外文翻译-低温低浊水处理中混凝和直接过滤工艺的优化

1OPTIMIZING COAGULATION AND DIRECT FILTRATION PROCESSESFOR LOW TURBIDITY, LOW TEMPERATURE WATERSKEYWORDSCoagulation, Velocity Gradients, Low turbidity Water,Natural Organic matter (NOM), Direct FiltrationABSTRACTThis study aims at exploring optimum coagulant dose and optimum pH for a low turbidity, low temperature water. Coagulation of low temperature and low turbidity waters generally requires addition of clays for adequate quantities of hydro-oxide precipitate to remove colloidal matter. This study focuses on particulate and natural organic matter removal from low temperature and low turbidity water without the addition of clays. An optimum combined dose of ferric sulfate and a cationic polyelectrolyte at an optimum pH using pilot scale direct filtration train was achieved. This work, not only recommended the elimination of a sedimentation basin but substantially reduced chemical cost for coagulation and pH adjustment. It was observed that use of color removal instead of turbidity removal as a variable of direct filtration performance was well suited to low turbidity waters.INTRODUCIONAll surface waters that are used for water supply require treatment to remove contaminants both for health as well as aesthetic reasons. Major contaminants that are found in surface waters are clays, microorganisms, natural organic matter (NOM) and trace metals in a few cases. Solid-liquid separation processes such as coagulation and filtration are often employed to remove these undesirable contaminants. These processes when optimized, can remove all organic, inorganic and suspended matter to a level below national water quality standards.Low turbidity waters are hard to coagulate due to low concentrations of stable particles. Before the discovery of Giardia problem, many treatment plants in US used to stop adding coagulant during winter and spring months as the turbidity of the incoming water was already below the national primary drinking water standards. A few treatment plants would add only token coagulants to, presumably, take advantage of filtration. Filtration, though in place, was effective only after effective coagulation Al-Jadhai, 1993.Color in water stems from decayed vegetation, humic substances and natural as 2well as anthropogenic organic sources. Effective color removal is an indication of effective coagulation. Lefebvre et. al.2003 reported that NOM can be best removed at pH around 6 while using Ferric Iron as sole coagulant.On one hand color in water can serve as surrogate for NOM Edzwald, 1979, on the other hand only effective coagulation would result in effective filtration as transport and attachment are the dominant steps in both operations Habibian et. al. 1975. Whilst turbidity represents the bulk water characteristics and cannot provide any insight on gradation of colloidal and suspended matter, removal of microscopic pathogens of certain size range cannot be ensured. Particle counting and particle size distribution (PSD) reveal a clear picture about the size distribution of particles that have been removed or reduced. One disadvantage of particle counting is the coagulation of submicron particles into microscopic zone. Rise in total number of particles may not correlate with decrease in turbidity. This study was conducted to optimize pH and coagulant dose using a pilot scale direct filtration plant in parallel with an existing conventional treatment plant in Rhode Island USA. Stakeholders sought to explore cost reduction measures including changes in the treatment train, if economical. Rather than traditional variable of effluent turbidity, color and particle counts were used to determine treatment effectiveness.Major objectives of the study were to determine the optimum dose of ferric sulfate with and without polymer addition using color and particle counting as performance indicators, and, to examine the effectiveness of direct filtration scheme compared with conventional treatment train. This study covers a comparison between the pilot plant and the full scale water treatment plant for low turbidity water for winter months only. No data was collected or compared for the summer time. Water temperature in Rhode Island varies between -10 to + 20 Co. No data was collected for biological variables.MATERIALS AND METHODSExperimental investigations of this study were divided into three phases: Phase-1: Laboratory scale investigations involving optimum coagulant dose and pH determination. Phase-2: Pilot scale confirmation of the effectiveness of the optimum dose at optimum pH. Phase-3: Comparison of pilot plant results with actual plant.Twenty-five jar testing experiments were conducted in phase-1 within a pH range of 6-10 and ferric sulfate dose of 0-20 mg/L. Color, remaining after coagulation and 3settling, was used as the indicator of performance.Polymer was added as secondary coagulant in the second phase of jar testing experiments. A total of twenty-six experiments were conducted using ferric sulfate dose of 10, 8, 6, and 4 mg/L along with polymer dose ranging from 0-1.0 mg/L at pH of 6-10. Total number of particles/mL and color, remained after coagulation and settling were used as performance indicators.In phase-3, a pilot filtration plant was used to assess the effectiveness of selected coagulant dosage in filtration. Another set of two filtration experiments was conducted to compare filter response of pilot plant with the real plant. Remaining color and number of particles per mL in the effluent were used as performance indicators.Chemicals and equipment used in the above mentioned experiments is briefly described in the subsequent section.ChemicalsLab scale investigations involved extensive jar tests using ferric sulfate, Aqualyte 505 (a cationic polyelectrolyte) and quicklime. All of these chemicals were included in the list of chemicals approved by the American Water Works Associations (AWWA, 1991) for