Problems of Water Purification and Softening
Some of the Difficulties Experienced in System of Newark, Ohio— How They Were Overcome—Use of Clarifier-Some Soot Troubles
THE various problems that arose in the working of the newly installed water softening and purifying system of Newwark, Ohio, as described in the following paper with the methods employed to overcome them, will no doubt offer some valuable hints to other water works which are struggling with similar troubles or are about to adopt a parallel method of dealing with the water supply:
The original works of the purification plant of Newark consisted of an infiltration system in the bed of the Licking River used in time of flood, a direct intake for the low service pumps, a small suction reservoir for the high service pumps, and a high elevation storage reservoir which supplies the city by gravity. The new work includes a new concrete intake with gravity conduit to the newlow service pump house, new low lift pumps, and a six million gallon water softening and purification plant, the infiltration system and high elevation storage and high service pumps being all retained.
The Licking River, from which Newark takes its supply, has, above the intake, a water shed area of about 225 Sq. miles.
The stream flow is sufficient at all times so that no impounding reservoir has been built. A low dam just below the old intake built to submerge it also submerges the new one. The topography of the country is rolling, the soil loose and gravelly, and the resulting quick runoff, coupled with the fact that there is no impounding produces a very flashy and shifty stream.
The new low lift pump house contains three electrically driven pumps of the centrifugal type of two, four, and six million gallon capacity. Regulation of the water pumped into the plant is accomplished by throttling the pump discharges.
Chemical House and Filter House in One Structure
The chemical house and filter house are in one superstructure, located adjacent to the high service pumping station. Within the chemical house arc lime storage bins holding two cars, lime unloading and handling equipment, scales, slackers, and solution tanks. Additional storage space is provided for three cars of soda ash and alum in bags. The lime is unloaded by means of a power shovel and elevated into the bins with a bucket elevator and screw conveyor. Two Booth lime feeders measure the milk of lime into the water. Alum is applied by two dry feeders furnished by the International Filter Company. Also located in the chemical house are the 24-inch Venturi which measure the raw water, sutro weir box, and preliminary mixing tank or saturator with mechanical agitation. The sutro weir box is supplied with six weirs, and by means of these most any division of the flow is possible. At present 1/3 of the flow is diverted through the mixing tank just mentioned, receiving all the chemical application and is retained for about twenty minutes. This over treated portion then combines with the remaining 2/3 of the raw water and passes to the gravity mixing tanks. This arrangement of chemical treatment is called “split treatment.”
The gravity mixing chamber is located adjacent to the settling basins and is so arranged that the water can pass through continuously as in series flow or in a parallel manner, the water entering at two points through branch lines, or if desired either half of the chamber may be used alone. The tank is covered and provided with suitable manholes and inlets. The baffle system is a combination of the over and under and around the end types, there being 166 baffles, some of wood and some of concrete. At nominal capacity the tank affords a one-hour retention period and a velocity of five feet per second.
Description of the Dorr Clarifier
The Dorr Clarifier, located between the mixing tanks and settling basins, is a circular concrete tank forty feet in diameter and about eight feet deep, with a conical bottom sloping toward the center. The flow through the tank is radial, the softened water laden with floe from the mixing tanks, entering at the top and center and the clear water overflowing around the circumference. It is designed for a retention period of 22 minutes at nominal capacity. The Dorr mechanism consists of a slowly revolving rake which lays very close to the bottom of the tank and moves the settled solids slowly toward the center. These plows revolve four times per hour and are driven by a threehorsepower motor with suitable speed reducers. The sludge is drawn off continuously underneath by means of a six-inch line connecting with the center of the tank. Sludge discharge is regulated by throttling six-inch valve. The inlet and outlet lines to the clarifier are provided with valves so that the clarifier may be by-passd if desired.
There are two separate settling basins with center baffle of one million gallons capacity each, and which give an eight hour retention period at plant capacity. The velocity at 6 m. g. d. is about five-tenths foot per minute. They are arranged for series or parallel flow or may be operated singly.
