The total quantity of sewage to be dealt with per day can be ascertained by gauging the flow in those cases where the sewers are already constructed, but where the scheme is an entirely new one the quantity must be estimated. If there is a water supply system the amount of water consumed per day, after making due allowance for the quantity used for trade purposes and street watering, will be a useful guide. The average amount of water used per head per day for domestic purposes only may be taken as follows:—
DAILY WATER SUPPLY(Gallons per head per day.)
Dietetic purposes (cooking, drinking, &c.) 1 Cleansing purposes (washing house utensils, clothes, &c.) 6
If water-closets are in general use, add 3
If baths are in general use, add 5
Total 15
It therefore follows that the quantity of domestic sewage to be expected will vary from 7 to 15 gallons per head per day, according to the extent of the sanitary conveniences installed in the town; but with the advent of an up-to-date sewage scheme, probably accompanied by a proper water supply, a very large increase in the number of water-closets and baths may confidently be anticipated, and it will rarely be advisable to provide for a less quantity of domestic sewage than 15 gallons per head per day for each of the resident inhabitants. The problem is complicated in sea coast towns by the large influx of visitors during certain short periods of the year, for whom the sewerage system must be sufficient, and yet it must not be so large compared with the requirements of the residential population that it cannot be kept in an efficient state during that part of the year when the visitors are absent. The visitors are of two types—the daily trippers and those who spend several days or weeks in the town. The daily tripper may not directly contribute much sewage to the sewers, but he does indirectly through those who cater for his wants. The resident visitor will spend most of the day out of doors, and therefore cause less than the average quantity of water to be used for house-cleansing purposes, in addition to which the bulk of the soiled linen will not be washed in the town. An allowance of 10 gallons per head per day for the resident visitor and 5 gallons per head per day for the trippers will usually be found a sufficient provision.
It is, of course, well known that the flow of sewage varies from day to day as well as from hour to hour, and while there is no necessity to consider the daily variation—calculations being based on the flow of the maximum day—the hourly variation plays a most important part where storage of the sewage for any length of time is an integral part of the scheme. There are many important factors governing this variation, and even if the most elaborate calculations are made they are liable to be upset at any time by the unexpected discharge of large quantities of trade wastes. With a small population the hourly fluctuation in the quantity of sewage flowing into the sewers is very great, but it reduces as the population increases, owing to the diversity of the occupations and habits of the inhabitants. In all cases where the residential portions of the district are straggling, and the outfall works are situated at a long distance from the centre of the town, the flow becomes steadier, and the inequalities are not so prominently marked at the outlet end of the sewer. The rate of flow increases more or less gradually to the maximum about midday, and falls off in the afternoon in the same gradual manner. The following table, based on numerous gaugings, represents approximately the hourly variations in the dry weather flow of the sewage proper from populations numbering from 1,000 to 10,000, and is prepared after deducting all water which may be present in the sewers resulting from the infiltration of subsoil water through leaky joints in the pipes, and from defective water supply fittings as ascertained from the night gaugings. Larger towns have not been included in the table because the hourly rates of flow are generally complicated by the discharge of the trade wastes previously referred to, which must be the subject of special investigation in each case.
APPROXIMATE HOURLY VARIATION IN THE FLOW OF SEWAGE.Percentage of Total Flow Passing Off in each Hour.
