Chapter IV: Part 4

In connection with some elaborate studies of methods of distribution for percolating filters, which were carried out for the Sanitary Research Laboratory and Sewage Experiment Station of the Massachusetts Institute of Technology, by Messrs. Winslow, Phelps, Storey and McRae, a special series of investigations were made into the value of what is referred to as the “gravity method of spraying.” This consists in supplying the sewage from troughs or pipes supported above the filter, and discharging it through orifices on the under side of the pipes or troughs in the form of solid jets, which are made to impinge upon concave discs fixed near to the surface of the filter. The effect of this impingement is to cause the liquid to splash upwards and outwards in the form of a fine spray, and, as a result of a long series of tests with various forms of discs and various heads, it was found that the best average results were obtained with a concave metal disc 3 inches in diameter having a curvature corresponding to a radius of 2 inches. With a total head of 4 feet and a head of 3 feet from disc to the pipe above, it gave the best results obtained with that total head. With a total head of 6 feet and a head of 4 feet from disc to supply pipe, it gave the best results obtained by any of the forms of this type of distribution that were tested.

A method of distribution similar to this type of gravity disc spray is manufactured by Messrs. Glenfield and Kennedy, Ltd. The details of this disc, with the methods of supporting same and the method of supplying the sewage from an overhead pipe, are shown in Fig. 121.

In connection with sewage disposal works where chemical precipitation has been adopted for preliminary treatment, the tank effluent has been distributed over fine-grained percolating filters, by means of dosing tanks discharging comparatively large volumes intermittently over the surface of the filter. In such cases the aim has been to deliver the liquid at such a rate that it would flood the entire surface of the filter in a very short space of time, and then percolate through and leave the surface free to aerate for as long a period as possible before the next discharge took place. This process involved the use of small filters, in order that the area to be flooded should not be too large for the liquid to spread over it rapidly, and the use of very fine material to prevent the liquid from passing through too quickly. Among the difficulties encountered by this method were the displacement of the surface layer of the material, due to the high rate of delivery; and the labour involved in cleaning the surface of the filter. A further trouble which arises when effluents from septic tanks are treated by this process is the evolution of evil-smelling gases, which are certain to cause a nuisance if there are any houses in the vicinity. Efforts have been made to obviate this difficulty by distributing the sewage through open-jointed or perforated pipes laid a few inches beneath the surface of the filter, as described later in connection with contact beds (page 200). While this is satisfactory to a certain extent, especially where the volume to be treated is small, and the filter capacity is comparatively large, neither this method, nor that of flooding the surface of the filter, can be said to comply with “the desideratum in the application of sewage to percolating filters,” quoted previously (page 106), nor do they avoid the difficulties encountered in securing uniform distribution dealt with in the following pages. The defects of these methods of distribution may also have been aggravated by the very optimistic anticipations on the part of the designers of the schemes as to the volume of sewage which could be satisfactorily purified under such conditions. It is, therefore, desirable to point out that, in the event of either of these methods of distribution being adopted, care should be taken to secure a very high degree of clarification (reduction of matters in suspension) in the preliminary tank treatment, and to provide filters of ample cubic capacity.

_Methods of Feeding Percolating Filters._—Under the heading “Methods of Distribution” (page 106), reference is made to some experiments, from the results of which the conclusion was drawn that “a high efficiency in sub-surface distribution is fostered by a slow continuous rate of application rather than by an intermittent application at a higher rate.” This agrees in every respect with the author’s own experience, and confirms his opinion that intermittent discharges to percolating filters should only be resorted to in cases where it is absolutely necessary in order to assist in securing uniformity of distribution, or to ensure a sufficient volume for the operation of the appliances adopted for distribution. Among the former may be cited fixed spray nozzles and jets from fixed pipes, in which cases an intermittent supply is useful in causing a regular variation of the head upon the orifices, thus varying the distance to which the jets or sprays are thrown, and producing greater uniformity of distribution per unit of area covered. Examples of cases, where intermittent supply is necessary in order to discharge the volume required to operate the appliances adopted for distribution, are found in connection with most types of fixed distributors and nearly all types of automatic revolving distributors. In the case of the latter, it is well known that a certain minimum head is necessary to overcome the friction due to the resistance of the air and to the weight of the apparatus itself on its bearings. However small this friction may be, it needs a volume of sewage slightly in excess of that required to fill the spray-holes, which must be large enough to take the maximum flow of sewage when working under the maximum head. It frequently occurs, especially in small schemes, and in schemes where the percolating filters are preceded by contact beds or slate beds, that the rate of flow of the sewage is at times so low that it is not equal to the minimum volume required to operate the distributor, and, in the absence of any arrangement to overcome the difficulty, the distributor would cease to revolve and the sewage would trickle through the spray-holes without proper distribution. The same difficulty arises in connection with many of the fixed methods of distribution, and it is most readily overcome by the use of a dosing tank fitted with an automatic syphon or valve, by means of which the sewage is held-up in the tank until it reaches a certain predetermined level, and is then discharged at a given rate to the filter. When the tank is empty the discharge ceases, and the sewage is again held-up as before.

