Efficiency of Coagulating Basins
Shortly before the World’s Fair at St. Louis, as an emergency measure, coagulation with iron sulphate and partial softening with lime were added to the system of plain sedimentation which had been in use for years. The scheme made such a manifest improvement in the appearance of the effluent that it was accepted by the local public as the last word in water purification. At times the effluent has been excellent in every way; again, of doubtful character. It is only within the past year that the suggestion has been entertained to supplement with a filter plant the abridged form of treatment now in use. In an extended report on the “Water Supply of St. Louis,” published in an edition of 1,000 copies last January, the writer reviewed accumulated data touching the “quality of the raw water, the adequacy of the process of treatment, the present capacity of the plant, and such extensions and changes as give promise of providing for present and future needs of the community a supply ample and of a quality acceptable to a public whose judgment has been made keener by progressive improvement through eight years past.” At the suggestion of your secretary I am presenting this summary of the facts bearing upon the efficiency of coagulating basins. Such basins form an essential part of every plant for the purification of turbid waters. Without doubt everyone concerned with the operation of a filter plant has consciously or unconsciously had trouble from all the disturbing influences mentioned. The clarification plant of the St. Louis water works comprises a coagulant house ample for treating with lime and iron sulphate more than 160,000,000 gallons per day. with a series of unbaffled basins. We have no filters. There are six basins of about 25,000,000 gallons capacity, one of 40,000,000 gallons, and two of 20,000,000 each, through all of which in series the treated water flows. The combined capacity of the system is 230,000,000 gallons; the combined area about 52 acres.
Capacity of the Plant
It has been popularly assumed that the plant as now constituted is ample for adequate treatment of an indefinitely large volume of water. One writer estimates the capacity of the plant at 250,000,000 gallons per day. Operating results warrant a much lower estimate based on the degree of success attained in treatment of varying quantities of different waters, ranging in suspended solids from 14 to more than 8,000 parts per 1,000,000. There is reason to believe that even in the first year of operation the clarifying system was overtaxed. The records of this laboratory afford but 12 determinations of suspended solids in river and clear well samples during the year 1904-1905, of which two must be discarded for obvious reasons. All are given in the table, with consumption for their respective dates:
CLARIFICATION RESULTS—FIRST YEAR OF OPERATION.
In the absence of any evidence to the contrary it is assumed that the scattered analyses are representative of the working of the plant. Treatment was begun on March 22, 1904; the first recorded result nine days later suggests the inadequacy of the plant, which then comprised but the six basins in series, to clarify 67,000,000 gallons per day when the suspended matter was 2,400 parts per 1,000,000. For 97.5 parts per 1,000,000 of suspended matter in the finished water indicates the introduction into the distribution system on the corresponding day—April 1, 1904— of 27.5 tons of suspended solids. It is possible that this result was abnormal, since mud previously deposited in conduits and clear well may have been loosened by the flow. The second result. two months later, shows improved working. but suggests that 75,000,000 gallons could not be handled properly when the suspended solids in the river exceeded 1,700 parts. According to present standards, the only passable results in the first 12 months of operation are those of November 30 and April 1, 1905, when river solids were less than 1,000 parts and the daily consumption was under 80,000,000 gallons. Under other conditions the finished water contained notable amounts of solid matter. High consumption during the World’s Fair period overtaxed the new clarification plant the entire summer. Still these results show a vast improvement over conditions in the year prior to the introduction of the lime-iron method, when the public was expected to make no complaint when 450 parts per 1,000,000 of suspended water appeared in the tap water. In the second year of operation results are generally better. However, of 130 recorded determinations of suspended matter in clear well samples, 40 gave negative results. Figures given are averages of 130 determinations for river and 90 for clear well. Average daily consumption for each month is from high-Service pumping.
CLARIFICATION RESULTS SECOND YEAR.