drinking water treatment and were being used by Situate Water Treatment Plant.EquipmentA Phipps and Bird (Model No. 300) six paddle stirrer, equipped with a 115 volt AC motor, a 110 V, 60 Hz speed controller (0-150 rpm) illuminated table were used in this study. The function of the illuminating table was to aid in watching the floc formation.A PlatinumCobalt solution of 500 mg/L concentration, manufactured by HACH Chemicals, Colorado, was used to prepare color standards from 0 to 30 Pt CO Units in accordance with method described in “Standard Methods, Section 118”.A Fisher Accumet (Model142) pH meter manufactured by Fisher Scientific Company, Pittsburg, PA, was used to measure the pH of the treated and untreated waste.A HIAC (Model PC-320) automatic particle size analyzer was used to determine the number and the size frequency distribution of the particles present per unit volume of the sample. When a particle of projected area “a” passes through a window, the particle partially blocks the light beam, the reduction being equal to a/A where A is the area of the window. Since the preamplifier output without a particle is -10 volts, 4the amplitude E0 of the pulse produced by the particle is given as;(1)01aEVAThe voltage corresponding to the appropriate particle size is dialed into the threshold set controls for each channel of the analyzer and is displayed on the front panel as the number of particles per unit volume of the sample in the reach of the channel size. At least two samples of distilled water were analyzed before each set of samples were drawn during the experiment.The power transmitted to the water in the rapid mixing and slow mixing chambers of the pilot-plant was determined by Al-Abdulali 1987 using a Bexometer Model 38 Torque-meter and multiplying the observation with the rpm of the stirrer whereas:Power transmitted = torquerpm mixer (2) Jar testing results are affected by a large number of factors such as the coagulant solution strength, coagulant dosage, pH conditions, method and sequence of coagulant addition, intensity and duration of flocculation, method of sample withdrawal, temperature, the type of equipment and jars used for testing. Measures adopted for quality control over the jar testing results during the coagulation experiments were the same as suggested by various other researchers such as Amirtharajah 1978, Brink et. al., 1988, Stumm and Morgan, J. J.1991.Pilot Filter Column SpecificationsThe pilot filter consisted of a cylindrical plexy-glass column with the following properties:Table 1: Filter Column SpecificationsThe filter column was capped at both ends with two-7/8 inch thick Teflon plates. At the inner side of the Teflon plate, a thick stainless steel screen was fixed in order to distribute the flow uniformly. A flow control valve and a flow meter was installed at the downstream side of the filter to control the flow rate into the column. Nine-3/16 inch O.D steel tubes were embedded at different depths in the filter bed. These tubes were connected to the manometers on the manometer board. The filter ends of each 5tube were covered with a stainless steel screen of diameter less than the smallest grain diameter of the filter bed. The filter column was also fitted with a backwash flow valve which allowed the treated water from the treatment plant to flow upward in order to back wash the pilot filter bed as needed.Figure 1: The pilot filtration plant.The first sample in each run was collected about 15 minutes after the start of the filter run. Only the effluent samples were collected at different time intervals. Samples collected were analyzed for color and the number of particles per milliliter as soon as possible. The Influent pH and the flow rates were maintained at 7 and 0.49 gpm (2.5 gpm/ft2), respectively, throughout the filter run.Backwash CriteriaIves 1978 and Cleaseby 1995 have suggested that the filter should be so designed that it could reach the headloss limit at the same time as the effluent concentration breakthrough limit was reached. For this study, the filter run was terminated as soon as the total headloss reached the available headloss (approximately 200 cm of water). The rate of backwash was kept low at the start of the backwash and then raised slowly up to 6 to 9 gpm/ft2.A total of 8 experiments were conducted using this pilot filter column for pH 6levels of 7, ferric dosages of 4 and 10 and polymer dosages 0, 0.2, 0.6 and 1 mg/L.RESULTS AND DISCUSSIONAs mentioned earlier, phase-I of this study comprised of jar testing experiments meant to determine the optimum coagulant dose and pH. Ferric sulphate was the only coagulant used in these experiments. Experimental conditions for this phase are given below:Table-2: Experimental condition for phase-1 experimentsResults of this phase of experiments are illustrated in Figure-2:It is clear from Figure-2 that optimum dose of ferric sulfate lies between 4 10 mg/L for a wide Ph range. It was therefore decided to use a ferric sulfate dose of 4-10 mg/L for determining the optimum pH and the combined coagulant dose ie., ferric sulfate and the polyelectrolyte. Polyelectrolyte dose of 0-1.5mg/L is typically used in US water treatment facilities JM Montgomery, 1995. The Situate Water Treatment Plant was not using any polyelectrolyte at the time of this study. The objective of adding a polyelectrolyte was to eliminate the need for a sedimentation basin and 7develop stronger flocs that could penetrate into the filter bed leading to the longer filter runs. Figures-3 and 4 illustrate that number of remaining particles/mL as well as remaining color after 2 minutes of coagulation and 20 minutes of flocculation reduced continuously with the rise in polymer dose. Coagulant dose of 10 mg/L was used for these experiments. Most effective pH for color removal was pH 7 and for particulate removal was pH 10 followed by pH 7. Most effective polymer dose was 1 mg/L in both