CHAPTER XVII
FILTRATION AND IRRIGATION
=255. Theory.=—The cycle through which the elements forming organic matter pass from life to death and back to life again has been described in Chapter XIII. It has been shown in Chapter XVI that septic action occupies that portion of the cycle in which the combinations of these elements are broken down or reduced to simpler forms and the lower stages of the cycle are reached. The action in the filtration of sewage builds up the compounds again in a more stable form and almost complete oxidation is attained, dependent on the thoroughness of the filtration. In the filtration of sewage only the coarsest particles of suspended matter are removed by mechanical straining. The success of the filtration is dependent on biologic action. The desirable form of life in a filter is the so-called nitrifying bacteria which live in the interstices of the filter bed and feed upon the organic matter in the sewage. Anything which injures the growth of these bacteria injures the action of the filter. In a properly constructed and operated filter, all matter which enters in the influent, leaves with the effluent, but in a different molecular form. A slight amount may be lost by evaporation and gasification but this is more than made up by the nitrogen and oxygen absorbed from the atmosphere. The nitrifying action in sewage filtration is shown by the analysis of sewage passing through a trickling filter, as given in Tables 86 and 87. It is shown by the reduction of the content of organic nitrogen, free ammonia, oxygen consumed, and the increase in nitrites, nitrates, and dissolved oxygen. The reduction of suspended matter is interrupted periodically when the filter “unloads.” The suspended matter in the effluent is then greater than in the influent.
The nitrifying organisms have been isolated and divided into two groups—_nitrosomonas_, the nitrite formers, and _nitrobacter_, the nitrate formers. Experiments indicate that the growth of the nitrobacter organisms is dependent on the presence of the nitrosomonas organisms, which are in turn dependent on the presence of the putrefactive compounds resulting from the action of putrefying bacteria. The existence of these organisms is an example of symbiotic action in bacterial growth. The organisms have been found to grow best on rough porous material on which their zoögleal jelly can be easily deposited and affixed. Sewage filters were constructed to provide these ideal conditions before the action of a filter was thoroughly understood.
The action in irrigation is similar to that in filtration. Although more strictly a method of final disposal rather than preliminary treatment, the similarity of the actions which take place, and the grading of sand filtration into broad irrigation with no distinct line of difference has resulted in the inclusion of the discussion of irrigation in the same chapter with filtration.
=256. The Contact Bed.=—A contact bed is a water-tight basin filled with coarse material, such as broken stone, with which sewage and air are alternately placed in contact in such a manner that oxidation of the sewage is effected. A contact bed has some of the features of a sedimentation tank and an oxidizing filter. As such it marks a transitory step from anaërobic to aërobic treatment of sewage. A plan and a section of a contact bed are shown in Fig. 166.
Because of its dependence on biologic action a contact bed must be ripened before a good effluent can be obtained. The ripening or maturing occurs progressively during the first few weeks of operation, the earlier stages being more rapidly developed. The time required to reach such a stage of maturity that a good effluent will be developed will vary between one and six or eight weeks, dependent on the weather and the character of the influent. During the period of maturing the load on the bed should be made light.
The use of contact beds has been extensive where a more stable effluent than could be obtained from tank treatment has been desired, yet the best quality of effluent was not required. The sewage to undergo treatment in a contact bed should be given a preliminary treatment to remove coarse suspended matter. The efficiency of the contact treatment can be increased by passing the sewage through two or three contact beds in series. In double contact treatment the primary beds are filled with coarser material and operate at a more rapid rate than the secondary beds. Double contact gives better results than single contact, but triple contact treatment, though showing excellent results, is hardly worth the extra cost. An advantage which contact treatment has over all other methods of sewage filtration is that the bed can be so operated that the sewage is never exposed to view. As a result the odors from well-operated contact beds are slight or are entirely absent and there should be no trouble from flying insects. Such a method of treatment is favorable to plants located in populous districts and to the fancies of a landscape architect. Another advantage of the contact bed is the small amount of head required for its operation, which may be as low as 4 to 5 feet. This low head consumption by a sewage filter is equaled only by the intermittent sand filter.
The quality of the effluent from some contact beds is shown in Table 85. It is to be noted that nitrification has been carried to a fair degree of completion, and that the reduction of oxygen consumed has been marked. In comparison with the effluent from filters, contact effluent contains a smaller amount of nitrogen as nitrites and nitrates, and suspended solids. Contact effluent is usually clear and odorless, but it is not stable without dilution. The absence of nitrites and nitrates is sometimes advantageous as the effluent will not support vegetable growths dependent on this form of nitrogen. The absence of suspended solids obviates the use of secondary sedimentation basins which are needed with trickling filters. The head of 5 to 8 feet required for contact treatment is low in comparison to the 10 to 15 feet required for trickling filters, but is slightly higher than the head required for intermittent sand filtration. The cost of contact treatment is higher than the cost of trickling filters but is lower than the cost of intermittent sand filtration, as shown in Table 90.
TABLE 85
QUALITY OF EFFLUENTS FROM CONTACT BEDS
Report on Sewage Purification at Columbus, Ohio, by G. A. Johnson, 1905.