—————-+————————————————————————| Population.Hour. +——-+——-+——-+——-+——-+——-+——-+———|1,000|2,000|3,000|4,000|5,000|6,000|8,000|10,000—————-+——-+——-+——-+——-+——-+——-+——-+———Midnight | 1.0 | 1.0 | 1.2 | 1.3 | 1.5 | 1.5 | 1.8 | 2.01.0 a.m. | 0.7 | 0.7 | 0.7 | 0.8 | 0.8 | 1.0 | 1.0 | 1.02.0 " | nil | nil | nil | nil | 0.2 | 0.2 | 0.3 | 0.53.0 " | nil | nil | nil | nil | nil | nil | nil | 0.24.0 " | nil | nil | nil | nil | nil | nil | nil | nil5.0 " | nil | nil | nil | nil | nil | nil | nil | 0.26.0 " | 0.2 | 0.2 | 0.3 | 0.5 | 0.6 | 0.5 | 0.7 | 0.87.0 " | 0.5 | 0.5 | 1.0 | 1.5 | 1.6 | 1.7 | 2.0 | 2.58.0 " | 1.0 | 1.5 | 2.0 | 2.5 | 3.0 | 3.5 | 4.0 | 5.09.0 " | 3.5 | 4.5 | 4.5 | 4.8 | 5.5 | 5.8 | 6.0 | 6.510.0 " | 6.5 | 6.5 | 6.8 | 7.0 | 7.5 | 7.7 | 8.0 | 8.011.0 " |10.5 |11.0 |10.5 |10.0 | 9.6 | 9.3 | 9.0 | 8.8Noon |11.0 |11.3 |10.8 |10.3 | 9.3 | 9.5 | 9.2 | 9.01.0 p.m. | 6.0 | 5.5 | 6.0 | 6.7 | 7.0 | 7.2 | 7.3 | 7.52.0 " | 7.0 | 7.3 | 7.0 | 7.0 | 6.5 | 6.5 | 6.2 | 6.03.0 " | 6.8 | 6.5 | 6.5 | 6.5 | 6.5 | 6.3 | 6.3 | 6.04.0 " | 7.5 | 7.5 | 7.3 | 7.0 | 6.7 | 6.5 | 6.2 | 6.75.0 " | 6.5 | 6.5 | 6.5 | 6.3 | 6.0 | 6.0 | 6.0 | 5.86.0 " | 4.5 | 4.5 | 4.7 | 4.8 | 5.0 | 5.0 | 5.0 | 5.27.0 " | 6.5 | 6.2 | 6.0 | 5.8 | 5.5 | 5.5 | 5.5 | 4.78.0 " | 6.2 | 6.0 | 5.8 | 5.5 | 5.5 | 5.3 | 5.0 | 4.89.0 " | 5.0 | 4.8 | 4.7 | 4.5 | 4.5 | 4.5 | 4.5 | 4.010.0 " | 4.8 | 4.6 | 4.2 | 4.0 | 3.8 | 3.5 | 3.0 | 3.011.0 " | 4.3 | 3.5 | 3.5 | 3.2 | 3.2 | 3.0 | 3.0 | 2.8—————-+——-+——-+——-+——-+——-+——-+——-+———Total |100.0|100.0|100.0|100.0|100.0|100.0|100.0|100.0—————-+——-+——-+——-+——-+——-+——-+——-+———
Percentage of total flow passing off during period named.
——————————-+————————————————————————————————+| Population. |+———-+———-+———-+———-+———-+———-+———-+————+| 1,000 | 2,000 | 3,000 | 4,000 | 5,000 | 6,000 | 8,000 | 10,000 |——————————-+———-+———-+———-+———-+———-+———-+———-+————+7.0 a.m. to 7.0 p.m | 77.3 | 78.8 | 78.6 | 78.7 | 78.5 | 78.8 | 78.7 | 75.2 |7.0 p.m. to 7.0 a.m | 22.7 | 21.2 | 21.4 | 21.3 | 21.5 | 21.2 | 21.3 | 21.8 |Maximum 12 hrs. | 84.0 | 83.6 | 82.6 | 81.7 | 81.0 | 80.6 | 79.7 | 78.2 |" 10 " | 72.8 | 72.8 | 72.1 | 71.4 | 70.0 | 69.8 | 69.2 | 68.5 |" 9 " | 66.3 | 66.6 | 66.1 | 65.6 | 64.5 | 64.8 | 64.2 | 63.3 |" 8 " | 61.8 | 62.1 | 61.4 | 60.8 | 59.5 | 59.0 | 58.2 | 57.5 |" 6 " | 48.8 | 49.1 | 43.1 | 47.5 | 46.8 | 46.5 | 46.0 | 45.8 |" 3 " | 23.0 | 28.8 | 27.11| 27.3 | 26.8 | 26.5 | 26.2 | 25.8 |" 2 " | 21.5 | 22.3 | 21.3 | 20.3 | 19.3 | 18.5 | 18.2 | 17.3 |" 1 " | 11.0 | 11.3 | 10.8 | 10.3 | 9.8 | 9.5 | 9.2 | 9.0 |Minimum 9 " | 3.4 | 3.9 | 5.2 | 6.6 | 7.5 | 6.9 | 8.8 | 10.0 |" 10 " | 6.9 | 7.4 | 8.7 | 9.8 | 10.7 | 10.4 | 11.8 | 13.0 |——————————-+———-+———-+———-+———-+———-+———-+———-+————+
The data in the foregoing table, so far as they relate to populations of one, five, and ten thousand respectively, are reproduced graphically in Fig. 12.