Compared with the new standard set by the previous year, these results are good. Apparently lower consumption favored better operation; although the occurrence of so many impossible results (negative quantities) makes interpretation difficult. The records show that from June to January, inclusive, high caustic alkalinity was carried in the treated water. This is, perhaps, the explanation of the high suspended solids recorded. which seem to indicate somewhat less than 80,000,000 gallons per day as the maximum capacity of the plant when the river was carrying 1,800 parts per 1,000,000 suspended solids. The writer specifically di‘ laims any responsibility for the foregoing rc-ults of operation and analytical data. Results (or subsequent years were determined under his direction, and are believed to represent with a fair degree of accuracy the working of the plant. Averages are used in lieu of rehearsal of the full detail of determinations made on all save holidays. Inasmuch as total displacement of water in the entire basin system requires from seven to 15 days, the averages by months of operation better represent the blended waters issuing from the clear wells to the distribution system. They are platted in Diagram 1, in sequence with those of 1905-1906.
The diagram shows graphically that when suspended solids in the raw water were below 1,500 parts were 1,000,000 and consumption did not exceed 75,000,000 gallons per day, the average suspended solids in the clear well water were usually hot more than 1 or 2 parts per 1,000,000, increasing slightly with a rise in either river solids or consumption, and greatly with concurrent rise in both. It is noticeable that when pumpage was much above 75,000,000 gallons and river solids greatly exceeded 1,500 parts, the quality of the treated water was seriously changed for the worse. In general the curves of suspended matter in the treated water reflect the effect of high river solids and high consumption, following one or both in extreme cases, and illustrating the fact that these uncontrollable factors produce conditions which the present plant cannot meet. It appears that a raw water carrying not more than 1,500 parts of suspended matter can be made acceptable to the public of St. Louis so long as consumption does not exceed 75,000,000 to 80,000,000 gallons per day, but that the capacity of the plant varies inversely with the suspended matter. The diagram further shows that from 1905 to 1910 consumption exceeded 75,000,000 gallons only in the summer and fall months, while in 1911 and 1912 the average daily consumption has rarely fallen below this figure at any time The rise in river stage caused by rains of the late spring and summer brings high average suspended solids in the raw water, for suspended solids vary with the stage. Coincident with these times of high turbidity comes the heaviest draught upon the distribution system, when lawns arc to be sprinkled and the greatest waste of water occurs. It is unfortunately true that periods of highest turbidity are generally periods of greatest consumption.
Efficiency of Clarification
There is abundant evidence that efficiency of clarification and bacterial reduction are conditioned by:
Rate of flow through basins.
Amount of sludge already deposited.
Character and quantity of suspended solids in the raw water.
Temperature of water in river and basins.
Wind velocity and direction.
SUSPENDED SOLIDS IN EFFLUENT OF SUCCESSIVE BASINS.
When a water properly treated is passed through the settling basins, 97 to 99 per cent, of the suspended matter is precipitated in the first basin, the percentage removal depending upon the character and quantity of solids contained, the temperature, wind velocity and direction and the amount of sludge in the filling basin. In passing succeeding basins the remaining suspended matter undergoes a further reduction of one-half or one-sixth, corresponding to a few hundredths of 1 per cent, reckoned on suspended matter in the raw water, likewise dependent upon velocity, size of particles, wind and temperature. The major portion of clarification is, however, accomplished in the filling basin of 25,000,000 gallons working capacity. There is no provision for applying chemicals after water enters the basins. Efficiency of the plant, therefore, depends primarily upon the volume of water which can be satisfactorily clarified in the filling basin. The weight of suspended matter in tne effluent of successive basins varies with the weight of solids carried by the raw water. The percentage is approximately constant; the actual weight of solid matter remaining in the finished water is proportional to that originally present in the river water.