In the case of large installations, or schemes where the whole of the sewage is pumped and the rate of flow to the filters is thus under control, it is not necessary to use a dosing tank for the purpose of providing the rate of discharge to the filters necessary to keep the distributor in motion. It is, however, maintained in some quarters that intermittent supply is desirable in any case, in order to secure alternate periods of work and rest for aeration. There is one obvious disadvantage in this method of working. Assuming that during the maximum rate of flow of the sewage the volume which comes down in 5 minutes is stored in a dosing tank, and discharged to the filters in 2½ minutes, it is clear that the rate of distribution is twice as great as it would be if the distribution were continuous over the whole period of 5 minutes. It is claimed that the disadvantage of the higher rate of distribution is counteracted by the 2½ minutes of rest and aeration, but on this point there is room for doubt, especially when the conditions which come into play during the average and minimum rates of flow of the sewage are taken into consideration. Taking the average rate as equal to one-half the maximum rate, it will be seen that the dosing tank will discharge every 10 minutes, but the time in which its contents are delivered to the filter will still be 2½ minutes, so that under these circumstances the rate of distribution on the filter will be four times as much as it would be if the distribution were continuous. During the minimum rate of flow of the sewage the conditions are still worse, and it is difficult to accept the theory that periods of rest for 7½ minutes will compensate for excessive rates of distribution at four times the rate under continuous operation for periods of 2½ min. at a time.

Those who advocate intermittent supply under all circumstances would appear to have lost sight of the fact that very ample periods of rest are already provided by all revolving distributors. When working at the rate of one revolution per minute, the average time taken by each arm to pass over any one point on the surface of the filter may be taken as 0·5 second. Before the succeeding arm reaches the same point a period of 15 seconds will have elapsed, so that under these conditions the ratio of the periods of work to periods of rest is as 1 to 30. In other words, even during continuous distribution, the time allowed for rest and aeration is thirty times as much as that during which the sewage is actually being delivered to the filter, and there would thus appear to be no reason for unnecessarily adding to the periods of rest by means of dosing tanks, especially as such a course involves a greatly increased rate of distribution at all times when the discharge actually takes place.

There is, of course, the possibility of adopting the happy medium, which would consist of a dosing apparatus of such a type that it would provide a continuous supply during the maximum flow of the sewage, and only act as an intermitting appliance during the minimum rate of flow. This is probably the best arrangement to adopt in all cases where a dosing apparatus is absolutely essential.

As previously stated, however, there are many cases where the conditions render the use of a dosing tank absolutely necessary to prevent the stopping of the distributor, and these have brought about the introduction of several types of automatic syphons and valves, all arranged specially for the purpose of giving intermittent discharges to filters, contact beds, areas of lands, and so forth. There are various methods of constructing dosing tanks, all dependent upon the arrangement of the preliminary processes and the filters which follow them. In nearly all cases, however, it is found necessary to reduce the fall taken up by the dosing tank to the minimum, and the different syphons and valves have thus been designed to work with the least possible head. One point in the construction of dosing tanks should not be overlooked, i.e. the provision of a washout valve for use when it becomes necessary to clean out the tank. It may be noticed here that neither the Fiddian type of distributor, on the water-wheel principle, nor the power-driven distributors, require any dosing tank to keep them in motion during the periods of minimum flow of the sewage. In the case of the Fiddian distributor the apparatus remains standing until the buckets are filled, and each time these discharge their contents the distributor is moved forward a short distance. Under these conditions, however, the disadvantage of long intervals of rest between comparatively large discharges of sewage still remains. Undoubtedly the most even rate of distribution under all conditions is secured by the use of power-driven distributors, whether of the rotary or travelling type, but these can only be economically adapted for use in large installations or where power is available at a very low cost.

T = tank. C = carrier.
P = supply pipe. F = filter.
E = effluent channel. H = humus tank.]

T = tank. C = carrier.
P = main supply pipe. P_{1} = branch supply pipe.
V = regulating valve. W = washout valve.
F = filter. E = effluent channel.
H = humus tank.]