──────┬──────┬─────────┬───────┬────────┬───────────────────────────────── Filter│Depth,│ Size of │ Rate, │ Oxygen │ Nitrogen as │ Feet │Material │Million│Consumed│ │ │in Inches│Gallons│ │ │ │ │ per │ │ │ │ │ Acre │ │ │ │ │per Day│ │ ──────┼──────┼─────────┼───────┼────────┼───────┬───────┬────────┬──────── │ │ │ │ │Organic│ Free │Nitrites│Nitrates │ │ │ │ │ │Ammonia│ │ ──────┼──────┼─────────┼───────┼────────┼───────┴───────┴────────┴──────── │ │ │ │ │ Parts per Million │ │ │ │ │ │ │ │ A │ 5│0.25–1.00│ 0.953│ 23│ 3.5│ 8.7│ 0.20│ 1.6 B │ 5│0.25–2.00│ 1.514│ 21│ 4.0│ 8.4│ 0.15│ 1.4 C │ 5│0.25–1.50│ 1.222│ 24│ 3.5│ 10.8│ 0.11│ 0.6 D │ 5│0.50–1.50│ 1.405│ 22│ 3.3│ 9.5│ 0.13│ 0.9 │ │ │ │ │ │ │ │ │ │ │Per Cent Removal of Constituents of Applied Sewage │ │ │ │ │ │ │ │ A │ 5│0.25–1.00│ 0.953│ 48│ 49│ 10│ │ B │ 5│0.25–2.00│ 1.514│ 52│ 40│ 11│ │ C │ 5│0.25–1.50│ 1.222│ 47│ 31│ 12│ │ D │ 5│0.50–1.50│ 1.405│ 46│ 37│ 19│ │ ──────┴──────┴─────────┴───────┴────────┴───────┴───────┴────────┴────────
──────┬──────┬────────────────────┬───────── Filter│Depth,│ Suspended Matter │Dissolved │ Feet │ │ Oxygen │ │ │ │ │ │ │ │ │ │ │ │ ──────┼──────┼─────┬────────┬─────┼───────── │ │Total│Volatile│Fixed│ │ │ │ │ │ ──────┼──────┴─────┴────────┴─────┴───────── │ Parts per Million │ │ │ │ │ A │ 5│ 832│ 94│ 737│ 0.3 B │ 5│ 831│ 85│ 746│ 0.1 C │ 5│ 826│ 92│ 734│ 0.8 D │ 5│ 810│ 91│ 717│ 0.9 │ │ │ │ │ │ Per Cent Removal of Constituents of Applied Sewage │ │ │ │ │ A │ 5│ 73│ 70│ 76│ B │ 5│ 80│ 77│ 83│ C │ 5│ 70│ 70│ 70│ D │ 5│ 67│ 61│ 72│ ──────┴──────┴─────┴────────┴─────┴─────────
The depth of the contact bed is generally made from 4 to 6 feet. The deeper beds are less expensive per unit of volume, to construct, as the cost of the underdrains and the distribution system is reduced in relation to the capacity of the filter. The increased depth reduces the aëration, and the periods of filling and emptying are so increased as to limit the depths to the figures stated. The other dimensions of the bed are controlled by economy and local conditions, as the success of the contact treatment is not affected by the shape of the bed. Contact units are seldom constructed larger than one-half an acre in area, as larger beds require too much time for filling and emptying. A large number of small units is also undesirable because of the increased difficulty of control. In general it is well to build as large units as are compatible with efficient operation, elasticity of plant, and which can be filled within the time allowed at the average rate of sewage flow, or from dosing tanks in which the storage period is not so long as to produce septic conditions.
The interstices in a contact bed will gradually fill up, due to the deposition of solid matter on the contact material, the disintegration of the material, and the presence of organic growths. The period of rest allowed every five or six weeks tends to restore partially some of this lost capacity through the drying of the organic growths. It is occasionally necessary to remove the material from the bed and wash it in order to restore the original capacity. It may be necessary to do this three or four times a year, in an overloaded plant, or as infrequently as once in five or six years in a more lightly loaded bed. The period is also dependent on the character of the contact material and the quality of the influent. This loss of capacity may reduce the voids from an original amount of 40 to 50 per cent of voids to 10 to 15 per cent. If the bed is not overloaded the loss of capacity will not increase beyond these figures.
The rate of filtration depends on the strength of the sewage, the character of the contact material, and the required effluent. It should be determined for any particular plant as the result of a series of tests. For the purposes of estimation and comparison the approximate rate of filtration should be taken at about 94 gallons per cubic yard of filtering material per day on the basis of three complete fillings and emptyings of the tank. This is equivalent to 150,000 gallons per acre foot of depth per day, or for a bed 5 feet deep to a rate of 750,000 gallons per acre per day. The net rate for double or triple filtration is less than these figures, but on each filter the rates are higher.
The material of the contact bed should be hard, rough, and angular. It should be as fine as possible without causing clogging of the bed. Materials in successful use are: crushed trap rock or other hard stone, broken bricks, slag, coal, etc. Soft crumbling materials such as coke are not suitable as the weight of the superimposed material and the movement of the sewage crushes and breaks it into fine particles which accumulate in the lower portion of the filter and clog it. Roughness, porosity, and small size are desirable, as the greater the surface area the more rapid the deposition of material. After a short time, however, the advantages of roughness and porosity are lost, as the sediment soon covers all unevenness alike. The minimum size of the material is limited by the tendency towards clogging. The sizes in successful use vary between ¼ and ¾ of an inch, ½ inch being a common size. The same size of material is used throughout the depth of the bed except that the upper 6 inches may be composed of small white pebbles or other clean material, which does not come in contact with the sewage and which will give an attractive appearance to the plant. In double or triple contact beds 3 or 4–inch material is sometimes used for the primary beds, and ¼-inch material in the final bed.
Sewage may be applied at any point on or below the surface. The sewage is withdrawn from the bottom of the bed. It is undesirable to have too few inlet or outlet openings as the velocity of flow about the openings will be so great as to disturb the deposit on the contact material. The distribution system and the underdrains for the bed at Marion, Ohio, are shown in Fig. 166.
The cycle of operation of a contact bed is divided into four periods. A representative cycle might be: time of filling, one hour; standing full, 2 hours; emptying, one hour; standing empty, 4 hours. The length of these periods is the result of long experience based on many tests and are an average of the conclusions reached. Wide variations from them may be found in different plants, and tests may show successful results with different periods. The combination of these four periods is known as the contact cycle.
The period of filling should be made as short as possible without disturbing the material of the bed nor washing off the accumulated deposits. The sewage should not rise more rapidly than one vertical foot per minute. During the contact or standing full period sedimentation and adsorption of the colloids are occurring on the area of surface exposed to the sewage. This period should be of such length that septic action does not become pronounced, and long enough to permit of thorough sedimentation. The period of emptying should be made as short as possible without disturbing the bed, on the same basis that the period of filling is determined. During the period of standing empty, air is in contact with the sediment deposited in thin layers on the contact material, and the oxidizing activities of the filter are taking place. The filter is given a rest period of one or two days every five or six weeks, in order that it may increase its capacity and its biologic activity.
The control of a contact bed may be either by hand or automatic, the latter being the more common. Hand control requires the constant attention of an operator and results in irregularity of operation, whereas automatic control will require inspection not more than once a day and insures regularity of operation. A number of automatic devices have been invented which give more or less satisfaction. The air-locked automatic siphons, without moving parts, have proven satisfactory and are practically “fool-proof.” The operation of these devices is explained in Chapter XXI.
=257. The Trickling Filter.=—A trickling or sprinkling filter is a bed of coarse, rough, hard material over which sewage is sprayed or otherwise distributed and allowed to trickle slowly through the filter in contact with the atmosphere. A general view of a trickling filter in operation at Baltimore is shown in Fig. 167. The action of the trickling filter is due to oxidation by organisms attached to the material of the filter. The solid organic matter of the sewage deposited on the surface of the material, is worked over and oxidized by the aërobic bacteria, and is discharged in the effluent in a more highly nitrified condition. At times the discharge of suspended matter becomes so great that the filter is said to be unloading. The action differs from that in a contact bed in that there is no period of septic or anaërobic action and the filter never stands full of sewage.