These results are from the records of periods when pumping was constant. Since the course of currents will vary with changing rates, and the velocity and consequent carrying power of currents increases with increase in pumping, the weight of suspended matter carried over the first weir is subject to wide fluctuations. A change in pumping from a rate of 60,000,000 gallons per day to 90,000,000 has increased the suspended matter in the treated water (clear well) by from 2 to 7 parts per 1,000,000, and a further sudden increase to 120,000,000 per day has caused a further rise of 10 to 13 parts, which was carried through the entire series of basins and conduits to the clear well, where 25 to 30 parts of solids in suspension occurred. Slight changes in temperature suffice to alter the course of currents through the basins. In the spring, when the temperature of water in the river and basins is rising, the sludge is less subject to disturbance than in the fall and early winter, when, with tailing temperature, the influent water, more dense than the warmer water of the basins, passes downward over the sludge, causing it to carry over weirs. In the fall, with lowering atmospheric temperatures, the sludge and water in the bottom of the basins are sometimes one degree or more Fahrenheit warmer than surface water of basins and river. Circulation is then effective in changing the course of influent water currents, making them deeper and increasing the scour. The sludge is further subject to disturbance by wave action when high winds prevail, a frequent occurrence in March. In such case the amount of suspended matter carrying over the weir from the filling basin shows a marked increase. Our basins have a working depth of about 14 feet. Similar effects of wave action have been noted in other reservoirs 18 feet deep. The basins at the Chain of Rocks (52 acres) are all uncovered, all used in series, and, therefore, subject to disturbance by each of the agencies affecting their successful working.
Character of Sludge
Suspended matter with the coagulum produced by chemical treatment subsides rapidly, undergoing a change in volume during its accumulation in the bottom of basins. When freshly formed it is loose, disseminated through the full volume of water in which it forms; under average conditions, after an hour and a half it occupies about 3 per cent., and after 24 hours about 2 per cent, of the original volume. After this lapse of time only the newly precipitated portions are disturbed by gentle currents. Opening mud gates at eight-hour intervals seems to reduce the sludge only near the gates, since it follows in a general way the contour of the bottom of the basin and is of such constituency that it does not flow readily over the compact material of earlier subsidence. The tendency is for each new deposit to collect more thickly upon the highest points of previous deposits.
Bacterial purification is proportional to the degree of clarification, falling a little below the percentage removal of suspended solids for the reasons which are cited below :
REMOVAL OF BACTERIA WITH SUSPENDED SOLIDS.
Bacteria entangled in the natural sediment of the water, and gathered into coherent masses with the coagulum, concentrate in the sludge to the extent of 1,000,000 or more per cubic centimeter; are subject to disturbance and dissemination through basin contents by varying currents, however produced, and are easuly carried through later basins to the clear well. It has been observed that slight changes in temperature of the influent water, causing almost infinitesimal differences in the density from that of water at different levels in the basins, give rise to turbid, polluted effluents from each basin in series, as the less compact layers oi sludge are moved; that overturning, which in large bodies of deeper water occurs but twice a year, may occur several times a week when warm and cold days alterate in spring and _____all, and that with a slight rise in suspended matter carried from the older sludge in the filling basin may come an altogether disproportionate increase in bacteria, by reason of alterations in the rate of pumping or the currents produced by high winds sweeping along the surface of a half mile of water. When it is considered that the combined surface of water exposed to winds and temperature changes is more than 52 acres, it will not seem idle to refer to what might at first appear as entirely negligible factors.
EFFECT OF HIGH WINDS ON BACTERIAL COUNTS.
“Bacteria of the B. coli group present.
An abrupt change in the rate of flow through basins may cause the sludge to carry its burden of bacteria through successive basins; e. g., on August 17, 1912, such a change occurred, followed on the 19th by the appearance in the clear well samples of contamination with organisms of the B. coli group. Change in the direction of flow incident to cleaning and restoring a basin to service affects both suspended solids and bacteria per cubic centimeter in the finished water. Basin 1 was thus put in service June 6, 1912. The rise of bacteria in clear well samples was from 150 per cubic centimeter on the 6th to 3.300 in the following week. In this case irregular pumping (at rates ranging from 70,000,000 to 120,000,000 gallons per day) was a factor in producing bad results.
BACTERIAL INCREASE IN CLEAR WELL AFTER BASIN CLEANING.
*Filling basin restored to service after cleaning June 6.