Apart from the question of continuous _versus_ intermittent methods of feeding percolating filters, there still remains the problem of deciding whether the actual connections between the tanks and the filters shall be in the form of channels (carriers) or pipes. For fixed sprays and revolving distributors, which involve the use of pressure due to a column of liquid, the actual connection to the apparatus must be in the form of a pipe extending at any rate to the outside of the filter. Whether it should be continued beyond this point in pipe form or in a channel depends entirely upon local conditions, such as the slope of the ground, the planning of the filters, and the relative cost of the two methods. There is one thing to be said in favour of pipes, i.e. that there is less chance of a nuisance being created by the evolution of foul gases from the sewage than in the case of open channels. On the other hand, channels may be, and frequently are, covered in. Whichever method is adopted, care should be taken to provide a valve or penstock on the connection to each separate filter, in order to be in a position not only to throw any filter out of work when desired, but also to regulate the rate of supply to each filter independently of the others. Washout valves should also be provided on the channels or pipes, in suitable positions. Where the filters are arranged in groups, as suggested in Figs. 75 and 77, the intervening spaces may be used to accommodate a simple receiving tank, or a dosing tank, to which the sewage or tank effluent may be conveyed by an elevated channel or by a pipe under pressure. The connections to the adjacent filters may then be provided in the form of pipes leading from this central tank direct to the distributor at the centre of each filter. If the receiving tank or dosing tank in such cases is built upon piers, the floor of the tank will come above the level of the surface of the filters, and the space below may be utilised for an effluent receiving chamber or even for a settling tank to arrest the solids in suspension in the effluent. There are many ways in which the supply to filters may be arranged, and the preceding notes are merely intended as suggestions, which may be elaborated as found desirable to suit conditions which vary almost in every case.

T = tank. D = dosing tank.
P = supply pipe. F = filter.
E = effluent channel. H = humus tank.]

Figs. 121_a_ and 121_b_ have been prepared to show conditions under which channels and pipes respectively would be suitable for supplying the tank effluent to the filters. Fig. 121_c_ is the plan of a works where a dosing tank is used to deliver the tank effluent intermittently through separate supply pipes to each of the filters. If the valves on the supply pipes are properly adjusted, each filter will receive an equal proportion of each discharge from the dosing tank. An alternative method is to provide an apparatus by means of which each separate discharge from the dosing tank is delivered in rotation to each of the filters.

In cases where the appearance of the works must be taken into consideration, the method of arranging the tanks and filters suggested in Fig. 121_d_ may be adopted. By this means the space covered by the works is utilised to the utmost, and the tanks are of the form by which uniformity of distribution and the maximum reduction of velocity is secured. On the other hand, it is probable that the slope of the site will only rarely suit this arrangement.

O = outfall sewer. S = screen chamber.
D = detritus tank. T = settlement tank.
C = carrier. P = supply pipe.
F = filter. E = effluent channel.
H = humus tank.]

_Dosing Apparatus._—Various methods of providing intermittent supply are in use. Among these one of the first was a simple form of balanced valve with float. This, however, has been superseded by other types, among which the syphons are the simplest form. One of the first of this type specially designed for the purpose in question was the low-draught syphon, manufactured by Messrs. Adams Hydraulics, Ltd., and shown in Fig. 122. The special feature of this apparatus is that it can be arranged for a small depth of sewage in the dosing tank. It will work with any depth from 6 inches upwards. These syphons can also be arranged to work singly, in pairs, each one coming into action alternately, or in numbers of three and upwards, when they can be adjusted to work in sequence or simultaneously.

Another type of syphonic dosing apparatus is that supplied by the Patent Automatic Sewage Distributors, Ltd., and shown in Fig. 123. In this case a syphon is combined with a mechanical valve operated by floats, hence the name “auto-mechanical syphon.”

A further type of syphonic dosing apparatus is manufactured by Messrs. George Jennings, Ltd., as shown in Fig. 124. These syphons are arranged to work in regular sequence, and are operated by means of air valves, which are shown mounted on a frame and connected by means of air pipes to the syphons. The sequence is obtained by means of cams, fixed in different positions on a shaft, which is rotated by means of a float actuated by the rise and fall of the liquid in the dosing tank.

A different type of dosing apparatus is the “Coleman” valve, manufactured by The Ames Crosta Sanitary Engineering Company, Fig. 125. This consists of a balanced valve, suspended from a hollow lever which contains mercury, and has a branch provided with a float actuated by the rise and fall of the liquid in the tank. As the sewage rises in the tank, the float overcomes the resistance of the balance weight on the lever, which is lifted, and the mercury falls to the lower end of the lever. The combined action of the float and the mercury opens the valve, and the sewage is discharged. When the tank is emptied, the apparatus resumes its former position ready for the next dose.

Another type of dosing apparatus is manufactured by Messrs. Mather and Platt, Ltd. This consists primarily of a flap valve, counterbalanced weights, and a balance vessel or drum, which is filled gradually with liquid, the whole being pivoted about a horizontal axle. When the measuring chamber and balance drum are both empty, the counterbalance weights raise the latter and allow the flap valve to close, the balance drum then being in its highest position. As the measuring chamber fills, a portion of the sewage is allowed to pass from it into the balance drum, which is thus gradually filled, and at a given level its weight will suffice to overcome the counterbalance weights and the pressure of the sewage on the flap valve, and will open the latter, thus allowing a measured quantity of sewage to be discharged into the troughs or pipes for spreading over the filter. The balance drum discharges its contents meanwhile, and thus, when the measuring chamber is empty, the flap valve closes again by the action of the counterweights, and the chamber is ready to receive another quantity of sewage (Fig. 126).