The effluent from a trickling filter is dark, odorless, and is ordinarily non-putrescible. Analyses of typical effluents are given in Tables 86 and 87. The unloading of the filter may occur at any time, but is most likely to occur in the spring or in a warm period following a period of low temperatures. It causes higher suspended matter in the effluent than in the influent and may render the effluent putrescible. The action is marked by the discharge of solid matter which has sloughed off of the filter material and which increases the turbidity of the effluent. Where the diluting water is insufficient to care for the solids so carried in the effluent, they can be removed by a 2–hour period of sedimentation. The effluent may become septic during this time, however. The nitrogen in the effluent is almost entirely in the form of nitrates, and the percentage of saturation with dissolved oxygen is high. The effluent is more highly nitrified than that from a contact bed, and its relative stability is also higher, thus demanding a smaller volume of diluting water.
The principal advantage of a trickling filter over other methods of treatment is its high rate which is from two to four times faster than a contact bed, and about seventy times faster than an intermittent sand filter. The greatest disadvantage is the head of 12 to 15 feet or more necessary for its operation. Sedimentation of the effluent is usually necessary to remove the settleable solids. During the period of secondary sedimentation the quality of the filter effluent may deteriorate in relative stability. In winter the formation of ice on the filter results in an effluent of inferior quality, but as the diluting water can care for such an effluent at this time the condition is not detrimental to the use of the trickling filter. In summer the filters sometimes give off offensive odors that can be noticed at a distance of half a mile, and flying insects may breed in the filter in sufficient quantities to become a nuisance if preventive steps are not taken. The dissemination of odors is especially marked when treating a stale or septic sewage. The treatment of a fresh sewage seldom results in the creation of offensive odors.
TABLE 86
ANALYSIS OF CRUDE SEWAGE, IMHOFF TANK, AND SPRINKLING FILTER EFFLUENTS AT ATLANTA, GEORGIA
(Engineering Record, Vol. 72, p. 4)
─────────┬───────────┬────────────────────────────────────────── │Temperature│ Parts per Million │Fahrenheit │ │ │ │ │ ─────────┼───────────┼─────────────────────────────────┬──────── │ │ Nitrogen as │ Oxygen │ │ │Consumed ─────────┼───────────┼───────┬───────┬────────┬────────┼──────── │ │Organic│ Free │Nitrites│Nitrates│ │ │ │Ammonia│ │ │ ─────────┴───────────┴───────┴───────┴────────┴────────┴────────
_Crude Sewage_
─────────┬───────────┬───────┬───────┬────────┬────────┬──────── 1913 │ │ │ │ │ │ Maximum │ 77│ 15.6│ 21.8│ 0.1│ 3.0│ 100.0 Minimum │ 61│ 10.4│ 16.5│ 0.1│ 1.4│ 78.3 Average │ 70│ 12.8│ 18.8│ 0.1│ 2.2│ 90.6 1914 (7 │ │ │ │ │ │ months)│ │ │ │ │ │ Maximum │ 74│ 16.0│ 33.4│ │ 2.3│ Minimum │ 60│ 9.5│ 18.1│ │ 1.6│ Average │ 66│ 13.4│ 27.1│ │ 2.0│ ─────────┴───────────┴───────┴───────┴────────┴────────┴────────
_Imhoff Effluent_
─────────┬───────────┬───────┬───────┬────────┬────────┬──────── 1913 │ │ │ │ │ │ Maximum │ 78│ 13.2│ 21.9│ 0.2│ 3.1│ 68.0 Minimum │ 58│ 6.5│ 16.8│ 0.1│ 1.1│ 53.1 Average │ 68│ 9.0│ 20.0│ 0.2│ 2.1│ 60.1 1914 (7 │ │ │ │ │ │ months)│ │ │ │ │ │ Maximum │ 77│ 10.3│ 30.3│ │ 2.0│ Minimum │ 59│ 4.1│ 18.0│ │ 1.5│ Average │ 65│ 7.7│ 25.9│ │ 1.8│ ─────────┴───────────┴───────┴───────┴────────┴────────┴────────
_Sprinkling Filter Effluent_
─────────┬───────────┬───────┬───────┬────────┬────────┬──────── 1913 │ │ │ │ │ │ Maximum │ 79│ 5.6│ 14.2│ 0.8│ 11.3│ 32.1 Minimum │ 55│ 2.6│ 6.2│ 0.5│ 5.8│ 23.6 Average │ 66│ 3.8│ 9.9│ 0.7│ 8.2│ 28.2 1914 (7 │ │ │ │ │ │ months)│ │ │ │ │ │ Maximum │ 77│ 8.5│ 20.7│ │ 11.2│ Minimum │ 55│ 4.4│ 8.8│ │ 3.6│ Average │ 63│ 5.7│ 15.2│ │ 7.2│ ─────────┴───────────┴───────┴───────┴────────┴────────┴────────
─────────┬────────────────────┬──────────┬───────── │ Parts per Million │ Per Cent │Relative │ │Saturation│Stability │ │Dissolved │ │ │ Oxygen │ ─────────┼────────────────────┼──────────┼───────── │ Suspended Matter │ │ │ │ │ ─────────┼─────┬────────┬─────┼──────────┼───────── │Total│Volatile│Fixed│ │ │ │ │ │ │ ─────────┴─────┴────────┴─────┴──────────┴─────────
_Crude Sewage_
─────────┬─────┬────────┬─────┬──────────┬───────── 1913 │ │ │ │ │ Maximum │ 371│ 154│ 163│ 47│ Minimum │ 222│ 98│ 112│ 11│ Average │ 285│ 126│ 138│ 28│ 1914 (7 │ │ │ │ │ months)│ │ │ │ │ Maximum │ 431│ │ │ 48│ Minimum │ 279│ │ │ 12│ Average │ 351│ │ │ 30│ ─────────┴─────┴────────┴─────┴──────────┴─────────
_Imhoff Effluent_
─────────┬─────┬────────┬─────┬──────────┬───────── 1913 │ │ │ │ │ Maximum │ 90│ 50│ 41│ │ Minimum │ 35│ 42│ 21│ │ Average │ 68│ 46│ 33│ │ 1914 (7 │ │ │ │ │ months)│ │ │ │ │ Maximum │ 73│ │ │ 48│ Minimum │ 49│ │ │ 34│ Average │ 65│ │ │ 43│ ─────────┴─────┴────────┴─────┴──────────┴─────────
_Sprinkling Filter Effluent_
─────────┬─────┬────────┬─────┬──────────┬───────── 1913 │ │ │ │ │ Maximum │ 60│ 31│ 28│ 76│ 99 Minimum │ 33│ 26│ 28│ 52│ 88 Average │ 49│ 28│ 28│ 64│ 89 1914 (7 │ │ │ │ │ months)│ │ │ │ │ Maximum │ 106│ │ │ 79│ 99 Minimum │ 40│ │ │ 55│ 89 Average │ 62│ │ │ 65│ 95 ─────────┴─────┴────────┴─────┴──────────┴─────────
TABLE 87
EFFICIENCY OF SPRINKLING FILTER CHICAGO, ILLINOIS
Depth of Filter 9 feet. Size of stone 2 in. to 3 in.