As tbe plant neared completion it was necessary to sacrifice some small refinements in order to cut down the total cost, and it was found impossible to construct a separate carbonation chamber. Accordingly, the diffusers were laid on the return or outlet side of the basins. Flue gas from the boilers was drawn through a scrubber and drier by a steam driven compressor and then forced through the diffusers into the water at the outlet of the basins. These diffusers were constructed of a very lean mixture of coarse sand and cement cast around a perforated pipe.
There are four filters of 1 1/2 million gallon capacity similar in design to those at Columbus. They are fitted with Simplex tables, gages, and controllers. A 50,000 gallon elevated steel wash water tank supplies water for about 1 1/4 washes at a 24-inch vertical rise per minute. The pipe gallery is interesting in that it is especially accessible. A small clear well of 225,000 gallon capacity, in two compartments, lays under the filters and empties into the suction well for the high service pumps through two 20-inch conduits. Disinfection is taken care of with a Wallace & Tiernan vacuum type chlorinator discharging into the high service suction well.
Some Minor Difficulties Through Night Operation
It is interesting to note at this time that the pumping station is also the municipal light plant, and a part of the current generated for the street lights is used in the plant. As the light plant operates only at night and operates more economically under a heavy load, it has been the practice to operate the softening plant at night also, and this has introduced minor difficulties, such as adjustment of chemical feeds daily, settling of the sludge in the mixing tanks, and settling of the lime solutions. The daily periods of operation vary between 12 and 20 hours, the plant starting up in the afternoon and shutting down early the next morning.
Overcoming Flood and Bad Weather Troubles
The plant was first started up during the second week in March, 1924. The weather was very bad with considerable rain and snow and frequent floods and high turbidities. The current is very swift around the intake during these floods and on two or three occasions threatened to undermine the intake and gravity conduit. To prevent this a break was constructed out of brush about 500 feet above the intake which it is hoped will straighten the channel somewhat. In addition to this a protecting embankment of heavy concrete has been burit up and down stream from the intake. For these first few months it was the practice to alternate between the 2 and 4 million gallon pumps to obtain a daily average of about 3 m. g. This required very careful adjustment of the chemical feeds and it was soon found that best results could be obtained by throttling the 4 m. g. down to the required daily amounts. This arrangement is being used at this time and very steady chemical treatment is obtained. The caisson system is not connected up yet with the new low lifts, but when completed will make it possible to escape the high turbidities in time of flood. Since the plant is operating intermittently, it is possible to shut down for short periods in order to allow these short flashes of mud to go by.
The average hardness of the raw water is about 275 p.p.m. and this is being reduced to an average hardness of 85 p.p.m. in the filtered water. During the summer and fall months the river water has been exceedingly clear with no apparent turbidity or organic material, and yet after a hard rain it is possible for it to change in two or three hours to a turbidity of 2,000 ppm. or more and to a hardness of possibly 80. These turbidities are usually of short duration, however, unless the rains are over prolonged periods. The chemical treatment requires about 9 grains of lime, 1.5 grains of soda ash, and 1 grain of alum. As long as the incrustant hardness is below 40 p.p.m. soda ash is dispensed with and lime used alone.
Problems in the Mixing Tanks
“Pebble” quick lime is being used at the present time and is stored in air tight bins above the solution tanks. It is weighed into the slackers by hand from which the milk of lime flows into the solution tanks below the slackers. The slackers are stirred by electrically driven rakes and are fitted with water and steam lines. The large solution tanks are agitated bv Albright-Nell “perfect circulators.” These are double propeller type stirrers with both propellers lifting, the propellers being fitted to concentric shafts. Soda ash, when used, is added to and fed along with the milk of lime solution. Alum is fed in granulated form through the two dry feeds. These consist essentially of a revolving hopper and a deflector knife. The hopper is rotated an arc of a circle five times a minute, the alum falling into a whirling spray of water which carries it to the reaction tanks. The arc is varied by the automatic proportional feed connected to a float in the weir box. The deflector knife may be adjusted over a wide range of settings by hand. Some trouble was experienced with the agitators in the lime tanks after a few weeks operation, the lime creeping up between the two shafts and causing them to bind frequently. After repeated attempts to prevent this, one of the drive motors burned out and it was thought best to use one propeller only and samples of the solution taken over a run of the tanks show that very good agitation is still obtained. The lime saturator or preliminary mixing tank is stirred by the same type of mechanism which has been arranged in like manner.