Messrs. George Kent, Ltd., have introduced several different types of dosing apparatus, which can be arranged to draw off less than 6 inches head of water and any greater depth. In one case the apparatus, Fig. 127, consists essentially of a seating S, through which the liquid discharges, a valve-cap C, attached to arms which rotate about the pin P and nearly balanced by the weights W, also by floats F F attached to a frame, also rotating about P and carrying two adjusting weights A A. The apparatus is actuated by the rise of the liquid, raising the floats and lifting the cap C, which returns to its position after the discharge has taken place. For greater depths a different type is used, as shown in Fig. 128. In this case the apparatus consists of a valve seating S, through which the liquid discharges, two floats A and B, a controlled device for giving a sudden release and for regulating the levels at which the valve is opened or closed, and an alternating gear G, which is fitted with two or more valves placed in the chamber and operated in rotation. The action of the apparatus is as follows: When the liquid in the dosing chamber reaches the top level, the float A has acquired sufficient buoyancy to cause the shoulders of the plate K to suddenly push aside the rollers at the ends of the weighted levers L L of the control device. The lugs T T then come in contact with the collar R, and the valve-cap C and the float B are lifted. The tank discharges until the floats A and B have lost sufficient buoyancy to cause the bottom shoulders of the plate to push aside the rollers and allow the valve to close.

This apparatus is applied in various ways for different purposes in connection with sewage works, and the same firm also manufacture special valves for giving a measured discharge under various heads.

The Ames Crosta Sanitary Engineering Company, Ltd., also supply a very ingenious method of providing intermittent supply to filters, as shown in Fig. 129, which is equally suitable for percolating filters, for contact beds, or for intermittent filtration on land. This machine is arranged to give any desired timed discharge to the beds in rotation, or to discharge measured volumes to the beds in rotation. By means of valves or weirs, the machine is arranged so that additional beds in the series are automatically brought into operation as the increase in flow demands, and the additional beds brought into action go out of action as the flow decreases. The illustration shows a machine for delivering timed volumes on a series of six beds. The square chamber is divided into six divisions by means of iron plates; each division is connected to one of the six beds by means of a pipe. A circular dish or vessel is suspended over these divisions, and revolves on ball-bearings carried on the bridge. The circular dish is provided with a weir at its circumference, and the sewage or tank effluent which enters the dish from the pipe in the centre is diverted by means of the weir or outlet to one of the divisions supplying the beds. The upright shaft from which the revolving dish is suspended is fitted with a six-toothed ratchet wheel, and, by the rise and fall of a float in the liquid of the actuating tank, a pawl is moved along until it engages with the next tooth of the ratchet wheel. When the pawl has caught the tooth of the ratchet wheel the outlet valve in the actuating tank is opened, and as the liquid flows out of the tank the float descends, and by means of levers the pawl is moved one-sixth of a revolution; the pawl being engaged with the ratchet wheel, the dish is revolved one-sixth of a revolution, and the weir is thus brought over the next division, and consequently the next bed receives the flow. Extra weirs can be arranged at various levels, so that if more sewage is coming than that required for one bed, a portion of the liquid would flow over one or more of the weirs, and so on, to the respective beds. The flow to the actuating tank is taken from the feed channel, and can be set to fill the actuating tank in any desired period of time. The discharge from the actuating tank can be conveyed either to a special plot of land, or into the revolving dish. An ingenious device is arranged so that any bed in the series can be instantly shut out of action should it become overworked. In an apparatus for feeding revolving sprinklers, arrangements are made for bringing the extra beds into operation by means of valves controlled by the flow of sewage.

Messrs. Glenfield and Kennedy also manufacture a balance-valve type of dosing apparatus operated by means of floats and buckets. This is shown in Fig. 130.

The “Uniform” automatic distributor, manufactured by Messrs. Whitehead and Poole, is designed for the automatic distribution of measured volumes of liquid to filters or other areas in regular rotation. This is shown in Fig. 131, and the principles on which it is constructed involve rotating arms carried on a float operated in a closed central chamber. The arms are rotated by the reaction of the discharge of the liquid, but are locked so as to remain stationary in one position, continuing to feed one filter until the head of liquid actuating a regulating cock and tipper releases the locking device and permits the rotation of the arms to the next bed, where the process is repeated, and so on to all the filters in rotation.