────────┬───────────────────────────┬─────────────────────────── Month │ Organic Nitrogen │ Free Ammonia │ │ ────────┼───────────────────────────┼─────────────────────────── │ │ ────────┼─────────┬─────────┬───────┼─────────┬─────────┬─────── │Influent,│Effluent,│ Per │Influent,│Effluent,│ Per │Parts per│Parts per│ Cent │Parts per│Parts per│ Cent │ Million │ Million │Removed│ Million │ Million │Removed ────────┼─────────┼─────────┼───────┼─────────┼─────────┼─────── 1910 │ │ │ │ │ │ October │ 5.1│ 2.8│ 45│ 12.0│ 4.6│ 62 November│ 5.9│ 2.5│ 58│ 12.0│ 5.9│ 51 December│ 4.6│ 3.0│ 35│ 12.0│ 6.9│ 42 │ │ │ │ │ │ 1911 │ │ │ │ │ │ January │ 6.3│ 4.8│ 24│ 11.0│ 7.0│ 36 February│ 9.0│ 4.8│ 47│ 10.0│ 7.2│ 28 March │ 8.3│ 3.5│ 58│ 9.9│ 6.4│ 35 April │ 6.4│ 4.0│ 37│ 8.3│ 3.6│ 69 May │ 7.6│ 5.4│ 29│ 9.2│ 2.4│ 74 June │ 5.9│ 3.2│ 46│ 11.0│ 0.6│ 95 July │ 6.2│ 4.2│ 32│ 11.0│ 1.3│ 88 ────────┴─────────┴─────────┴───────┴─────────┴─────────┴───────
────────┬───────────────────────────┬─────────────────────────── Month │ Oxygen Consumed │ Nitrites │ │ ────────┼───────────────────────────┼─────────────────────────── │ │ ────────┼─────────┬─────────┬───────┼─────────┬─────────┬─────── │Influent,│Effluent,│ Per │Influent,│Effluent,│ Per │Parts per│Parts per│ Cent │Parts per│Parts per│ Cent │ Million │ Million │Removed│ Million │ Million │Removed ────────┼─────────┼─────────┼───────┼─────────┼─────────┼─────── 1910 │ │ │ │ │ │ October │ 30│ 15│ 50│ │ .90│ November│ 35│ 15│ 57│ │ .76│ December│ 39│ 20│ 49│ .07│ .45│ 6.4 │ │ │ │ │ │ 1911 │ │ │ │ │ │ January │ 42│ 20│ 52│ .08│ .15│ 1.9 February│ 46│ 20│ 56│ .09│ .15│ 1.7 March │ 47│ 21│ 56│ .09│ .15│ 1.7 April │ 38│ 21│ 45│ .16│ .21│ 1.3 May │ 33│ 31│ 6│ .08│ .38│ 4.8 June │ 28│ 16│ 43│ .00│ .30│ ∞ July │ 34│ 26│ 24│ .00│ .36│ ∞ ────────┴─────────┴─────────┴───────┴─────────┴─────────┴───────
────────┬───────────────────────────┬─────────────────────────── Month │ Nitrates │ Dissolved Oxygen │ │ ────────┼───────────────────────────┼─────────────────────────── │ │ ────────┼─────────┬─────────┬───────┼─────────┬─────────┬─────── │Influent,│Effluent,│ Per │Influent,│Effluent,│ Per │Parts per│Parts per│ Cent │Parts per│Parts per│ Cent │ Million │ Million │Removed│ Million │ Million │Removed ────────┼─────────┼─────────┼───────┼─────────┼─────────┼─────── 1910 │ │ │ │ │ │ October │ │ 7.8│ │ 0.0│ 8.5│ ∞ November│ │ 5.9│ │ 0.0│ 8.1│ ∞ December│ .15│ 2.6│ 17│ 2.0│ 8.4│ 4.2 │ │ │ │ │ │ 1911 │ │ │ │ │ │ January │ .27│ 2.2│ 8.2│ 3.0│ 7.8│ 2.9 February│ .50│ 2.6│ 5.2│ 2.6│ 8.0│ 3.1 March │ .34│ 3.2│ 9.4│ 2.2│ 6.6│ 3.0 April │ .53│ 4.5│ 8.5│ 2.1│ 7.1│ 3.4 May │ .15│ 7.5│ 4.3│ 0.1│ 7.7│ 77 June │ .16│ 8.3│ 5.2│ 0.0│ 7.6│ ∞ July │ .09│ 7.7│ 8.0│ 0.0│ 6.5│ ∞ ────────┴─────────┴─────────┴───────┴─────────┴─────────┴───────
────────┬───────────┬─────────────────────────────────────────────────────── Month │ Per Cent │ Suspended Matter │Putrescible│ ────────┼───────────┼───────────────────────────┬─────────────────────────── │ │ Total │ Volatile ────────┼───────────┼─────────┬─────────┬───────┼─────────┬─────────┬─────── │ │Influent,│Effluent,│ Per │Influent,│Effluent,│ Per │ │Parts per│Parts per│ Cent │Parts per│Parts per│ Cent │ │ Million │ Million │Removed│ Million │ Million │Removed ────────┼───────────┼─────────┼─────────┼───────┼─────────┼─────────┼─────── 1910 │ │ │ │ │ │ │ October │ 0│ 75│ 40│ 47│ 54│ 25│ 54 November│ 5│ 61│ 16│ 74│ 52│ 15│ 71 December│ 35│ 85│ 40│ 53│ 60│ 26│ 57 │ │ │ │ │ │ │ 1911 │ │ │ │ │ │ │ January │ 38│ 112│ 43│ 63│ 68│ 29│ 57 February│ 29│ 100│ 49│ 51│ 64│ 32│ 50 March │ 28│ 106│ 37│ 65│ 63│ 22│ 65 April │ 9│ 113│ 68│ 40│ 59│ 35│ 41 May │ 6│ 88│ 150│ _1.7_│ 54│ 70│ _1.3_ June │ 1│ 92│ 77│ 18│ 56│ 36│ 36 July │ 4│ 155│ 130│ 16│ 74│ 61│ 18 ────────┴───────────┴─────────┴─────────┴───────┴─────────┴─────────┴───────
────────┬─────────────────────────── Month │ Suspended Matter │ ────────┼─────────────────────────── │ Fixed ────────┼─────────┬─────────┬─────── │Influent,│Effluent,│ Per │Parts per│Parts per│ Cent │ Million │ Million │Removed ────────┼─────────┼─────────┼─────── 1910 │ │ │ October │ 21│ 15│ 29 November│ 9│ 1│ 89 December│ 25│ 14│ 44 │ │ │ 1911 │ │ │ January │ 44│ 13│ 70 February│ 37│ 17│ 53 March │ 43│ 15│ 65 April │ 54│ 33│ 39 May │ 34│ 80│ _2.4_ June │ 36│ 41│ _1.1_ July │ 81│ 69│ 15 ────────┴─────────┴─────────┴───────
NOTE.—Italic figures represent increases.
