Advanced primary treatment: a positive alternative for new York's Owls Head Plant.
New York City's Owls Head Water Pollution Control Plant faced problems common to treatment facilities in many metropolitan communities. Placed in operation in 1952, the facility did not meet the 1972 Clean Water Act (CWA) requirements for secondary treatment. A major plant upgrading was undertaken to meet CWA requirements. While new secondary treatment facilities were being constructed, the New York State Department of Environmental Conservation (NYSDEC) mandated that primary suspended solids (SS) and biochemical oxygen demand (BOD) removals be enhanced through the addition of coagulants.
Because the chemistry of coagulation is complex, the type and dosage of coagulants used must be scrutinized carefully. Factors ranging from wastewater pH to oxidation-reduction potential can affect coagulation. At the same time, specific coagulants can affect existing wastewater treatment systems and equipment, including sludge management. To identify the optimal type and dosage of coagulant, and to establish achievable effluent quality permit requirements, a full-scale coagulant testing program was undertaken at the Owls Head WPCP. The resulting data, collected over the span of one year, formed the basis of the plant's advanced primary treatment during upgrading of the treatment facility.
Advanced primary treatment uses metal salts, such as ferric chloride or alum, in conjunction with newly developed polymers to coagulate raw wastewater solids. The basis of the technology is coagulation, which is the process by which small suspended particles are combined into larger aggregates. These aggregates then settle, allowing their removal from the raw wastewater.
In raw wastewater, suspended colloidal particles carry a like charge, which causes adjacent particles to repel each other and prevents them from agglomerating, or combining into larger masses. In nature, the charge carried by the suspended colloidal particles usually is negative. Chemical coagulants destabilize the charged particles by reducing the electrostatic charge, bridging the particles, and physically enmeshing the fine solids in gelatinous precipitates. The destabilized particles agglomerate into larger masses, called flocs.
Coagulation may involve different types of coagulant that trigger different chemical reactions. Metal salt coagulants, such as ferric chloride and alum, react with the alkalinity in water to form insoluble hydroxide complexes that are positively charged. These positively charged hydroxide complexes destabilize the negatively charged suspended colloidal particles, allowing the particles to precipitate, or grow larger, and form flocs. Polymer coagulants (polyelectrolytes) are long-chained organic molecules that can become highly charged when dissolved in water. The polymer molecules form bridges between the floc particles, which strengthen and enlarge the metal-salt floc. Polymers can be anionic (negatively charged), cationic (positively charged), or non-ionic (very low-charge density). Although polymers can be used as primary coagulants, more typically they are used in conjunction with metal salt coagulants.
Coagulation can be affected by a number of factors, including wastewater pH, alkalinity, oxidation-reduction potential, and the presence of dissolved inorganic compounds. These water quality conditions will determine both the type and dosage of coagulant required to achieve established effluent water quality limits.
Another factor in determining the type and dosage of coagulant used is the effect of the coagulants on other wastewater treatment operations, such as sludge processing.
Owls Head WPCP
The Owls Head Water Pollution Control Plant is located in the Bay Ridge section of Brooklyn and serves a design population of more than three-quarters of a million people. Brought on line in 1952, the secondary treatment plant was designed with a modified aeration process to remove 60 percent of BOD and 70 percent of SS. The CWA requirements for secondary treatment, promulgated long after the plant was designed and brought on line, call for 85 percent removal of both BOD and SS.
In 1978, Metcalf & Eddy was retained to develop plans for upgrading the facility. The resulting upgrading program includes the development of new facilities--a sludge processing complex, primary and secondary settling tanks, step aeration tanks, outfall sewer, and chlorination facilities--as well as the renovation of such facilities as the pump and power house, dock, and seawall. The scope of the upgrading project, which ultimately would cost more than $350 million, required an eight-phase construction schedule that spans twelve years and involves more than 40 construction contracts.
Before the new facilities could be constructed, it was necessary to demolish the existing secondary treatment plant, leaving the Owls Head facility with only primary treatment capabilities. The primary treatment system could normally achieve removal rates of 15 to 20 percent BOD and 25 to 40 percent SS. To boost those removal rates while the secondary treatment facilities were under construction, state and federal officials required the addition of coagulants to the primary treatment process. Before the proper type and dosage of coagulant could be determined, a plant-scale coagulant testing program was initiated.
A Plant Scale Test
The Owls Head plant-scale test was based on preliminary jar tests that identified metal salts and polymers as the most promising coagulants. The metal salts (ferric chloride, aluminum chloride, and aluminum chlorohydrate) were tested alone and in combination with the polymers (cationic alone, cationic combined with metal salts, and anionic combined with metal salts). In addition to the type and dosage of coagulants, the effect of varying the polymer feed location was evaluated. Polymers were fed at the primary influent gates, venturi flow element, force main, and primary influent pipes. Throughout the study, the effects on other plant processes, such as sludge treatment, were recorded.
