ) including temporal and spatial variations in DBP concentrations in the Stamford WDS. A computer model of the Stamford WDS was created by Haestad Methods. The field data and operational data supplied to the University of Massachusetts at Amherst allowed for the verification and application of this model. The model predictions of tracer concentrations and chlorine residual were similar to a majority of the field results from the three sampling events. Water demands during the three sampling events were calculated from operational data supplied by the AWC and used as the water demands in the model. Field tracer studies for all three sampling events were modeled by applying the step up and step down concentrations measured in the Stamford water treatment plant (SWTP) clearwell effluent. Field measured free chlorine residuals were modeled by applying the free chlorine residual measured in the SWTP clearwell effluent, and a calculated first order bulk water decay rate based upon field measurements. The model predictions of tracer concentrations and chlorine residuals were compared to field measurements of fluoride concentrations and free chlorine residuals at locations in the Stamford WDS. Mean hydraulic residence times (HR.1) were calculated via F curves of the tracer concentration data The model over predicted the mean HRT at the locations by an average of 33% during the summer 2002 event. During the summer 2002 event, the difference between model predictions and field calculations of mean HRTs ranged from - 38% to 80%. The model over predicted the mean HRT at the locations by an average of 11 % during the winter 2003 event. During the winter 2003 event, the difference between model predictions and field calculations of mean HRTs ranged from -58% to 54%. The model over predicted the mean HRT at the locations by an average of 15% during the summer 2003 event. During the summer 2003 event, the difference between model predictions and field calculations of mean HRTs ranged from -33% to 120%. On average, the model over predicted the mean HRT at locations in the Stamford WDS during all three events. The model most closely predicted the mean HRTs at locations in the Stamford WDS during the winter 2003 event. The model predictions of chlorine residuals were compared to field measurements of free chlorine residuals at locations in the Stamford WDS. The model over llIedicted the chlorine residual at the locations by an average of 9.8% during the summer 2002 event. During the summer 2002 event, the difference between model predictions and field calculations of chlorine residual ranged from -63% to 127"/0. The model over predicted the chlorine residual at the locations by an average of 15% during the winter 2003 event: During the winter 2003 event, the difference between model predictions and field calculations of chlorine residual ranged from -10% to 40%. The model over predicted the chlorine residual at the locations by an average of 49% during the summer 2003 event. During the summer 2003 event, the difference between model predictions and field calculations of chlorine residual ranged from -60% to 450%. On average, the model over predicted the chlorine residual at locations in the Stamford WDS during all three events. The model most closely predicted the chlorine residual at locations in the Stamford WDS during the summer 2002 event. A correlation between the free chlorine residuals and DBP formation was established. Applying this relationship to the model predictions of the chlorine residuals allowed for estimation of relative DBP concentrations in the Stamford WDS. The reccommendations for the IDSE monitoring sites by Kansas (2004) were confirmed based on the model predictions of water age and the estimations of relative DBP concentrations based on the model pnidictions of chlorine residuals.
DOI
https://doi.org/10.7275/gsfc-nn23