Raw sewage cannot be treated successfully on a trickling filter. Coarse solid particles should be screened and settled out, in order that the distributing devices or the filter may not become clogged. The effluent from an Imhoff tank has proven to be a satisfactory influent for a trickling filter. A septic tank effluent may be so stale as to be detrimental to the biologic action in the filter.
In the operation of a trickling filter the sewage is sprayed or otherwise distributed as evenly as possible in a fine spray or stream, over the top of the filtering material. The sewage then trickles slowly through the filter to the underdrains through which it passes to the final outlet. The distribution of the sewage on the bed is intermittent in order to allow air to enter the filter with the sewage. The cycle of operation should be completed in 5 to 15 minutes, with approximately equal periods of rest and distribution. Cycles of too great length will expose the filter to drying or freezing and will give poorer distribution throughout the filter. Cycles which are too short will operate successfully only with but slight variation in the rate of sewage flow. In some plants it has been found advantageous to allow the filters to rest for one day in 3 to 6 weeks or longer, dependent on the quality of the effluent.
The rate of filtration may be as high as 2,000,000 gallons per acre per day, which is equivalent to 200 gallons per cubic yard of material per day in a bed 6 feet deep. This is more than double the rate permissible in a contact bed. The exact rate to be used for any particular plant should be determined by tests. It is dependent on the quality of the sewage to be treated, on the depth of the bed, the size of the filling material, the weather, and other minor factors.
The filtering material is similar to that used in a contact bed. It should consist of hard, rough, angular material, about 1 to 2 inches in size. Larger sizes will permit more rapid rates of filtration, but will not produce so good an effluent. Smaller sizes will clog too rapidly.
The depth of the filter is limited by the possibility of ventilation and the strength of the filtering material to withstand crushing. The deeper the bed the less the expense of the distribution and collecting system for the same volume of material, and the more rapid the permissible rate of filtration. The depths in use vary between 6 and 10 feet, with 6 to 8 feet as a satisfactory mean. From a biologic standpoint the action of the filter seems to be proportional to the volume of the filtering material and therefore proportional to the depth of the bed, being limited to a minimum depth of about 5 feet, below which sewage may pass through the filter without treatment. The shape and other dimensions of the filter depend on the local conditions and the economy of construction. The filters need not be broken up into units by water-tight dividing walls. One filter can be constructed sufficient for all needs and various portions of it can be isolated as units by the manipulation of valves in the distribution system. Ventilation is provided by the air entrained with the sewage as it falls upon the surface. If the sides of the filter are built of open stone crib work the ventilation will be greatly improved, but it will not be possible to flood the filters to keep down flies, and in cold climates these openings must be covered in winter to prevent freezing. Filters have been constructed without side walls, the filtering material being allowed to assume its natural angle of repose. This has usually been found to be more expensive than the construction of side retaining walls, due to the unused filling material and the extra underdrains required.
The distribution of sewage is ordinarily effected by a system of pipes and spray nozzles as shown in Fig. 168 and 169. Other methods of distribution have been used. At Springfield, Mo.,[160] a moving trough from which the sewage flows continuously is drawn back and forth across the bed by means of a cable. In England circular beds have been constructed and the sewage distributed on them through revolving perforated pipes. At the Great Lakes Naval Training Station[161] the distributing pipes in the plant, now abandoned, were supported above the surface of the filter. The sewage fell from holes in the lower side of these pipes on to brass splash plates 14 inches above the filter. It was deflected horizontally from these plates over the filter surface. Pipes and spray nozzles have been adopted almost universally in the United States. Splash plates, traveling distributors, and other forms of distribution have been used only in exceptional cases. In a distributing system consisting of pipes and nozzles, a network of pipes is laid out somewhat as shown in Fig. 168, in such a manner that the head loss to all points is approximately equal. The number of valves required should be reduced to a minimum. The pipes may be laid out with the main feeders leading from a central point and branches at right angles to them, somewhat on the order of a spider’s web, or they may be laid out on a rectangular or gridiron system. The radial system is advantageous because of the central location of the control house, but it does not always lend itself favorably to the local conditions, and the piping and nozzle location are not so simple. The gridiron system lends itself favorably to the equalization of head losses. The pipes used should be larger than would be demanded by considerations of economy alone, both for the purpose of reduction of head loss and ease in cleaning. No pipe less than 6 inches in diameter should be used, and the average velocity of flow should not exceed one foot per second. Cast-iron, concrete, or vitrified clay pipe may be used, but cast iron is the material commonly used. The system should be arranged for easy flushing and cleaning and the pipes so sloped that the entire system can be drained in case of a shut down in cold weather.
The pipes are placed far enough below the surface of the filling material so that the top of the spraying nozzle is 6 to 12 inches above the surface of the filter. If the pipes are placed near the surface they are accessible for repairs, but are exposed to temperature changes. If the pipes are large their presence near the surface of the filter may seriously affect the distribution of the sewage through the filter. If the distributing pipes are placed near the bottom of the filter they are inaccessible for repairs and the nozzles must be connected to them by means of long riser pipes. The distributing pipes should be supported by columns extending to the foundation of the filter bed, there being a column at every pipe joint with such intermediate supports as may be required. In some plants the pipes have been supported by the filtering material. Although slightly less expensive in first cost the practice of so supporting the pipes is poor, as settling of the material may break the pipe or cause leaks, and if the bed becomes clogged, removal of the material is made more difficult. Valves should be placed in the distributing system in such a manner that different sets of nozzles can be cut out at will, thus resting those portions of the filter and permitting repairs without shutting down the entire filter.
The spacing of the nozzles is fixed by the type and size of the nozzle, the available head, and the rate of filtration. Various types of sprinkler nozzles are shown in Fig. 169 and the discharge rates, head losses, and distances to which sewage is thrown for the Taylor nozzles, are shown in Fig. 170. Nozzles are available which will throw circular, square, or semicircular sprays. In the use of circular sprays there is necessarily some portion of the filter which is underdosed if the nozzles are placed at the corners of squares with the sprays tangent, and there is an overdosing of other portions if the sprays are allowed to overlap so that no portion of the filter is left without a dose. Rectangular sprays will apparently overcome these difficulties, but studies have shown that circular sprays with some overlapping, and the nozzles placed at the apexes of equilateral triangles as shown in Fig. 172 will give as satisfactory distribution as other forms.