The “Ponding or Intermitting Valve,” supplied by the Septic Tank Co., is shown in Fig. 132. The essential features of this apparatus are two cylindrical vessels of cast-iron or other suitable material, approximately of the same shape, weight and displacement. One vessel is used as a bucket and the other as a bell. They are suspended from two equal arms of a pivoted lever at approximately the same distance from the standard carrying the lever. On the same end of the lever to which the bell is hung, an ordinary lift-up valve is attached, the seat of which is at or below the low-water level in the ponding chamber. This valve is connected with the lever by means of a linked rod, chain, or other suitable arrangement. The cubical content of the bell and bucket is sufficient to ensure the displacement of the liquid, in excess of that required to lift the valve from its seat, by rocking the lever to which it is attached. In the bottom of the bucket a draw-off valve is provided for the purpose of emptying same. This valve is constructed so as to prevent the liquid passing through it into the bucket as it rises in the chamber, and to open and allow the contents of the bucket to escape when the water in the ponding chamber has almost been discharged. The operation of the apparatus is as follows. The ponding valve being closed, and the supply of liquid being turned into the ponding chamber, the liquid will rise. On reaching the small ball in the bottom of the valve of the bucket, the ball will float and close this bucket valve, so as to prevent the entrance of the liquid through and into the bucket. The liquid continuing to rise, will gradually and simultaneously submerge the bell and the bucket. The upper edge of the bucket is fixed at the predetermined height to which the liquid is to rise in the ponding chamber. The bell and bucket being of the same size and weight, and being hung so that the one is level with the other, the displacement of both will be equal up to the time the liquid reaches the upper edge of the bucket, and consequently no movement of the lever will take place. On the liquid rising above the top edge of and filling the bucket, it will cause the displacement of the bucket to be reduced to much less than the displacement of the bell, and this difference is sufficient to lift the valve from its seat, and discharge the contents of the ponding chamber. As the liquid in the chamber recedes, the bucket and the bell will both be gradually left suspended in the air. Until the liquid in the chamber drops below the bottom of the bucket, the latter will remain full; on the water continuing to fall, the ball in the small valve at the bottom of the bucket will drop, allowing the contents of the bucket to escape. On the bucket being emptied of its contents, the weight of the bell, together with the valve and chain or rod, will overcome that of the bucket, the valve will close, and the liquid commence to pond in the chamber again.

The special syphons manufactured by Messrs. Burn Bros., described under contact bed apparatus, are also suitable for use as dosing syphons, and work singly or in sequence, and give intermittent discharge to filters or to land (see Fig. 145, page 218).

An auto-mechanical type of syphon for dosing tanks is supplied by the Carlton Engineering Co., and is illustrated in Fig. 133.

_EFFLUENT SETTLING TANKS OR HUMUS PITS._

Reference has already been made, under the heading “Grading of Filtering Material,” to the advisability of using coarse material, for the reason that the converted organic matter will in that case readily pass away in the effluent, and thus prevent the choking of the filter. Even with finer material, a certain amount of solids in suspension will be found in the effluent, and, in order to produce a final effluent suitable for discharge into any stream or watercourse, it is necessary to arrest and remove these suspended solids. This fact has been recognised by the Royal Commission on Sewage Disposal, who recommend the adoption of effluent settling tanks with a capacity equal to two hours’ flow of the sewage, and provided with means for removing the deposit.

The solids in effluents from percolating filters are rather difficult to arrest, as they are in the form of very finely divided matters in suspension. Many methods have been tried in various places, but in the author’s experience he has found that the chief factor in securing a satisfactory settlement of these solids, is the reduction to the minimum of the velocity of the effluent in its passage through the settling tank. If this principle is adopted, the simplest form of tank is similar in construction to that suggested for detritus tanks, so long as the outlet end is constructed in the form of a weir of the greatest possible length under the circumstances. Such a tank is illustrated in Fig. 134. By this means the rate of flow over the weir may be reduced to the minimum; and if, in addition to this, the outlet at the bottom of the tank is arranged in the form of a plug valve fixed in a pocket below the lowest point of the floor proper, near to the inlet end, and the floor is laid with a sharp slope towards this outlet, it will be found possible, as a rule, to draw off the deposit without discharging the entire contents of the tank, as long as it is done at frequent intervals. As there is a tendency for a scum to form on the surface of the liquid in these tanks, it is desirable to provide a scum-plate of wood, slate, or other material, as shown in the illustration, and in all cases, except the smallest schemes, these tanks should be constructed in duplicate. It may be mentioned here that the hydrolytic tank, Fig. 34, page 52, and the separator of the Septic Tank Co., Ltd., Fig. 47, page 65, have both been adopted for use as effluent settling tanks.

The deposit from effluent settling tanks, as a rule, rapidly dries without creating a nuisance when it is spread out in a thin layer upon a suitable draining bed similar to that suggested for dealing with sludge from settling tanks (Fig. 58, page 83). Under favourable conditions as to fall, the draining bed can be constructed below the level of the sludge outlet from the effluent settling tank, and the deposit can then be drawn off by gravitation. There is, however, still the problem of disposing of the liquid flowing from the bed, and as this should have a free outlet, it usually happens that the levels do not permit of the discharge of the deposit by gravitation. Under these circumstances the outlet should still be arranged as shown, Fig. 58, but it should be connected to a sludge well fitted with a chain-pump, or other means of raising the deposit to the draining bed, which may thus possibly be situated at the same level as the similar beds for the sludge from the settling tanks. When dry, this deposit may be spread out on the land, or used in gardens and on farms as a manure.