As a result of the intermittent operation, it was observed that a number of the channels in the mixing tanks did not have the proper velocity, and on investigation some were found to be very badly sludged up. They were at first operated in parallel and as the pumping averaged 3 m. g. the velocity was rather low, so that conditions favored partial settling in the tanks. In a short time it was found that a great part of the water was finding its way through a series of flap gates provided as a safety measure in each channel wall, instead of traversing the entire distance through the tank.
This served to by-pass the mixing tanks and rather poor reactions resulted as the water was not retained for the proper interval. On emptying the tanks we found debris of different kinds and among it several sacks of sand and gravel, an overcoat, and two ladders. After cleaning and before refilling the tanks, all the flap gates were fastened tightly shut, and only one tank was put into service. As a result a very satisfactory velocity is obtained in every channel and there’is no evidence of settling, the interior being scoured clean.
Removing Sludge from the Dorr Clarifier
Some difficulty was experienced in getting the sludge to carry away from the Dorr clarifier the first few days, and after examining the sludge line, a piece of wood was found obstructing the entrance to the drain. On removing this we had trouble throttling down the discharge enough. The round feed well had a tendency to produce excessive inlet velocities and short circuits, and to break these up a baffle of two-inch iron strips spaced about two inches apart was hung in the feed well opposite the inlet pipe. This apparently has been very effective as there is now no evidence of short circuiting. The velocity and retention at 3 m. g. or even 4 m. g., appears to be satisfactory but when the pumping exceeds 4 m. g. considerable floe is carried over the outlet weir. To regulate more carefully the sludge discharge, a decant pipe has been arranged so that a sludge containing as high as 15 per cent, solids has been produced and discharged by gravity. The amount of water wasted along with the sludge has not as yet been accurately measured, but is estimated at about 10,000 gallons daily. The highest solid content obtained by throttling the six-inch valve was around 4 or 5 per cent., but the decant has made it possible to maintain an average of about 12 per cent. It is recommended for best results that a blanket of sludge about one foot deep be maintained over the rakes to effect the maximum concentration and prevent any stirring action. Daily observations may be made at the time when the plant is not in operation, as the water is very clear and every detail of the tank easily seen through the day. Repeated tests made on the influent and effluent water show that an average removal of suspended solids of 98.5 per cent, is obtained. However, to maintain this figure, it is necessary to keep a close check on the chemical treatment and carry at least 5 p.p.m. caustic alkalinity.
Size of Basins Could Be Reduced
As a result of the performance of the clarifier, it would seem that the basins are possibly somewhat larger than necessary. After four months operation only about two feet of sludge was found on the inlet side of the baffle, and a few inches on the outlet side. There were intervals, too, during these first months, when the clarifier was not operating. It is certain, however, that it will be necessary to wash them no oftener than once a year. The basins were operated in parallel at first, and carbonation took place in both of the outlet compartments. The upward rise of the gas produced a boiling effect at the surface which set up currents reaching half the length of the basin, and it was found that almost the entire basin was carbonated. This had the effect of cutting down the period of contact for proper sterilization with the lime, and finally the basins were changed to series operation and carbonation practiced in the outlet of the last basin.
