| Watertown, South Dakota currently uses conventional softening to treat ground water to remove hardness, iron and manganese, and natural organic matter. The distribution system has experienced elevated concentrations of disinfection by-products, which will require additional treatment to comply with future limits set by the Stage 2 Disinfectant/Disinfection By-Products Rule. A nanofiltration membrane system has been proposed to remove additional natural organic matter from the softened water. Removing the natural organic material disinfection by-product precursors should reduce the concentration of disinfection by-products. The use of nanofiltration membranes has been previously documented to reject total organic carbon, while allowing desirable levels of hardness and alkalinity to remain in the water.A study using a 4" diameter, single element pilot plant study compared six nanofiltration membranes from various manufacturers. Koch TFC-SR2, Koch TFC-SR3, Trisep XN45-TSF, Hydranautics ESNA1-LF, Hydranautics HydraCoRe-70pHT, and Dow/Filmtec NF270 were operated at 15, 50, and 80 percent recovery to analyze each membrane's performance. Water analyses were performed and membranes were selected for further studies if they exhibited high total organic carbon rejection, low simulated distribution system disinfection by-product formation, and reject stream total dissolved solids concentrations below 1000 mg/L. The Koch TFC-SR3, Trisep XN45-TSF, and Hydranautics ESNA1-LF membranes were selected for further study. A larger pilot plant was utilized to further test the selected membranes using a 2:2:1:1 array of twenty-one 4" diameter elements. Three membrane models were operated at flux rates of 9, 12, and 15 gallons per day per square foot to determine an optimum flux setting for full scale design. Water quality parameters were also analyzed to provide additional basis for membrane selection. The efficacy of nanofiltration to reduce disinfection by-products was also compared to chloramination.Koch TFC-SR3, Trisep XN45-TSF, and Hydranautics ESNA1-LF exhibited total organic carbon rejections in excess of 90 percent for every test setting during final testing. The excellent organic rejections corresponded to disinfection by-product removals in excess of 90 percent for many of the flux settings. No trends were established between flux rate and disinfection by-product removal. However, increasing flux rates were shown to cause decreasing specific flux rates. A membrane fouling study was performed for 30 days to analyze the specific flux decline from organic and inorganic foulants. No significant fouling was observed for Hydranautics ESNA1-LF or Trisep XN45-TSF. Dependent on several economical factors, an optimum operating setting was not established. |