_SAND FILTERS._

It would probably be more correct to use the term “fine-grain filters” to describe the alternative methods occasionally adopted to deal with the effluents from percolating filters, as they do not always consist of sand. Fine clinker, ashes, broken saggars, and similar material, is equally suitable so long as it is of a gritty nature and not wholly dust. The term “sand-filters” is, however, used here, as it is well known in connection with the filtration of drinking water, and the method of construction is practically the same.

Although primarily designed for the purpose of arresting the suspended solids in final effluents, and thus required to act simply as mechanical strainers, sand filters have the additional advantage of increasing the degree of purification, especially from a bacterial point of view. The most important factor to be considered in constructing these filters is the grading of the material. It must not be too fine or contain too great a proportion of dust, or it will rapidly become clogged and involve much labour and expense in cleaning the surface layer. The nearest approach to perfection in material for this purpose is the coarse Leighton Buzzard filter sand. As in the case of the best material for the percolating filters themselves, it may cost a little more than other less satisfactory kinds, but it will generally be found to be the cheapest in the end.

In constructing filters of this type, whether composed of sand or other material, it is essential that the bottom layer should be of a coarser grade, in order to provide free drainage. In a general way it will be found satisfactory to have a series of 2-inch agricultural drain pipes laid on the floor and converging towards the outlet. Over the whole floor should then be laid a layer 6 inches deep of gravel, broken bricks or stones of the size of walnuts, upon this a layer 3 inches deep of pea gravel, and at the top a layer not less than 9 inches deep of suitable sand or other fine grade material. The surface of the filter should be well below the level of the inlet, in order to allow the liquid to pond up on the surface 6 inches to 9 inches in depth without backing up to the level of the floor of the percolating filter. One of the difficulties encountered in operating filters of this type is to secure even distribution over the whole area. The means to be adopted for this purpose should be simple and easily cleaned, and it is usual to find troughs of wood or iron, glazed ware channel-pipes and similar arrangements in use. Unfortunately these do not effectively cover the whole area until the sand has been saturated and the surface slightly coated, thus preventing the liquid from passing through as fast as it comes in. When this occurs, however, the time for cleaning the surface is not far distant, and when the filter is brought into use again the whole preliminary process has to be repeated. The best way of avoiding these difficulties would appear to be to arrange the filter in such a way that the liquid must cover the whole area from the very beginning. This can be accomplished by fixing the normal outlet _above_ the level of the surface of the filter as suggested, Fig. 135, where the final effluent discharge is normally from the end of the swivel-jointed pipe when in its vertical position. When it becomes necessary to clean the filter and drain it for purposes of aeration, the swivel-jointed pipe is simply lowered to the floor of the outlet-chamber, as shown in plan, and raised again when the filter is brought into operation.

Filters of this type should never be less than two in number, so that one may be in work while the other is being cleaned. It would probably be advisable to have even three or four filters for schemes of moderate size, so as to provide longer or more frequent intervals of rest for aeration. It will be obvious from the preceding observations that these filters must be substantially constructed and made absolutely watertight. When dealing with a good effluent from percolating filters or contact beds, these final sand filters may be provided at the rate of 1 square yard for every 500 gallons of the daily dry weather flow.

Where ample fall is available, careful consideration should be given to the advisability of operating these sand filters in the same manner as percolating filters, i.e. by using revolving sprinklers for the purpose of distribution without submerging the filtering material. This applies particularly in cases where it is desirable to secure a very high degree of bacterial purification. Recent investigations have shown that sand filters for drinking-water, when operated in this manner, are highly efficient and involve less expense for maintenance. In addition to this they require less cleaning, so that a much smaller area is thrown out of work during cleaning operations, and a smaller total area of filter surface is needed than in the case of similar filters operated on the submerged system. The additional fall required for the revolving sprinklers will usually be a serious difficulty in the case of sewage disposal works, but where it is available, and the extra cost entailed is not of great importance, the idea deserves consideration. Filters of this type should be preceded by an effluent settling tank as previously described.