The nozzles should be selected to give the best distribution, to consume all of the head available, and to give the proper cycle of operation. The entire head available should be consumed in order that the fewest number of nozzles may be used. An excellent study of the characteristics of various types of nozzles has been published in Bulletin No. 3 of the Engineering Experiment Station at Purdue University, 1920. As a result of the tests on the nozzles shown in Fig. 169, it was determined for all nozzles, except No. 8, that
_Q_ = _Ca_√(2_gh_);
in which _Q_ = the rate of discharge in cubic feet per second;
_C_ = a coefficient shown in Table 88;
_a_ = the net cross-sectional opening of the nozzle in square feet;
_h_ = the pressure on the nozzle in feet of water.
TABLE 88
COEFFICIENTS OF DISCHARGE FOR SPRINKLER NOZZLES SHOWN IN FIG. 169
──────────────────────┬──────┬──────┬──────┬──────┬──────┬──────┬────── Nozzle Number │ 1 │ 2 │ 3 │ 4 │ 5 │ 6 │ 7 ──────────────────────┼──────┼──────┼──────┼──────┼──────┼──────┼────── Coefficient │ .648 │ .756 │ .696 │ .666 │ .675 │ .598 │ .569 ──────────────────────┴──────┴──────┴──────┴──────┴──────┴──────┴──────
It is evident that if the head on the nozzles is constant and the nozzle throws a circular spray, the intensity of dosing at the circumference will be greater than nearer the center. This difficulty is overcome by so designing the dosing tank from which the sewage is fed that the head on the nozzle and the quantity thrown will vary in such a manner that the distribution over the bed is equalized. Intermittent action is obtained by an automatic siphon which commences to discharge when the tank is full and empties the tank in the period allowed for dosing. Under such conditions the tank should discharge for a longer time at the higher heads than at the lower heads as there is more territory to be covered at the higher heads. The design of the tank to do this with exactness is difficult, and the construction of the necessary curved surfaces is expensive. Where a dosing tank is used for such conditions it has been found satisfactory to construct the tank with plane sides sloping at approximately 45 degrees from the vertical (or horizontal). A tank with curved surfaces is shown in Fig. 171. The dosing siphon is usually placed in the tank as shown in the figure. The head and quantity of discharge through the nozzles can be varied also by maintaining a constant depth in a dosing tank by means of a float feed valve, and varying the head and quantity discharged to the nozzles by a butterfly valve in the main feed line, or by the use of a Taylor undulating valve designed for this purpose. The butterfly valve is opened and closed by a cam so designed and driven at such a rate that the required distribution is obtained. The Taylor undulating valve is opened and closed at a constant rate, the shape of the valve giving the required variations in head and discharge. Other methods of control have been attempted but have not been used extensively.
An example of the design of the nozzle layout and dosing tank for a sprinkling filter follows:
Let it be required to determine the nozzle layout for one acre of sprinkling filters with 5 feet available head on the nozzles.
The selection of the type of nozzle and the size of opening is a matter of judgment and experience. Nozzles with large openings are less liable to clog and fewer nozzles are needed than where small nozzles are used, but the distribution of sewage is not so even as with the use of small nozzles. In this example Taylor circular spray nozzles will be selected. Fig. 170 shows that a Taylor circular spray nozzle will discharge 22.3 g.p.m. under a head of 5 feet, and that the economical nozzle spacing will be 15.3 feet. The least number of nozzles at this spacing required for a bed of one acre in area is found as follows: In Fig. 172, let _n_ equal the number of nozzles in a horizontal row, counting half-spray nozzles as ½, and let _m_ equal the number of rows counting rows of half-spray nozzles as half rows.[162] Then the number of nozzles, _N_, equals _mn_, and 15.3_m_ × 13.2_n_ equals 43,560 or _mn_ equals 215.
The next step should be the design of the dosing tank and siphon. It is possible to design a tank which will give equal distribution over equal areas of filter surface. It has been found, however, that the expense of this refinement is unwarranted as there are a number of outside factors which tend to overcome the theoretical design. The effect of wind, unequal spacing, and irregularities in the elevation of the nozzles have a tendency to offset refinements in the design of a dosing tank. It is therefore the general practice to slope the sides of the tank at an angle of about 45 degrees as previously stated. The dosing tank is generally designed to have a capacity which will give a complete cycle of operation once in 15 minutes. In the ordinary design the factors given are the rate of inflow and the given time of filling. In the following example the time of filling will be taken as 10 minutes, the time of emptying as 5 minutes, and the rate of flow as 1,000,000 gallons per day. The capacity of the tank will therefore be (1,000,000)⁄24 x 6 = 7,000 gallons. The diameter of the siphon to be selected can be computed as follows:
Let _Q_ = the capacity of the tank in cubic feet; _q__{1} = the rate of discharge of the siphon in cubic feet per second; _q__{2} = the rate of inflow to the tank in cubic feet per second; _q_ = the rate of emptying the tank in cubic feet per second = (_q__{1} − _q__{2}); _A_ = the cross-sectional area of the free surface of the water in the tank at any instant, in square feet; _a_ = the cross-sectional area of the siphon in square feet; _b_ = the small dimension of the base of the tank in feet; _h_ = the head of water, in feet, on the discharge siphon; _h__{1} = the initial head of water, in feet, on the siphon; _h__{2} = the final head of water in feet, on the siphon; _t_ = the time, in seconds, required to empty the tank,
then _dQ_ = -_Adh_ = _q__{1}_dt_ − _q__{2}_dt_,
and _dt_ = (_dQ_)⁄_q_ = − _Adh_⁄(_q__{1} − _q__{2}),
but _q__{1} = 0.4 _A_ √((2_gh_)),[163]
therefore _t_ = ∫_{_h__{2}}^{_h__{1}} -_Adh_⁄(0.4_a_√(2_gh_) − _q__{2}),
but _A_ = 4_h_^2 + 4_bh_ + _b_^2,
therefore _t_ = ∫_{_h__{1}}^{_h__{2}} ((_b_^2 + 4_bh_ + 4_h_^2)_dh_)⁄0.4_a_√(2_gh_) − _q__{2}.
The integration of this expression is tedious. Its solution for siphons between 6 inches and 12 inches operating under heads commencing from 3 feet to 6 feet, with a time of emptying of 5 minutes and time of filling of 10 minutes is given in Fig. 173. In the example given the rate of inflow is 1.55 sec. feet and the head is 5 feet. Then from Fig. 173 the size of the siphon to be used is 12 inches. Where a siphon of the size required to empty the tank in the time fixed is not available, combinations of available sizes can sometimes be used.
For example, if the given head is 6 feet, and the rate of inflow is 1.4 sec. feet, it is evident from Fig. 173 that a 6,300–gallon dosing tank and two 8–inch siphons will give the required cycle.
The method used for the design of the setting of Taylor nozzles by the Pacific Flush Tank Co., is less rational but more simple and probably as satisfactory. In this method the steps are as follows:
(1) Divide the maximum daily rate of sewage flow by 1,000 to get the maximum minute inflow.