(Continued on page 276)
Water Purification and Softening Problems
(Continued from page 266)
Some Troubles with Soot
After only a few days operation the flow of gas from the diffusers seemed to diminish, although there was more than sufficient pressure on the lines, and the second week it was concluded that the diffusers were becoming clogged with soot. At the first opportunity the basins were drained and the cement diffusers broken away from the pipes. The interiors were found to be packed solidly with soot and tarry matter. Thereafter the gas was allowed to flow from the perforations in the pipes, and no difficulty was encountered in fully carbonating the water. The trouble did, however, crop up at this time in the boiler room. The boilers are hand fired with soft coal, and consequently the flue gas carries considerable amounts of soot. This would not cause any material damage in the process if it would pass readily through the lines to the basins, but such is not the case, for the soot congeals out on the oily interior of the compressor and produces a knock. It became necessary to clean the compressor almost every day and finally it was possible to run only a few hours before it became clogged. To bring about a better scrubbing the coke was removed from the scrubber and broken down to 1/2 inch size and the water sprays were increased, but without apparent improvement. Then a light oil. such as is used in refrigeration machines, was used in the compressor, and this gave slightly longer runs. Next the drying tower which was filled with excelsior, was packed with alternate layers of burlap and excelsior. This proved a very good filter, the hairy fibres of the burlap serving to catch and hold the soot particles. This filter held for about two weeks and then broke. Just at this time, too, a very disagreeable taste was noticeable throughout the distribution system, and was easily identified as phenols.
In hand fire boilers there takes place more or less distillation and phenols are produced in small quantities where this occurs. Slight tastes and odors had been reported here and there about the city, and were occasionally noticed in the laboratory, but they existed over only short periods and could not be definitely identified. Immediately after the new vacuum type chlorinator was set up and in use the tastes became very pronounced, phenols being greatly exaggerated by chlorination. For the time being carbonation with boiler flue gas has been given up. As soon as the carbonator was discontinued the taste disappeared.
Carbon Dioxide Generated in Coke Furnace
It was then planned to construct a coke furnace to generate the carbon dioxide, and fortunately a used household furnace with fire brick lining was obtained at small cost and is proving very satisfactory. The grate area is 14 x 40 inches, and about 100 pounds of coke are required per million gallon. The process is controlled by hourly test of the phenolphthalein alkalinity, and it is aimed to carry about 1 part per million of COa in the applied water. Regulation of the concentration of the gas going to the diffusers is accomplished by air cocks, stoking, dampers, or by varying the speed of the compressor. Samples of the gas taken at the compressor show an average of 5 per cent. COa, while that emerging from the carbonation chamber (bubbles) shows only .5 per cent., or in other words 90 per cent of the available carbon dioxide is used up. If the free COa in the applied water is present in amounts exceeding 2 p.p.m., there is a rise in the total alkalinity of the filtered water, due no doubt to the fact that some solution of the mat on the filters occurs.
Sand samples from the filters after six months operation show that no change has taken place in the filtering material since the original installation. The sand has an effective size of .35 mm. and a uniformity coefficient of 1.20. No incrustation is noticeable although the carbonator has not operated continuously.
Filter runs are from 40 to 60 hours in length and the loss of head is very low. Length of run is determined not by the loss of head but rather by the appearance of the filters, the mat usually breaking after 40 or more hours service.
Until the first of September an old type chlorinator discharging into a dead end of the suction line was in use, and necessarily gave very irregular treatment with heavy slugs at times. This accounted for the presence of tastes at intervals.
Operating Results. Average of 10 Sets of Samples Taken from the Dorr Clarifier Over a Period of Three Weeks
Average of One Month’s Operation of Plant with Carbonator in Service the Greater Part of the Time. Soda-Ash not Being Used at this Time
Excellent Results from New Chlorinator
When the new chlorinator was installed discharging into the suction well a very even treatment was obtained, and it was found that very light doses were necessary to prevent tastes from chlorine, due possibly to the very low organic content of the water. In fact, the consumption of chlorine is so small at present that occasionally the residual equal to the applied.
It has been suggested that settling basins and filters might well be dispensed with as a result of the performance of the Dorr clarifiers. There is no doubt that the size of the settling basins may be greatly reduced, or it is possible that they might be entirely eliminated from the design of future plants, but it would appear that filters were practically a necessity. When washing a filter it is surprising to observe the amount floe caught during a run, when the extreme clarity of the applied water is considered.
(Excerpts from a paper read before the fourth annual Conference on Water Purification at Columbus, Ohio.)