It is apparent that bacterial reductions are subject to disturbance from too many factors to give constant results. We have no safeguard against turbid, contaminated water under these conditions. The table shows the disproportionate increase of bacteria released by stirring previously deposited sludge. River samples usually give evidence of the presence of bacteria of the B. coli group in 1/100 cubic centimeter. The finished water has given counts of more than 30,000 bacteria per cubic centimeter on gelatine at 20 degrees C., and members of the B. coli group have been found in 6 per cent, of tests on 1 cubic centimeter samples in a single month. In reviewing bacteriological results for past years there is noticed a very wide divergence at any one sampling point, and extremely irregular counts for a given period at various points in the clarification system. The utmost we can hope for is a large percentage reduction of bacteria. We can have no assurance that the water which enters the distribution system will he free from pathogenic organisms regularly. While the improvement in the character of the effluent since the introduction of the clarification scheme seems to have reduced the typhoid death rate, the quality is still far from that of a good filter plant.
Residual Solids in Distribution System
Water leaving the clear well contains small quantities of suspended and dissolved iron compounds. small particles of calcium and magnesium compounds, and larger quantities of silt and silicious matter too fine to be deposited during rapid flow through the sedimentation system. The amount of this material daily introduced into the distribution lines during the first year of operation. calculated from suspended matter and daily consumption, was as high as 27.5 tons, averaging 8.7 tons per day for the 10 analyses referred to above.
AVERAGE WEIGHT OF SUSPENDED MATTER IN DAILY SUPPLY.
This material is intermittently discharged from taps over the city in a very irregular way. At the laboratory of the city chemist samples are collected daily for analysis. Comparison of suspended solids in the clear well and at this tap illustrates the extent of this irregular deposition and displacement.
TABLE 7.—INTERMITTENT SEDIMENTATION IN MAINS.
It is a matter of common observation that after unusual draught upon the mains in a portion of the system, very high turbidity appears, local, or affecting large sections of the city, according to the degree of the disturbance. Following a large fire, complaints of turbid water are very numerous. So long as our practice continues sedimentation in the mains, the department cannot resent protests of consumers at turbid water when the accumulated solids are intermittently flushed out.
Incrustation in Distribution System
Because ours is a partially softened water there is always a certain variable amount of calcium carbonate present in the finished product. The softening process is completed slowly at summer temperatures, and in winter is incomplete even when the water passes to the distribution system. There is. therefore, more or less deposit of calcium carbonate in mains and service pipes. Even with high bicarbonate alkalinity in the filling basin, the water leaving the sixth or ninth basin is still supersaturated with calcium carbonate. Connection with a 7-foot steel flow line was made in January, 1908, and the city supply drawn through it for 74 days. Examination at the end of the period disclosed a deposit principally of calcium carbonate. 1/16 inch thick when moist, which sank to 1/32 inch in drying. The water had an average total alkalinity of 59 parts per 1,000,000, of which 25 were due to neutral carbonates. and 34 to bicarbonates. Temperatures ranged from 32 to 47 degrees Fahrenheit. During the earlier years of operation there were notable deposits in meter gears, fishtraps and the like. Two views arc given of the train from a meter set at Vista and Spring avenues, October 26, 1905, and removed November 16, 1909. It is apparent that the heavy incrustation had interfered with the meters’ operation long before the train was removed. This instance—one of many which might be cited—is important in considering the installation of meters in lieu of extending or enlarging the present plant.