(2) The number of nozzles required is determined by dividing the preceding figure by 6. Generally a Taylor nozzle with an orifice of ⅞ of an inch will discharge about 20 g.p.m. at the high head and about 8 g.p.m. at the low head, and as the nozzles must have a capacity which will take care of the inflow at the low head, the divisor 6 is used as a factor of safety instead of using 8 as the divisor.
(3) The type of nozzle to be used is selected from experience or as a matter of judgment. Circular-spray nozzles are more generally used.
(4) The spacings are determined from Fig. 170.
(5) The dosing tank of the shape described is then designed. The capacity is such as to give a complete cycle once every 15 minutes. The method of this design is similar to that followed previously.
(6) The dosing siphons are designed so that they will have a capacity at the minimum head of from 40 to 50 per cent in excess of the maximum minute inflow, and the draining depth of the siphon will be limited to a maximum of 5 to 5½ feet. The siphons are all made adjustable with a variation of 6 inches or more on either side of the normal discharge line so that the spraying area and cycle can be varied to secure the best results.
The underdrainage of a trickling filter should consist of some form of false bottom such as the types shown in Fig. 174. Where possible the underdrains should be open at both ends for the purpose of ventilation and flushing. It is desirable that the drains be so arranged that a light can be seen through them in order that clogging can be easily located. The drains should be placed on a slope of approximately 2 in 100 towards a main collector. The length of the drains is limited by their capacity to carry the average dose from the area drained by them. The main collecting conduits must be designed in accordance with the hydraulic principles given in Chapter IV. No valves, or other controlling apparatus, are placed on the underdrains or outlets from the filter.
Covers have been provided in winter for some trickling filters in cold climates. The Taylor sprinkling nozzle has been found to work successfully in extremely cold weather, and it is generally accepted that the covering of filters is unnecessary, if the filter is not to be shut down for any length of time in cold weather.
The operation of devices for automatically controlling the operation of a trickling filter is explained in Chapter XXI.
=258. Intermittent Sand Filter.=—An intermittent sand filter is a specially prepared bed of sand, or other fine grained material, on the surface of which sewage is applied intermittently, and from which the sewage is removed by a system of underdrains. It differs from broad irrigation in the character of the material, the care and preparation of the bed, and the thoroughness of the underdrainage. A distinctive feature of the intermittent sand filter is the quality of the effluent delivered by it. In a properly designed and operated plant the effluent is clear, colorless, odorless, and sparkling. It is completely nitrified, is stable and contains a high percentage of dissolved oxygen. It contains no settleable solids except at widely separated periods when a small quantity may appear in the effluent. The percentage removal of bacteria may be from 98 to 99 per cent. Some analyses of sand filter effluents are given in Table 89. The dissolved solids, the remaining bacteria, and the antecedents of the effluent are the only differences between it and potable water. An effluent from an intermittent sand filter is the most highly purified effluent delivered by any form of sewage treatment. The effluent can be disposed of without dilution, on account of its high stability. The treatment of sewage to so high a degree is seldom required, so that the use of intermittent filters is not common. Other drawbacks to their use are the relatively large area of land necessary and the difficulty of obtaining good filter sand in all localities.
TABLE 89
QUALITY OF EFFLUENTS FROM SAND FILTERS
(Report on Sewage Purification at Columbus, Ohio, by G. A. Johnson, 1905) ────────────┬───────────────────────────────────────────────────────┬─────── Source of │ Parts per Million │Rate of Sample │ │Filtra- │ │ tion │ │Gallons │ │ per │ │ Acre, │ │per Day ────────────┼────────────────────────────────────┬────────┬─────────┼─────── │ Nitrogen as │ Oxygen │ Oxygen │ │ │Consumed│Dissolved│ ────────────┼───────┬──────────┬────────┬────────┼────────┼─────────┼─────── │ Free │Albuminoid│Nitrites│Nitrates│ │ │ │Ammonia│ Ammonia │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 11.0 │ 8.6 │ │ │ 59. │ │ influent │ │ │ │ │ │ │ from grit │ │ │ │ │ │ │ chamber │ │ │ │ │ │ │ Filter │ 1.12 │ 0.88 │ 0.08 │ 11.5 │ 6.9 │ 6.3 │ 0.081 effluent │ │ │ │ │ │ │ Filter │ 0.81 │ 0.88 │ 0.10 │ 12.6 │ 6.5 │ 6.2 │ 0.118 effluent │ │ │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 9.7 │ 5.4 │ │ │ 33. │ │ influent │ │ │ │ │ │ │ from plain│ │ │ │ │ │ │ settling │ │ │ │ │ │ │ tank │ │ │ │ │ │ │ Filter │ 0.62 │ 0.77 │ 0.11 │ 14.9 │ 6.0 │ 8.2 │ 0.139 effluent │ │ │ │ │ │ │ Filter │ 0.99 │ 1.10 │ 0.10 │ 12.6 │ 7.8 │ 6.5 │ 0.274 effluent │ │ │ │ │ │ │ Filter │ 2.61 │ 1.39 │ 0.09 │ 9.0 │ 9.7 │ 3.9 │ 0.357 effluent │ │ │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 10.7 │ 5.6 │ │ │ 38. │ │ influent │ │ │ │ │ │ │ from │ │ │ │ │ │ │ septic │ │ │ │ │ │ │ tank │ │ │ │ │ │ │ Filter │ 1.63 │ 1.16 │ 0.09 │ 11.2 │ 8.0 │ 5.8 │ 0.357 effluent │ │ │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 13.4 │ 4.7 │ │ │ 40. │ │ influent │ │ │ │ │ │ │ from coke │ │ │ │ │ │ │ strainer │ │ │ │ │ │ │ Filter │ 2.24 │ 1.35 │ 1.03 │ 14.6 │ 10.1 │ 6.9 │ 0.372 effluent │ │ │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 8.6 │ 3.6 │ 0.19 │ 1.6 │ 24. │ 0.3 │ influent │ │ │ │ │ │ │ from │ │ │ │ │ │ │ contact │ │ │ │ │ │ │ bed │ │ │ │ │ │ │ Filter │ 2.62 │ 1.35 │ 0.31 │ 8.1 │ 8.3 │ 5.8 │ 0.516 effluent │ │ │ │ │ │ │ Filter │ 2.44 │ 2.41 │ 0.16 │ 9.4 │ 12.5 │ 5.0 │ 0.525 effluent │ │ │ │ │ │ │ Filter │ 3.40 │ 1.15 │ 0.20 │ 10.9 │ 9.7 │ 5.2 │ 0.525 effluent │ │ │ │ │ │ │ ────────────┼───────┼──────────┼────────┼────────┼────────┼─────────┼─────── Filter │ 9.0 │ 4.8 │ 0.42 │ 1.3 │ 27. │ 3.4 │ influent │ │ │ │ │ │ │ from │ │ │ │ │ │ │ sprinkling│ │ │ │ │ │ │ filter │ │ │ │ │ │ │ after │ │ │ │ │ │ │ sedimen- │ │ │ │ │ │ │ tation │ │ │ │ │ │ │ Filter │ 2.95 │ 1.25 │ 0.19 │ 7.0 │ 8.8 │ 3.8 │ 0.675 effluent │ │ │ │ │ │ │ Filter │ 4.77 │ 2.63 │ 0.51 │ 4.6 │ 11.8 │ 2.5 │ 0.749 effluent │ │ │ │ │ │ │ Filter │ 3.47 │ 1.61 │ 0.31 │ 7.2 │ 11.9 │ 3.7 │ 1.129 effluent │ │ │ │ │ │ │ ────────────┴───────┴──────────┴────────┴────────┴────────┴─────────┴───────
The action in an intermittent sand filter is more complete than in other forms of filters because a greater surface is exposed to the passage of sewage by the fine sand particles, and the sewage is in contact with the filtering material a longer time due to the lower rate of filtration and the slow velocity of flow through the filter. It is essential that the sewage be applied to the bed intermittently in order that air shall be entrained in the filter. The period between doses should not be so long that the filter becomes dry.