Based on preliminary jar testing to select the most promising chemicals, the full-scale study was conducted under five phases:
* Base polymer testing, injected at different locations
* Combinations of polymers and metal salts
* Coagulant dosage optimization
* Cold weather testing
* Additional dosage optimization
Base Polymer Testing
A base polymer was tested at full scale to obtain baseline data for comparing the effectiveness of alternative coagulants and to determine if polymer alone could improve effluent quality. Primary influent and effluent samples were taken six times a day and analyzed for BOD and SS concentrations. Each coagulant or combination of coagulants was tested for approximately one week. In addition, the performance of the coagulant at different feed locations was analyzed.
Early in the study, it became clear that the base cationic polymer, which was fed to the primary influent flow at dosages of 5 to 20 ppm, did not provide a significant improvement in effluent quality. Dye testing confirmed that the polymer added into the plant flow was inadequately mixed to develop proper floc formation.
Upon startup of a new feed system using gate nozzles on two tanks, polymer was injected at 15 ppm. The new feed system produced improved distribution of the polymer, but test results showed no improvement in removal rates when compared to the control tanks. In a final attempt to improve mixing and dispersion of the base polymer in the wastewater, the polymer was injected into the venturi chamber located on the force main upstream of the primary tanks at dosages ranging from 5 to 20 ppm.
To determine whether influent channel aeration was a factor in coagulant performance, the air was shut down for varying increments of time. While shutting off channel air improved surface clarity in the tanks and reduced scum accumulation with no significant grit buildup in the channels, effluent quality did not improve.
Phase I Conclusion: Data produced by Phase I indicated that the base cationic polymer alone was ineffective in achieving any net improvement in effluent quality. Regardless of the polymer's application point or dosage, removal efficiencies for BOD and SS never exceeded that of the control tanks (15 percent and 45 percent).
Ferric chloride was tested alone and in combination with the base polymer. At periods of relatively low flow (less than 80 percent of design average), the ferric chloride performed well by itself. As plant flow increased, however, the ferric chloride floc tended to rise over the effluent weirs. Evidently, a coagulant-aid polymer was required to settle the floc at these higher flows.
Following the addition of ferric chloride at the pump and power house, polymers were added at the venturi chamber. Anionic polymers tested in combination with ferric chlorides reacted rapidly to improve settling characteristics. However, during late morning and early evening hours, even the addition of the polymer did not improve floc settling characteristics. During this period, floc formation appeared finer and more granular, and the wastewater clarity decreased.
Ferric chloride also affected the plant's sludge processing system, which comprises gravity thickening and anaerobic digestion. Ferric chloride also decreased grease and floatables on the primary tank surface and reacted with hydrogen sulfide to reduce odors at the gravity thickeners.
Aluminum salts also were tested in combination with polymers. Although the addition of aluminum salts produced removals comparable with those achieved with ferric chloride, the effect of aluminum salts on sludge thickening prohibited its use. When aluminum salts were not added, the thickened sludge concentration averaged 5.2 percent; when ferric chloride was added, the average solids concentration was 5.5 percent. Aluminum salt addition, however, produced an average solids concentration in the thickened sludge of 3.4 percent during the first testing. A second, four-day test showed that thickened sludge concentration was reduced to 1.9 percent. Following aluminum salt addition, periods of no coagulant addition were required to allow the thickeners and digesters to recover. Excessive quantities of solids overflowed thickener weirs and returned to the headworks. Digestion process upset was indicated by the ratio of digester volatile solids to alkalinity and reduced gas production.
Conclusions: Ferric chloride and anionic polymer were selected as the optimal coagulants to achieve the proposed effluent limits of 30 percent BOD and 45 percent SS removals. The effluent limits were accepted by the NYCDEP and the NYSDEC.
Identifying the optimal coagulants answered only half the question. Remaining was identification of the optimal coagulant dosage, not only to achieve the required BOD and SS removals, but to accommodate existing treatment systems. While high coagulant dosages produced better solids removal, for example, the resulting sludge did not thicken well. In addition to the thickened sludge concentration decreasing under high dosages, the sludge blankets overflowed weirs and digester gas production and quality dropped.
Because of its effect on sludge treatment, the dosage of ferric chloride was limited. At the same time, Phase II testing showed that a higher metal salt dosage was needed during the daytime "hot spot" period when floc formation was poor. As a result, Phase III focused on a two-level daily coagulant dosage schedule. During each 24-hour period, ferric chloride was injected for 14 hours at a level of 15 ppm. During the remaining 10 hours, ferric chloride was added at levels ranging from 30 ppm to 60 ppm. The results of the testing were:
* Thickener and digester performance decreased at ferric chloride dosages of 50 ppm for ten-hour intervals.
* A dosage of 50 ppm ferric chloride for four hours per day produced enhanced BOD and SS removals without significantly affecting sludge processing.
* Addition of ferric chloride at a dosage of 60 ppm for four hours per day produced no appreciable benefit when compared to the results produced by the 50 ppm dosage.
Investigating Additional Factors
Even after the optional type and dosage of coagulant had been identified, several other factors required investigation. These factors represented the less-than-ideal conditions of the "real" world and the potential effects of those conditions on polymer performance.
