Over the past couple of years, several water suppliers in the U.S. have selected membrane technology for new or upgraded water systems Although each system inevitably is confronted by individual conditions, one water supplier-Plantation, Florida, which relies on wells for its source-faced issues that are familiar to water suppliers around the country. The work of the city and its consultant, Metcalf & Eddy, to address those interconnected issues illustrates the factors that must be considered and accommodated in the selection of a membrane treatment process. Not surprisingly, those factors begin with the water supply and end at the tap.
Looking To UpgradeLike water supply systems across the country, Plantation's East Water Treatment Plant was approaching the end of its useful life. The plant's three Dorr-Olive clarifier/filters, lime-handling system, clearwells, and 0.6 million gallon storage tank were built between 1955 and 1965. A 1.0 million gallon storage tank was added in 1967 and a 3.0 million gallon tank constructed in 1984. Treatment equipment for sludge dewatering, chlorination, hydofluosilicic acid feed, a second lime-handling system, and a 6 mgd Infilco softener and Greenleaf filter were installed in 1977. Although an ammonia handling system was added in 1984, that system, as well as the Dorr-Oliver treatment units and the original lime storage/feed system, had not been used in recent years.
The nominal capacity of the treatment system was 13 MGD. With actual production during the past three years between 2 and 6 mgd, the East Plant was relegated to serving as a peaking facility for Plantation's water system, which centered on the recently-renovated Central Treatment Plant.
The broad issues facing the city were not uncommon: deteriorated equipment, increased service demand, and heightened treatment requirements inflicted by new or anticipated Safe Drinking Water Act (SDWA) controls on pathogens and disinfection/disinfection by-products (D/DBP). Technical choices available ranged from replacing existing equipment with upgraded lime softening to employing membrane softening. Potential upgrades to the lime-softening system included incorporating ozonation, activated carbon, ultraviolet irradiation, or a combination of those. Based on earlier experience with membrane technology at its Central Treatment Plant, Plantation initiated a program to upgrade the East Treatment Plant using membrane technology. A need for flexibility and cost-efficiency fueled the decision to develop all the support facilities that would be required for the full build out of a 12 mgd facility, but to limit initial installation of new membrane treatment units to those required for 6 mgd production capacity initially. Until the new system was operable, the existing lime-softening system and all essential support systems had to remain in service.
The city's utility director, Mel Entus, wanted to use as much of the existing system as possible in the new design. Still, the new reverse osmosis (RO) softening system selected required a new building with approximately 16,000 sq ft of floor space on one level. Constructing this facility required the demolition of the north Dorr-Oliver treatment unit, the washwater recovery basin, the lime sludge lagoon, the sludge thickener, the vacuum filter, and the sludge-dewatering building. The demolition, in turn, required the use of a temporary lime sludge lagoon and backwash recovery basin to maintain the Infilco lime-softening system in service. The 5-mgd Dorr-Oliver unit was converted for sludge storage and the 0.6 million gallon storage tank was assigned to backwash recovery.
Unlike many selection processes where the optimal treatment technology is the result of extensive evaluations of alternative technologies, Plantation's upgrading project began with deciding on a scheme capable of achieving very high treatment levels. Targeting reverse osmosis was not the last step in the process; rather, it was Plantation's first step, based on previous experiences with RO. And, like any first step, the effects and needs triggered by Plantation's decisions extended far beyond the treatment process.
The Starting Point: Water SupplyThree primary water supply issues were identified immediately: raw water quality, adequate quantity, and the withdrawal/transmission system. Each factor could be expected to affect the applicability, design and performance of an RO system.
Jay Ameno, M&E Project Director, explained, "Many people are familiar with the fouling potential of membrane systems and of the effect raw water quality will have on that potential. What is less often recognized, though, is the expected effect of an RO system on withdrawal capacity or the interdependence of the treatment and supply facilities. Assessing the water supply system goes far beyond identifying raw water quality parameters."
In fact, the assessment extended beyond gathering water quality data, both chemical and biological, to include the following: configuration of original wells, materials of construction, pumps, piping, valving, electrical components, geology, hydrogeology, capacity, well efficiency, sand and colloidal material, maintenance history, age and life expectancy of wells, and permitted capacity. After each well was investigated, the wellfield and the raw water transmission system as a whole required evaluation.