In the operation of an intermittent sand filter one dose per day is considered an ordinary rate of application, although some plants operate with as many as four doses per day per filter, and others on one dose at long and irregular intervals. It is not always necessary to rest the filter for any length of time unless signs of overloading and clogging are shown. The intermittent dosing action may be obtained by the action of an automatic siphon as is described in Chapter XXI. The sewage is distributed on the beds through a number of openings in the sides of distributing troughs resting on the surface of the filter. The sewage is withdrawn from the bottom of the filter through a system of underdrains, into which it enters after its passage through the bed. There are no control devices on the outlet, as the rate of filtration is controlled by the action of the dosing apparatus and the rate at which sewage is delivered to it. The action of the dosing apparatus should respond quickly to variations in sewage flow. As the doses are applied to a sand filter, a mat of organic matter or bacterial zoöglea is formed on the surface of the bed. The mat is held together by hair, paper, and the tenacity of the materials. It may attain a thickness of ¼ to ½ an inch before it is necessary to remove it. So long as the filter is draining with sufficient rapidity this mat need not be removed, but if the bed shows signs of clogging, the only cleaning that may be necessary will be the rolling up of this dried mat. It is believed that the greater portion of the action in the filter occurs in the upper 5 to 8 inches of the bed, but occasionally the beds become so clogged that it is necessary to remove ¾ of an inch to 2 inches of sand in addition to the surface mat, or to loosen up the surface by shallow plowing or harrowing. The necessity for such treatment may indicate that the filter is being overloaded as a result of which the rate of filtration should be decreased or the preliminary treatment should be improved. The plowing of clogging material into the bed should be avoided as under these conditions the final condition of the bed will be worse than its condition when trouble was first observed.
In winter the surface of the bed should be plowed up into ridges and valleys. The freezing sewage forms a roof of ice which rests on the ridges and the subsequent applications of sewage find their way into the filter through the valleys under the ice. In a properly operated bed the filtering material will last indefinitely without change. If a filter is operated at too high a rate, however, although the quality of the effluent may be satisfactory, it will be necessary at some time to remove the sand and restore the filter.
The rate of filtration depends on the character of the influent, the desired quality of the effluent, and the depth and character of the filtering material. Filters can be found operating at rates of 50,000 gallons per acre per day and others at eight times this rate. For sewage which has had some preliminary treatment, the rate should not exceed 100,000 gallons per acre per day, whereas the rate for raw sewage should be less than this. For rough estimates made without tests of the sewage in question, the rate should not be taken at more than 1,000 persons per acre. If the preliminary treatment of the sewage has been thorough and the material of the sand filter is coarser than ordinary the rate of filtration can be high. For less careful preliminary treatment and fine filtering material the rates must be reduced. The sewage must undergo sufficient preliminary treatment to remove large particles of solid matter which would otherwise clog the dosing apparatus and the filter. This treatment should include grit removal, screening, and some form of tank treatment. Some plants have operated successfully with a stale sewage and no preliminary treatment, as at Brockton, Mass. Septic tank effluent can be treated successfully on an intermittent sand filter, but not so satisfactorily as the effluent from a tank delivering a fresh sewage.
The material of the filter should consist of clean, sharp, quartz or silica sand with an effective size[164] of 0.2 to 0.4 mm., preferably about 0.25 to 0.35 mm., and a uniformity coefficient[165] of 2 to 4. Within the limits mentioned no careful attention need be given to the size of the material. Natural sand found in place has been underdrained and used successfully for sewage treatment. The size of the sand is fixed by the rate of filtration rather than the bacteriological action of the filter. A coarse sand will permit the sewage to pass through the bed too rapidly, and a fine sand will hold it too long or will become clogged. The same size of material should be used throughout the bed, except that a layer of gravel from 6 to 12 inches thick, graded from very small sizes to stones just passing a 2–inch ring should be placed at the bottom to facilitate the drainage of the bed.
The thickness of the sand layer should not be less than 30 inches to insure complete treatment of the sewage. In shallower beds the sewage might trickle through without adequate treatment. Beds are ordinarily made from 30 to 36 inches deep, but when deeper layers of sand are found in place there is no set limit to the depth which may be used. The shape and overall dimensions of the bed should conform to the topography of the site and the rate of filtration adopted. A plan and cross-section of an intermittent sand filter showing the distribution and under drainage systems are given in Fig. 166 and 175.
The distribution system consists of a system of troughs on the surface of the filter, laid out in a branching form, as shown in the figure. The openings in the troughs should be so located that the maximum distance from any point on the bed to the nearest opening should not exceed 20 to 30 feet. If the filters are small enough, troughs need not be used, the sewage being distributed from one corner, or from mid-points on the sides. Where troughs are used they should be supported from the bottom of the filter in order to prevent uneven settling due to the washing of the sand. The openings in the troughs are made adjustable by swinging gates as shown in Fig. 176, or by other means so that after the filter is in operation the intensity of the dose on any portion of the filter can be changed. The troughs may be placed with their bottoms level with the surface of the sand and with sides of sufficient height to give the required gradient to the water surface, or they may be built up above the surface of the filter and given the required slope so that the surface of the flowing water is parallel to the bottom of the trough. In either case a splash plate should be placed at each opening, so that not less than 2 feet of the surface of the sand is protected in all directions from the opening. A stone or concrete slab 2 to 4 inches thick makes a satisfactory splash plate. Either wood or concrete may be used for the construction of the troughs. The former is less durable, but also less expensive in first cost. The capacity of the troughs may be computed by Kutter’s formula with the quantity to be carried equal to the maximum rate of discharge of the feeding siphon, with a reduction in size below each branch or outlet proportional to the amount which will be discharged above this point.
The operation of automatic devices for dosing the bed is explained in