The city of Kennewick is located in southeastern Washington, and its primary drinking water sources are two Ranney Collector Wells that are near the Columbia River.
In 1979, a water filtration plant treating the Columbia River source was built to augment the Ranney Wells in the months when demand was highest, typically March through October. The initial capacity of the treatment plant was 7.5 mgd, but it was built with the idea that over the years its capacity would be expanded in 7.5 mgd increments up to 30 mgd.
By 2002, it became clear that it was time to plan an expansion of the water treatment plant.
Kennewick selected HDR to undertake a study of how best to expand the water treatment plant, which treated Columbia River water using ozone, rapid mix, flocculation, sedimentation, media filtration and chlorine disinfection. Not only did they want to expand the capacity from 7.5 to 15 mgd, they also wanted to be able to operate the plant year-round, relying less on the wells in future years. Of course, the treated water had to meet both current and anticipated Safe Drinking Water Act (SDWA) regulations.
HDR investigated three options:
Submerged membrane retrofit
The third option was chosen, partly on the basis of cost—no new buildings would be necessary—and it was referred to as the “no concrete” option.
However, this solution had other strong advantages as well. It provided the greatest flexibility to meet future regulations; it could be easily automated; and it had the shortest construction time.
The construction time issue was critical because the work would have to be done in the winter months when the plant was not in use.
The submerged membrane retrofit approach required a change in treatment technology at the plant, a conversion from ozonation/flocculation/sedimentation/filtration to flocculation/sedimentation/microfiltration.
The key technology change involved eliminating the ozone contact basin and equipment, and the addition of low-pressure membrane microfiltration equipment. The following changes were necessary:
Selecting a membrane vendor
Two vendors that supply submerged membrane systems for drinking water treatment were investigated—Zenon Environmental and USFilter/Memcor.
Both vendors provide hollow-membrane modules in a low- or high-density fiber packing configuration.
For this application, a high membrane-fiber packing density was preferred to maximize the amount of membrane surface area that could be installed in the existing filter basins.
Because each membrane system is unique, it is necessary to design around a specific system. The city of Kennewick pre-selected the membrane vendor based on head-to-head pilot tests, procured the membrane system, and then completed the water treatment plant improvement design.
Side-by-side pilot testing with Zenon and USFilter/Memcor systems was done at the water treatment plant between April and July 2003.
Pretreatment with an inclined plate settler was provided upstream of the membrane pilot units. Both vendors successfully demonstrated performance for the first and second clean-in-place demonstration periods. Based on the flux rates demonstrated during the pilot testing, it became clear there was more than adequate space in each of the four filter basins to provide 15 mgd of capacity.
Both vendors submitted bids. Equipment capital costs for the two bids for a capacity of 15 mgd were very close, $.20 per gallon for USFilter and $.22 per gallon for Zenon. Both units met performance goals, but based on 20-year lifecycle costs, USFilter was the low bidder and was chosen for this project.
The challenge was to implement the changes with a limited budget and within a very aggressive schedule to bring the plant online in 2005.
There were also some physical challenges—the depth of the existing filter basins had to accommodate the height of the membrane units, and the existing headroom in the building had to accommodate the lift height of the membrane units plus a new bridge crane for installation and removal of the membrane units.
There was limited space in the pipe gallery of the filter building, and the permeate pumps had to be located adjacent to the membrane tanks in the gallery.
Finally, the backwash pumps had to be located on the outside of the filter building.
Also, although not necessary for near-term operations, because the pilot testing showed that the plant membrane system could easily be expanded from 15 to 20 mgd by adding more membrane modules, Kennewick elected to install the larger capacity permeate pumps, backwash pumps, and air and permeate pipe headers—which would be able to accommodate this future expansion—as part of this construction project.
Because of the limited construction window, the design and construction of this project was split into two phases.
Phase 1 involved the membrane system installation, which was to be completed in April 2005. At that point, the facility became a membrane filtration plant. The pretreatment upgrade will take place between November 2005 and April 2006.
Phase 2 involves upgrading the intake and finished water pump stations, modifying the clearwell basin, and remodeling the operation and chemical building. This work will be done between November 2006 and April 2007. At that point, the facility will be operated as a year-round 15 mgd membrane filtration plant.
Careful design and construction coordination is mandatory for a successful retrofit, but there are several good reasons for utilities to consider retrofitting when they need to upgrade their facilities or expand capacity.
The cost of membrane systems are coming down. Retrofitting avoids the cost of building new structures, which can save time as well. The ability to expand again at a later date in a short time without a major expense, simply by adding new modules to the membrane basins, provides flexibility.
Based on Kennewick’s experience of doubling the water treatment capacity in the same footprint, other utilities should feel encouraged to consider membrane retrofits.