Challenges of Selecting NF Membranes for Removal of DBPs Precursors from Softened Groundwater: Watertown Utilities Experience
选择NF膜去除软化地下水中DBPs前体的挑战:Watertown公用事业公司的经验
The majority of the drinking water provided to Watertown, South Dakota, is treated at the 8
million gallons per day (MGD) Town Water Treatment Plant (WTP), which employs a
conventional groundwater treatment system, comprised of aeration, lime softening, recarbonation,
media filtration and disinfection. The plant has experienced a few occurrences of elevated levels
of disinfection byproducts (DBPs) in the distributed water, with concentrations of Total
Trihalomethanes (TTHM) up to 120 µg/L. A nanofiltration membrane system is proposed as
enhanced treatment for the existing softened water. The purpose of the NF membrane treatment
system is to further treat a stream of the softened water in order to reduce the concentration of
disinfection byproduct (DBP) precursors. The NF permeate will be blended with the remaining softened water, which is not
routed through the NF system, as well as with additional 4 MGD from a new conventional
groundwater softening plant. Final blend will represent a total of 12 MGD of drinking water
meeting Stage 1 and Stage 2 of D/DBP Rules.
One of the main challenges of this project is selecting the NF membrane which can provide the
most economical solution while satisfying several competing goals: high TOC rejection to reduce
the level of DBP precursors; low calcium and alkalinity removal to avoid the need for final blend
stabilization treatment; and, low inorganic rejection in the concentrate water for cost-effective local
disposal. Selection of the NF membrane is performed based on the results of a pilot study program
composed of three segments. First, a membrane screening study was performed with a single 4"
element pilot unit. This allowed screening of six candidate NF membranes in a quick but effective
manner. The screening criteria were high TOC removal, low simulated distribution system (SDS)-
DBP formation in the product water and in the blend, and low total dissolved solids (TDS) passing
rates. In no preferred order, the selected membranes were Koch TFC-SR3, Trisep XN45, and
Hydranautics ESNA1-LF.
These membranes underwent a second phase of testing which was performed with a two stage
pilot unit, modeling the full scale plant design, and which was just completed. One membrane will
be selected based on operating performance parameters as well as quality of product water (TOC,
SDS-DBP, TDS, alkalinity and calcium). A third and last phase of the piloting will focus on
fouling and cleaning requirements and will begin in January 2009. The results of this pilot study
will not only allow selection of the most suitable NF membrane for this application but will also
provide a set of hydraulic parameters for the design of the full scale NF membrane plant. This
paper focuses on the results and the selection process for the first two phases of the piloting
program. Includes tables, figures.