Water Quality: The nature of membrane softening triggers particular concern over raw water quality because of the potential for membrane biofouling. Identifying the source and type of any biological contamination also was critical here in assessing the potential remediation of wells or the wellfield. Seven of the ten wells serving the East Plant are located within approximately 50 ft of the East Holloway canal. Under current regulations issued by the Florida Department of Environmental Protection (FDEP), the South Florida Water Management District (SFWMD), and Broward County, these wells would not be permitted because of that proximity. The canal, which acts as a source of recharge, could have a hydraulic connection to the wells, with the possibility of biological contamination from a surface water source. Exacerbating that possibility is the Miami Limestone Formation, a very porous formation through which the canal is cut. Finally, the species of biological contaminant was of critical concern, not only for well protection, but also for the protection of the cartridge filters and membranes.
Supply System: The existing supply system showed signs of deterioration. Some well pumps showed leaking packing glands, which allowed the development of algal, bacteriologic and fungal growths on the outside of the wells. One pump bowl assembly, which had been left on the pad of a well, showed severe corrosion, indicating galvanic incompatibility between the water and well materials. In addition, concerns about water hammer arose. This could possibly cause silting and sanding problems, and even contamination from the nearby canal.
Based on the analysis, the newest existing 1 mgd well was to be modified, and joined by seven new 2.4 mgd wells with 100 hp submersible, stainless steel pumps to provide a minimum pressure in the 50p;60 psi range in the RO feed-pump suction header. The new wells are approximately 150 ft deep, cased for the first 110 ft with a 24-in. diameter, 1-in thick PVC casing. The open hole portion has the ability to have a stainless steel screen installed if required, according to project hydrogeologist G. M. Witt.
Permitting: Groundwater withdrawals in Broward County are controlled by two regulatory agencies: SFWMD, which regulates the quantity of water that can be withdrawn by any individual entity from a given area and issues FDEP permits for well construction, repair and abandonment; and Broward County, which controls the placement of wells and wellfields through the county's Wellfield Protection Ordinance.
The Plantation utility already had a consumptive use permit for its wellfield but several other permitting issues emerged.
First, the city's SFWMD permit, with an expiration date of June 10, 1998, had a 1995 permitted allocation of 4.451 billion gallons per year, with a maximum daily allocation of 16.33 million gallons. For 1998, the allocation shows a 21 percent increase, for an annual withdrawal of 5.631 billion gallons a year. Although this is adequate to meet projected service demands, an RO system would be expected to show a loss of approximately 15 percent of the supply as concentrate, or brine. To obtain the 12 mgd of finished water required by the city from the East Plant, 14 mgd of raw water is needed.
The Back End: Waste DisposalThe use of membrane technology in Plantation depended on the city's ability to dispose of the brine, or concentrate, generated by the membrane process. Concentrate from any RO process is classified by the FDEP and the USEPA as an industrial waste by-product, and its disposal requires an industrial discharge permit from the state. In addition, the concentrate quality must be compatible with the receiving body's water quality standards-and RO concentrate typically contains at least one constituent that exceeds acceptable water quality standards for surface water discharge. For drinking water, the parameters that may exceed acceptable levels include hydrogen sulfide, total dissolved solids, dissolved oxygen, calcium/magnesium carbonates, pH, specific conductance, chlorides, radionuclides, fluoride, and metals.
Plantation considered several options for concentrate disposal. Discharging directly into the ocean was unlikely, so consideration shifted to tapping into an existing ocean outfall. However, the nearest outfall systems are located at least 12 miles away from the treatment plant. Other surface water disposal possibilities included the withdrawal of canal water from the East Holloway Canal, mixing the concentrate with the canal water, and discharging the mixture back into the North New River Canal; or, mixing the concentrate with the canal water and incorporating it into a reuse system.
Deep well injection is a viable alternative when surface water discharges are not feasible due to such conditions as receiving water body acceptability, source availability, or regulatory constraints. Because of its classification as an industrial waste, RO concentrate requires a Class I injection well for disposal, defined by the FDEP as a well used to inject hazardous waste beneath the lower-most formation containing an underground source of drinking water within one quarter mile of the well bore.