_CONTACT BEDS._

The almost universal adoption at the present time of biological methods of sewage purification by means of artificial filters, is due entirely to the original experimental work of Mr. W. J. Dibdin, at the Barking Outfall Works of the London County Council. These experiments were carried out with a contact bed, and during the subsequent ten years an enormous number of works were constructed upon this principle. At the present time, however, it is a somewhat rare occurrence to find contact beds proposed for sewage disposal schemes of any size. It has been stated that the principle upon which they are operated is unscientific, that they rapidly become clogged and useless, and that, in any case, they are not capable of dealing with sewage at the same rate as percolating filters, or of producing such a high degree of purification. With regard to the first point it would be futile to endeavour to explain what is and what is not scientific. This must be left to the scientists. That contact beds have in many cases become clogged and useless cannot be denied, but there is also very little doubt that this unsatisfactory result has been due to one or more of the following causes: (_a_) overwork, (_b_) improper methods of operation, (_c_) the use of unsuitable material for filling the beds, (_d_) insufficient sub-drainage. It was most unfortunate that for some years the general idea of a contact bed was that it consisted of a simple excavation in the ground, filled with coke or similar material, into which the sewage was discharged, held up for two hours, and then drawn off; a very simple but crude affair altogether. It is now known that contact beds, like other systems, can only deal with limited volumes of sewage, the actual amount depending upon the character of the sewage and other factors; that there is a proper method of operating the beds, and that it must be strictly adhered to if the best results are to be produced; that unsatisfactory material is worse than useless, and that very ample means of sub-drainage are absolutely essential to the continued efficiency of the beds. It is probable that if these essential factors had been properly understood and acted upon from the outset, there would have been very few failures to record.

It has been stated that contact beds are obsolete, but there are engineers who even now recommend this system, and consider it satisfactory under some, if not under all conditions. In the opinion of the author contact beds are not obsolete, and there are cases where the conditions preclude the adoption of any other method of purification. Under these circumstances, it is considered desirable to describe in the following pages the details of design and construction which have been found by experience to be necessary to ensure satisfactory results.

_General Principles of Design._—The first point to be decided before commencing the design of a scheme of contact beds is whether single, double, or triple contact is necessary to produce the desired degree of purification, and this will depend upon the strength of the sewage and the destination of the final effluent. Single contact alone will not be sufficient, except in a very few cases where the sewage is weak (highly diluted), and even then it will necessitate the use of fine-grade material for filling the beds, and consequently a tank effluent of exceptional quality as regards the matters in suspension in order that the fine material may not be rapidly choked. Where a sufficient area of land of a suitable character can be procured at a convenient level for treating the effluent from the beds, single contact may be adopted with material of medium-grade, but even in this case special attention must be devoted to the preliminary process in tanks, so as to reduce the amount of solids in suspension in the tank effluent to the minimum. As a rule it will be found safer to adopt double contact, as the primary beds may then be filled with coarse grade material, which will be less liable to choke, and it will not be necessary to rely so much upon the land or any other final process that may have been provided. In special cases, and particularly where the sewage is strong, or an exceptionally high degree of purification is essential, triple contact should be adopted, but the tertiary beds may consist of a set of sand filters similar in construction to those described on pages 185 to 188. In some quarters the question of the grading of the material is considered of slight importance, and very little difference has been made in the size of the material for the primary and secondary beds, but in the author’s opinion it is absolutely essential that each series of beds should be filled with finer material than the preceding series, and the material in the final stage of treatment should be as fine as possible, so long as it does not contain any dust. In making these statements, it is assumed that the question of sub-drainage will be dealt with on the lines recommended later under that heading.

Another factor which has an important bearing upon the general design of a scheme of contact beds, is the method of operation which is to be adopted. It is generally assumed that all contact beds are worked in what is known as eight-hour cycles: viz. 1 hour filling, 2 hours standing full, 1 hour emptying, and 4 hours standing empty for rest and aeration. There has, however, been a tendency in the past to overlook the fact that the periods of standing full, and of emptying the beds, are the only sections of the cycle which are, as a rule, under absolute control. Unless special provision is made for the purpose, the time taken to fill each bed depends entirely upon the rate of flow of the sewage to the works, and the period of rest empty also depends upon the frequency with which the beds are filled, and thus indirectly upon the rate of flow of the sewage. For example, a set of four beds designed to receive each three fillings per day in wet weather, should not receive more than one filling per day in dry weather. Assuming that one-half the total flow comes down in six hours, it will be found that it takes six hours to fill two beds in the middle of the day, or three hours to fill one bed. During the remaining eighteen hours the other two beds are filled, one of them in say six hours and the other in twelve hours. The times taken to fill the four beds in this scheme would therefore be—No. 1, three hours; No. 2, three hours; No. 3, six hours; and No. 4, 12 hours. In each case the period of filling is thus much in excess of the one hour prescribed under the eight hours cycle. The obvious remedy is to subdivide the total area into a larger number of smaller beds, but if this is carried to its logical conclusion it will be seen that there must be 24 beds if the time taken to fill any one bed is not to exceed one hour. While it is very desirable to arrange this subdivision in order to secure the proper cycle of operations, the number of schemes where it is economically practicable are few, and recourse must be had to some other method of reaching the same end. This has already been recognised by most engineers, and provision is now usually made for a tank known under various names, such as dosing tank, collecting tank, equalising tank or holding-up tank, in which the tank effluent is stored until the volume accumulated is equal to the capacity of one contact bed, and the latter is then filled within the regulation time of one hour. If the necessity for making provision on these lines to ensure the proper working cycle had been recognised in the early days of contact beds, it would doubtless have prevented the troubles which have arisen in many places.