Polymer Feed Location: Although the optimal feed point for the anionic polymer was close to the primary tank influent channel, operators found it difficult to balance the flow from the injector nozzles at the 16 primary influent gates. In response, diffusers were installed across the four influent channels. The response improved balancing, but sacrificed equal distribution of the polymer. The feed point then was moved to a 10-in. overflow connection to the 96-in. force main upstream, which eliminated balancing problems and provided adequate mixing.
Wet Weather Flows: The optimal dosage was identified for dry-weather flows. During peak wet-weather flows, when high dosages of ferric chloride were being added, ferric chloride floc occasionally was washed out over the effluent weirs. A visible floating slick, which was detrimental to receiving water quality, formed over the outfall. To prevent floc washout, coagulant addition was stopped when wet-weather flow exceeded 160 mgd. The NYSDEC allowed a waiver of effluent limits during these wet-weather flows.
Cold-Weather Testing: Just as wet weather flows affected the optimal coagulant dosage, the colder wastewater temperatures produced during winter months was expected to affect coagulant performance. Settling tank efficiency, for example, decreases in colder weather. The cold-weather test data indicated an average BOD removal of 5 percent lower than warm-weather removals, while SS removals were 4 percent lower.
After the cold-weather testing was completed, additional tests were conducted to refine coagulant dosages. The base line assumption was that the addition of ferric chloride at 50 ppm for 10 hours per day upset the sludge treatment process, while the addition of ferric chloride at 50 ppm for four hours per day did not adversely affect sludge treatment.
A ferric demand study (conducted by MIT/Delta Chemical) confirmed that the maximum daytime ferric demand ranged from 50 to 60 ppm, with nighttime demand ranging from 15 to 30 ppm. Additional trials were conducted with ferric chloride fed at 50 ppm for six hours per day and at 15 or 20 ppm for the remainder of the day. Dosages of 50 ppm for six hours per day produced preliminary results of 35 percent BOD removal and 57 percent of SS removal. Signs of sludge process upset developed when 20 ppm of ferric chloride were fed during the remaining 18 hours.
Over the long-term, dosages of 50 ppm ferric chloride for four hours or more per day inhibited digester performance and a dosage decrease to 20 ppm was necessary to allow digesters to recover. To meet the established effluent levels without endangering the digestion process, 20 ppm of ferric chloride were applied throughout the 24-hour period. The drop in ferric chloride eliminated the potential for digestion upset while producing no significant drop in effluent quality.
Although advanced primary treatment has a low capital cost, chemicals represent a significant operating cost.
Construction costs for the interim polymer feed system (two 5000-gallon liquid polymer storage tanks, three packaged polymer preparation and feed systems, a dilution water system, and associated piping, valves, and electrical system) was approximately $920,000. The construction cost of the interim metal salt coagulant feed facility (two 13,000-gallon ferric chloride storage tanks, concrete containment dikes, titanium gear pumps, controls, instrumentation, a prefabricated fiberglass pump house, piping, valves, and associated work) was about $200,000.
The price of ferric chloride was $0.30/gallon, while the price of the anionic polymer was $1.60/gallon. Based on the use of 7000 gallons per day of ferric chloride and 130 gallons per day of anionic polymer, the average cost of coagulants was about $2,300 per day during the last stages of the testing program.
Producing Required Results
To achieve the effluent limits accepted by the NYCDEP and NYSDEC of 30 percent BOD removal and 45 percent SS removal while avoiding disruption of the sludge processing system, ferric chloride and an anionic polymer coagulant were added to the existing treatment facilities. The optimal dosage of ferric chloride was established at 20 ppm and 1 ppm of anionic polymer throughout the 24-hour period.
The average monthly BOD and SS removals were 36 percent and 55 percent, respectively, compared to 15 to 20 percent BOD removal and 25 to 40 percent SS removal normally achieved without chemicals at the design overflow rate of the primary settling tanks.
The provision of alternative coagulant injection points was an important design feature for coagulant feed facilities. Before the optimal feed location was identified, five different injection locations were analyzed. For the ferric chloride floc to form, a reaction time of between 1.5 and 2 minutes is necessary between the addition of ferric chloride and the addition of the anionic polymer.
Advanced primary treatment will not produce effluent of secondary treatment quality but it can significantly improve primary settling tank efficiencies and allow high SS and BOD removals, provided that the sludge processing system can handle the resulting sludge. As a result, it can be an effective method of upgrading treatment to an intermediate level without a large capital investment.
Messrs. Chack and Rubino are Associates and Mr. Florentino is an Operations Specialist with Metcalf & Eddy; Mr. Krasnoff is Chief, Division of South Operations, and Mr. Liubicich is Chief Process Control Engineer, New York City Department of Environmental Protection.
|Printer friendly Cite/link Email Feedback|
|Author:||Chack, John J.; Rubino, Vincent; Florentino, Richard; Krasnoff, Paul J.; Liubicich, James|
|Date:||Sep 1, 1994|
|Previous Article:||Rehabilitation of a low-volume county bridge.|
|Next Article:||Using silicates to lower lead levels in drinking water.|