A Class I well requires certain subsurface conditions, including a highly permeable injection zone that has been classified by the FDEP as suitable for waste injection. In southeast Florida, the Lower Oldsmar Formation, or Boulder Zone, has been used successfully as an injection zone. To protect the potable and potentially potable water supplies of the Biscayne and Floridan aquifers, brine concentrate disposal wells have to be constructed with approximately seven strings of casing in various diameters. Each casing string must be completely cemented in place, with the exception of the pit casing and the monitor zones of the intermediate casing, using API Class H cement and lost circulation additives to ensure a good bond. The final selection of casing depths, monitor zones, and injection zones depends on information collected during the drilling, sampling, and testing of a mandatory pilot hole.
Although Plantation was using an RO concentrate injection well at the Central Water Treatment Plant, the geology at that site differed significantly from the geology at the East Plant's proposed injection well site -just as the raw water feeding the two plants differed.
The Centerpiece: TreatmentThe East Treatment Plant's existing processes included lime precipitation, clarification, gravity filtration, and chlorine disinfection. Raw water was extracted from the Biscayne aquifer. Three of the clarifier/filter units, representing about half the 13 mgd capacity of the plant, had reached the end of their useful lives. At the very least, they required major renovation or replacement. Even with that type of response, however, the resulting system would have been incapable of meeting the SDWA rules for D/DBP and organics control. To meet the requirements the existing lime-softening system would have to be upgraded. Candidate technologies to accomplish this included ozonation, activated carbon, ultraviolet irradiation, or combinations such as ozonation/GAC.
Because of the city's successful experience with the Central Plant, which had been upgraded with a membrane system, Plantation's interest in using similar technology at the East Plant was understandable. Yet, different considerations affected selection of the appropriate solution for each facility, precluding a "twin" design.
Three major differences deserve mention. First, the Central plant had an existing wellfield designed around lime softening, whereas the East plant has a wellfield developed specifically for membrane processing. Second, lime facilities at the East plant were in good condition, allowing conversion from quicklime to hydrated lime for pH and alkalinity control. This change enables the East plant to produce a stable finished water, reducing the need for much additional treatment for scale prevention or corrosion control. Third, a wider selection of membranes is available now than at the time the Central Plant was upgraded with a membrane system.
Despite these broad differences, there was considerable interest in maintaining reasonable uniformity between the membrane softening systems at the two plants, which led to Metcalf & Eddy's development of an upgrade design for the East Plant that shared as many characteristics as possible with the Central facility. Among these were the size and capacity of the membrane modules (standard 8-in. diameter by 40-in. spiral wound membrane modules with 360 sq ft or more of effective surface area); the pretreatment equipment (depth-type, plain-end, 40 in. polyethylene cartridge filter elements in industrial style housings located ahead of the membrane softening arrays); and the Allen-Bradley programmable logic controller hardware and software.
The similarities between the two systems were accompanied by significant differences, based on the different design considerations of the two plants. For example:
- The East Plant was to be equipped with 2 mgd arrays instead of the 3 mgd arrays in the Central Plant. Normal maintenance activities (such as membrane cleaning) remove a smaller fraction of the plant's capacity from service with the smaller arrays in place.
- Side-entry pressure vessels were used at the East Plant rather than the end-entry vessels installed at Central. Side-entry, fiberglass-reinforced, thermosetting resin pressure vessels were not commercially available when the Central Plant was upgraded. They are easier to maintain.
- The arrays at the East Plant are top-fed to maintain a wetted condition for the membranes during idle periods.
- Single-stage, horizontal split-case centrifugal pumps are used at the East Plant instead of the multi-stage vertical turbine type installed at the Central Plant. The relatively low head requirements for softening type membranes can be satisfied by standard ANSI design, split case, or API pumps that have lower capital and operational costs, and that are readily available in all-stainless construction. The vertical turbine units used at Central were designed for the higher head requirements of cellulose acetate and brackish water thin-film composite membranes, and to accommodate a lime-softening process with a low-head requirement. The East plant, on the other hand, was designed to provide a higher head, which was needed for membrane applications.