From the preceding observations, it will be seen that it is very necessary to come to a decision as to the method of operation to be adopted, before designing any scheme of contact beds. If the method of subdivision into a large number of small beds is preferred, the planning of the separate series, and the probable cost of the additional work involved, must be taken into consideration. On the other hand, if a smaller number of larger beds with a suitable dosing tank are preferred, the extra fall required for the latter must be provided for, even if it involves the reduction of the depth which would otherwise be available for the beds themselves.

There is still another matter which has a considerable influence upon the general design of a scheme of contact beds, viz. the slope of the ground upon which they are to be constructed. If it has a fairly rapid and even slope, the tanks and beds may be arranged close together, as shown in Fig. 136. The only part that needs special care in this case is the cross-wall between the primary and secondary beds, which will need strengthening, especially in its lower half, in order to resist the extra pressure it is required to take.

Where the slope is not so great, a saving in the cost of excavation may be effected by arranging the separate tiers of beds at some distance apart, as indicated in Fig. 137, and connecting one with the other by pipes. The aim to be attained in arranging the beds under these conditions is to have the entire area of the floors on solid soil, with the walls half in and half out of the ground.

Another set of conditions occasionally met with, is where the site of the works is perfectly flat and the position of the outlet for the final effluent involves the construction of the secondary beds either wholly or partly below the surface. In such cases the primary beds will come above ground, and it will then be found economical to arrange each set of beds in two rows end to end, with a central combined supply channel and effluent carrier, the latter being formed in the space between the walls which support the former, as shown, Fig. 138. If there is not sufficient head to allow of the supply channel being made deep enough to serve as the dosing tank, the latter may be constructed across the ends of the settling tanks, as suggested in Fig. 139, or in any other convenient position. A dosing tank in this position lends itself to the method of feeding the beds by means of closed pipes instead of by open channels, whether in sets of four, with a central chamber for the inlets and outlets illustrated in Fig. 139, or in series as Figs. 136 and 137.

There are doubtless other alternatives, or combinations of methods, which may be adapted to meet the exigencies of peculiar conditions of site and fall, but the foregoing details will probably suffice to suggest ideas to those in need of them in designing schemes of contact beds.

_General Construction._—The most important point to be borne in mind in constructing contact beds is, that they must be absolutely watertight. Should they leak in any way, the sewage may pass away untreated; or it may find its way into adjacent beds, and thus prevent these from being properly aerated during the periods of rest when empty. It is therefore evident that they should be constructed in a substantial manner. The floors are usually of concrete, and the thickness of the floor will depend upon the nature of the subsoil. If for any reason the floors have to be laid upon made-up ground, provision should be made by means of piers or cross-walls, carried down to the solid subsoil to support the floor, independent of the made-up ground which is, in all cases, absolutely untrustworthy. The walls of the beds may be constructed either of concrete throughout or of brickwork in cement, and they should be of such a thickness that they will withstand the pressure of the head of water which would result if the beds were filled to the top of the walls.

It will not be found satisfactory to place reliance either upon brickwork or upon concrete alone to form a watertight bed, and in both cases the whole of the floors, as well as the walls, should be rendered with cement mortar in the proportion of 2 parts of sharp clean sand to 1 part of Portland cement. No rendering should be done during frosty weather or during excessive heat, as in both cases it will usually be found defective, and it is better to stop the work altogether for a time than have to patch it up afterwards.

One safeguard which can be adopted to prevent difficulties later on, is to insist upon testing all such beds with water to the full height before any of the filtering material is placed in position. Any slight defects which may appear can be made good then at very little expense. If the defects are not discovered until after the beds are filled with the filtering material, they can only be properly rectified by removing the material, and this involves a considerable outlay. In order that these tests may be carried out without friction, it is necessary to stipulate clearly in the specification for the work that each bed is to be tested with water to the full depth before any material is placed in the bed, and that the contractor must take full responsibility for making the beds absolutely watertight. It is, of course, understood that the method of construction adopted by the engineer is such that, if properly executed, the beds will be absolutely watertight; and, in order to prevent any misunderstanding, an item should be included in the quantities for the contractor to provide whatever sum he may consider necessary to allow for making these tests. It may be thought sufficient to state simply that the contractor should make all absolutely watertight, and to leave it to him to provide the means for doing so. It will, however, be found more satisfactory to all concerned to provide all means both in specification and in the quantities for attaining the desired results. It is not sufficient even to use the word “watertight” alone in this connection, as the interpretation of this word may be subject to differences of opinion which are obviated by the addition of the word “absolutely.”

The foregoing observations refer not only to contact beds but to the tanks and other portions of the work which are required to hold water.