While the deposit in the distribution system does not now seem to be increasing rapidly, there is still some incrustation in progress, due to the blending of unequally softened waters. Trouble from this source can be lessened by longer storage. which will equalize the quality of the water before passing the high-service pumps—a very costly expedient: or by so regulating the degree of softening that the finished water shall show a high degree of uniformity—a difficult matter when the raw water is changing quickly. Lower regular velocity through the settling basins would allow longer time for softening reactions; this entails the use of several filling basins, instead of but one with the rest in senes Finally, further reduction of this trouble could be effected by changing the order of chemical treatment, adding lime first and agitating the treated water before tke charge of iron sulphate is applied. There are several methods of arriving at a standard of judgment for considering the efficiency of coagulating basins as full and final preparation of water for distribution to consumers: Comparison of the quality of the effluent with respect to clarity and bacterial content (a) with that of the previous supply; (b) with untreated water from day to day, or (c) with the effluent of a well-operated filter plant. During the first year of operation at St. Louis the weight of solid matter carried injo the distribution system averaged 25 parts per 1,000,000—ranging from none to 97 parts; the reduction was about 98 per cent., reckoned on the weight of solids in the raw water. While this was a decided improvement over conditions prior to treatment, the introduction of an average of 8.7 tons of suspended matter per day into the mains left much to be desired if one compares this result with the standard set by a satisfactory filter plant, whose effluent contains no weighable quantity of solid matter in suspension, and shows no visible turbidity or opalescence in bright sunlight. The local standard of purity has advanced from year to year since the clarification system has been in operation. Prior to 1904 sedimented water containing 60 to 450 parts per 1,000,000 of solids in suspension was accepted, if not approved. Compared with that standard, the quality of water furnished in 1904-1905 was excellent. The next year showed a marked improvement, and established in turn a standard for comparison of the succeeding year’s supply. With each subsequent year the quality of water furnished has been progressively better, until 1910-1911. when consumption so far exceeded the plant’s capacity that there was a falling off in the quality of the effluent, although it was still superior ⅛0 that supplied before 1907-1908. The capacity of the present plant must be increased. The extensions and changes made must provide a better effluent than the present public has been educated to demand. It is apparent that the addition of a filter plant to the coagulating basins is essential if the residual sediment is to be finally removed, and the high color which sometimes characterizes the raw water is to be reduced to an acceptable degree. Operations under prevailing conditions in a plant of this size and character do not admit of the close control possible in a filter plant, where the units (filters) are small, subject to immediate supervision and washing, and their output regulated by rate controllers. Furthermore, with filters the solid matter collected with entangled bacteria is quickly and permanently removed from water passing through them; whereas coagulating basins give too frequently only a temporary separation. It is manifestly impossible to interpret the flow through any one basin at will, should the water in it prove unfit for use, without seriously affecting the contents of other adjacent units by altering the course of currents through them. Nor is it possible to wholly eliminate the previously accumulated sludge m hlltng and sedimentation basins, so that after cleaning the first water passing through them shall be faultless.
With seven filling basins and six additional new reservoirs for sedimentation, as considered in the report above referred to, there would be effected at best only a percentage reduction of suspended solids and bacteria, with no assurance of a safe, clear, sparkling effluent. Irregularities incident to a plant of this character, where pumping and drawing and the consequent period of sedimentation are subject to wide variations directly affecting the finished water, can be avoided only by filtration. The abridged form of treatment adopted prior to the World’s Fair has, beyond question, served a very useful purpose. The question has arisen whether the present system of partial softening, coagulation and sedimentation shall be extended by adding further sedimentation basins to the already existing plant, or supplemented by a filter plant which shall afford a perfectly clear water at all times, with fairly constant bacterial removals and the possibility of immediate control of operating conditions. The cost of constructing two basins of 40,000,000 gallons capacity each was about $7,500 per 1,000,000 gallons capacity. It is hardlv possible to make the necessary changes in the present basins and add the required reservoir capacity to provide for clarification of the volume of water which will certainly be consumed within the next 10 years at a cost of less than $1,500,000. Whereas, a filter plant added to the present basins with certain changes outlined, comprising 40 filters of a normal rating of 44,000,000 gallons in 24 hours, can be constructed at a cost of about $1,250,000. There are locations where basin construction is cheaper, where prolonged sedimentation in very large reservoirs may be advantageously used in producing an acceptable effluent for a time—until the public forgets its earlier satisfaction with an improved water supply, and clamors for further betterment Before such installations are begun there is need of very careful study of the quality of the untreated water, to show its adaptation to coagulation methods and good judgment as to arrangement of the plant with a view to the ultimate addition of filters.