- In-line static mixers are employed in the new East Plant process scheme to ensure uniform dispersion of chemical additives. The other plant depends on the natural turbulence created in the piping and pumping components. This has a slow response and corresponding difficulty in controlling the process.
- Raw well water is used at the East Plant instead of membrane-softened water for flushing membranes on shutdowns. The membrane array is equipped with special control valves to eliminate the need for a separate flush-water system.
- The standby generator installation at East Plant qualifies the facility for a load-control rate from the electric power utility. Application of such a rate can reduce power costs by about 15 percent. Anticipated savings are $30,000 to $40,000 per year.
- The upgraded treatment scheme (see diagram) begins with an acid/antiscalent addition and proceeds to cartridge filtration, membrane softening, sequesterant addition, and forced draft degas. The degassified water is transported to the existing clearwells for blending with hydrated lime for pH control, fluoride, and chlorine for disinfection.
Site Specific ApplicationsEven small differences in water quality or existing design parameters can have a profound effect on the selection of appropriate treatment technology. Membranes come in various configurations, materials of construction, and removal efficiencies, and designing an effective system is a complex process of addressing issues that range from raw water quality and volume to finished water taste. Although the two water treatment processes at Plantation share several components and characteristics, they are different.
Just a few miles away, in Hollywood, Florida, a completely different system, also designed by Eddy, was built. That city's original water supply, like that of Plantation, was the Biscayne aquifer, but a growing population demanded additional supplies. Because of the withdrawal limits imposed on Southeast Florida water users, the city was forced to seek out new water supplies to supplement its existing allotment from the Biscayne Aquifer. The Floridan Aquifer, a deeper and poorer-quality source, proved adequate in quantity. Since the city's existing lime-softening equipment had reached the end of its expected lifespan and was incapable of meeting upcoming SDWA regulations, Hollywood set about developing a new system.
The design response in that community was vastly different from that in Plantation, involving two separate treatment trains to augment the existing lime-softening system. The first new section used nanofiltration, while the second train (reserved for use on the poor-quality Floridan Aquifer water) employed reverse osmosis. The more "obvious" design response to the two-source raw water supply would have been to combine the flow from the two sources and subject it to one treatment process. In the case of the combined Biscayne and Floridan water, that single treatment process would have been reverse osmosis, which is the most intensive method-and usually the most expensive. Instead, the designers chose to treat each separate raw water supply according to its specific needs, and then to blend the finished water prior to distribution for a uniform taste and odor. The innovative, parallel treatment response produced considerable savings for the client and ensured that the finished water would be virtually indistinguishable from that previously provided by the city's lime-softening system.
Because membrane technology is relatively new in the United States, especially for drinking water applications other than desalination, many water suppliers are unfamiliar with the intricacies of membrane design and application. That unfamiliarity is counteracted by an understandable interest in the technology, based on its ability to address such critical drinking water issues as pathogen and D/DBP control. It deserves note that the technology's promise as the treatment technology of choice for many water utilities must be balanced with a recognition that its application is site-specific, dependent on variables that include raw water quality, required treatment levels, existing infrastructure, disposal methods, and cost.
In fact, "membrane technology" is not a single product or system, but a category of systems, processes, materials and products. The four primary types of membrane technology-ultrafiltration, microfiltration, nanofiltration and reverse-osmosis (also called hyperfiltration)-are available in a variety of configurations and materials of construction. Any specific membrane type may or may not require pretreated water. The selection of the wrong type, configuration, or material for any individual application is likely to trigger problems ranging from frequent fouling to unacceptable removal efficiencies.
It is because of these variables that membrane technologies-perhaps more than treatment techniques with which we in the U.S. are more familiar-virtually demand front-end testing in bench-scale, pilot-scale, or both. Without doubt, membranes represent a significant advance in drinking water technology and are likely to play a growing role in the U.S. water industry. When properly selected, designed and installed, they outperform many conventional processes and virtually eliminate the potential for pathogen breakthrough. But, as with any technology employed in the unforgiving business of providing safe drinking water-and as those of us in the industry realize on a day-to-day basis -the success of any individual application rests as much on the capabilities of the designers as on the structure of the equipment and facilities: the "thinking" is as critical as the "constructing."