cases.In the next phase of experiments, Ferric sulfate dose was reduced to 8 mg/L while keeping all other experimental conditions the same. Figures 5 and 6 show the results 8of these experiments. Maximum color removal took place at a polymer dose of 1 mg/L and at pH 7. Quiet the same is evident from figure-6. Max particulate removal takes place at a polymer dose of 1 mg/L irrespective of coagulated water pH. Since pH=7 is close to the raw water pH, it is safe to say that particulate removal is also maximum at pH=7.Further reduction in ferric sulfate dose to 6 mg/L did not improve the color removal or particulate removal (Figures 7 & 8). Results of this experimental set were however comparable with the previous one where 8 mg/L of ferric sulfate was used for the same level of color and particulate matter reduction. Most effective pH was again pH = 7.9Reduction in coagulant dose to 4 mg/L resulted in color removal to 7 Pt Co. Units and particulate removal to 200 particles/mL at pH 7 as shown in Figures 9 and 10 respectively.The most effective polymer dose was 1 mg/L at pH levels. Results of the above investigations are tabulated below:Table 3: Summary of results from Phase-II experiments-Table 4: Experimental conditions for Phase-III experiments10Further reduction in coagulant dose to 0.2 mg/L deteriorated both color removal and particulate removal to a poor level. Keeping in view the above results, coagulant dose, for the direct filtration experiments was decided as follows:Using the above mentioned coagulant dose and pH conditions, 8 direct filtration experiments were conducted to confirm that effective coagulation results in effective filtration and to determine the finalcombination of primary and secondary coagulants. No. of remaining particles per mL, remaining color and headloss were used as the variables of performance. Results of first 4 experiments are shown in Figures 11, 12 and 13 in terms of filter run time vs remaining particles/mL, remaining color and headloss. Primary coagulant dose and pH were fixed at 4 mg/L and 7 respectively, in all the four experiments. Polymer dose was however varied from 0.2 to 1.0 mg/L. 11Results in Figure 11 demonstrate that number of particles per mL remain below 10 particles /mL throughout the experiments at 0.2, 0.6 and 1 mg/L of polymer. Remaining color drops at higher polymer dosage. A polymer dose of 1 mg/L is most effective in removing color. Headloss is not much affected with the change in polymer dose. A polymer dose of 1 mg/L shows a steady rise in head-loss though.The second set of experiments was conducted using coagulant dose of 6 mg/L as shown in figures 14-16. Remaining color stays between 6-10 Pt CO unit for most part of the filter run. Polymer dose of 0.2 mg/L and 0.6 mg/L showed comparable results. Remaining color after coagulation was also within the same range. This shows that filtration through uniformly graded sand does not have a very astonishing effects on color removal. No. of particles per mL remained below ten for most part of the 12experiment, showing that sand filtration is an effective means of particulate removal. Polymer doses of 0.2 and 0.6 mg/L showed effective particulate removal. Headloss rose steadily, showing that floc penetration into the filter bed continued throughout the filter run at full range of polymer dose.Keeping the above results in view, a combined dose of ferric sulfate = 6 mg/L and polymer = 1 mg/L at pH 7 was concluded to be the optimum coagulant dose for direct filtration treatment of the low turbidity, low temperature water from the Situate Reservoir.A final comparison of the effluent characteristics using selected optimum coagulant dose at the optimum pH of the direct filtration pilot plant and the conventional filtration plant at Situate are illustrated in figures 17 and 18. These results show that the effluent quality produced under optimum direct filtration conditions at the pilot plant is comparable with the effluent quality of the actual treatment plant.CONCLUSIONSExperimental work carried out in this study lead to many useful conclusions. Whilst the results are based upon water quality of winter months, they are applicable throughout the year as polymer addition is expected to strengthen flocs and a combination of surface and depth filtration would prevail. Important conclusions from this study are delineated below:The optimum coagulant dose and pH for Situate Water Treatment Plant in Winter 13months is as follows:The Situate Treatment Plant should switch to direct filtration train during low turbidity season. The plant would save approximately US $ 27,000 per year in terms of chemicals and energy in moving and collecting water to and from the sedimentation basin.Coagulation at optimum dose is effective in color removal. However sand filtration does have major impact in color removal. Sand filtration, however, plays a dominant role in particulate removal after effective coagulation with optimum coagulant dose. The hypothesis that effective filtration of low turbidity water would result only after effective coagulation proved to be correct.Use of polymer does not only reduces ferric sulfate dose but also the color imparted by ferric sulfate.RECOMMENDATIONSAs a result of this study many new ideas came into being. For example, the use of on-line particle counter and the development of a computerized system that would provide instant information to the operator for quick response. Testing of other polymers currently being used by the water treatment industry may further economize the plant operations. Revisiting the existing filter media being used at the Situate water treatment plant may lead to a better effluent quality and longer filter runs.ACKNOWLEDGEMENTThis work could not have been possible without the courtesy of the management at the Scituate Water Treatment Plant, Providence. Authors gratefully acknowledge the support of the treatment