John E. Tobiason
The Trap Falls Water Treatment Plant (TFWTP) is a 25 million gallons per day (MGD) conventional drinking water treatment facility in Shelton, CT. Bench-scale experiments were performed to evaluate a potential advanced treatment upgrade of post-filtration water at the TFWTP, prior to chlorination and distribution. Two advanced treatment options were evaluated, granular activated carbon (GAC) and membrane filtration. The goal of advanced treatment was to remove natural organic matter (NOM), which would lead to a decrease in disinfection by-product (DBP) formation. DBPs are formed when NOM is oxidized by chlorine, which is used by TFWTP for primary disinfection and to maintaiu a disinfectant residual throughout the distribution system. Two classes ofDBPs are regulated, trihalomethanes (THMs) and haloacetic acids (HAAs) which have maximum contaminant levels of 80 and 60 )..Ig/L, respectively. To account for seasonal fluctuations, treatment of sununer and winter post filtration waters was evaluated. Bench-scale testing of GAC via rapid small-scale column tests (RSSCT) was used to estimate activated carbon treatment using far less time and water as compared to fullscale or pilot testing. RSSCTs simulating both 10 minute and 20 minute empty bed contact times (EBCT) were conducted. GAC treatment effectiveness was evaluated by monitoring dissolved organic carbon (DOC), ultraviolet absorbance at the wavelength of 254 urn (UV254), and potential formation ofTHMs, and HAAs. Rapid bench-scale membrane tests (RBSMT) were used to evaluate nanofiltration membrane treatment using far less time and water as compared to full-scale or pilot testing, and to screen several potential membranes. Two nanofitration membranes were evaluated: Dow's NF270, and Hydranautic's ESNA1. Nanofiltration effectiveness was evaluated by monitoring DOC, UV254, conductivity, calcium, and the potential to form THMs, and HAAs. DOC breakthrough to 1 mg/L occurred at approximately 15,000 to 18,000 bed volumes in the RSSCT tests. An effluent DOC of 1 mg/L yields THM and HAA formation of approximately 30 and 25 J.lg/L at a 1.8 mglL dose of chlorine, well below the 80/60 J.lg/L MCLs. This effluent goal, corresponding to approximately 50% DOC breakthrough, was used to design a full-scale GAC treatment system. The RBSMT results showed rejection of90 to 70% of NOM at recoveries of30 to 90%. Of the two membranes tested, Dow's NF270 membrane performed better than Hydranautic's ESNAI membrane in terms of DOC and UV254 rejection. Both membranes performed similarly in terms of conductivity and calcium rejection. A full-scale design and cost analysis was performed based on a 1 mglL effluent DOC goal. The GAC contactor design was based on a 10 minute EBCT, with 8 full-scale contactors in parallel. For the membrane system, approximately 67% of the TFWTP effluent would be treated by nanofiltration. Results of the cost analysis showed that GAC is a more economical option, costing approximately $0.23 per 1000 gallons, as compared to $0.45 per 1000 gallons for membrane filtration.