This table and diagram relate only to the flow of sewage—that is, water which is intentionally fouled; but unfortunately it is almost invariably found that the flow in the sewers is greater than is thus indicated, and due allowance must be made accordingly. The greater the amount of extra liquid flowing in the sewers as a permanent constant stream, the less marked will be the hourly variations; and in one set of gaugings which came under the writer's notice the quantity of extraneous liquid in the sewers was so greatly in excess of the ordinary sewage flow that, taken as a percentage of the total daily flow, the hourly variation was almost imperceptible.
[Illustration: Fig 12 Hourly Variation in Flow of Sewage.]
Provision must be made in the scheme for the leakage from the water fittings, and for the subsoil water, which will inevitably find its way into the sewers. The quantity will vary very considerably, and is difficult of estimation. If the water is cheap, and the supply plentiful, the water authority may not seriously attempt to curtail the leakage; but in other cases it will be reduced to a minimum by frequent house to house inspection; some authorities going so far as to gratuitously fix new washers to taps when they are required. Theoretically, there should be no infiltration of subsoil water, as in nearly all modern sewerage schemes the pipes are tested and proved to be watertight before the trenches are filled in; but in practice this happy state is not obtainable. The pipes may not all be bedded as solidly as they should be, and when the pressure of the earth comes upon them settlement takes place and the joints are broken. Joints may also be broken by careless filling of trenches, or by men walking upon the pipes before they are sufficiently covered. Some engineers specify that all sewers shall be tested and proved to be absolutely water-tight before they are "passed" and covered in, but make a proviso that if, after the completion of the works, the leakage into any section exceeds 1/2 cubic foot per minute per mile of sewer, that length shall be taken up and relaid. Even if the greatest vigilance is exercised to obtain water-tight sewers, the numerous house connections are each potential sources of leakage, and when the scheme is complete there may be a large quantity of infiltration water to be dealt with. Where there are existing systems of old sewers the quantity of infiltration water can be ascertained by gauging the night flow; and if it is proved to be excessive, a careful examination of the course of the sewers should be made with a view to locating the places where the greater part of the leakage occurs, and then to take such steps as may be practicable to reduce the quantity.
A method frequently adopted to gauge the flow of the sewage is to fix a weir board with a single rectangular notch across the sewer in a convenient manhole, which will pond up the sewage; and then to ascertain the depth of water passing over the notch by measurements from the surface of the water to a peg fixed level with the bottom of the notch and at a distance of two or three feet away on the upstream side. The extreme variation in the flow of the sewage is so great, however, that if the notch is of a convenient width to take the maximum flow, the hourly variation at the time of minimum flow will affect the depth of the sewage on the notch to such a small extent that difficulty may be experienced in taking the readings with sufficient accuracy to show such variations in the flow, and there will be great probability of incorrect results being obtained by reason of solid sewage matter lodging on the notch. When the depth on a l2 in notch is about 6 in, a variation of only 1-16th inch in the vertical measurement will represent a difference in the rate of the flow of approximately 405 gallons per hour, or about 9,700 gallons per day. When the flow is about lin deep the same variation of 1-16th in will represent about 162 gallons per hour, or 3,900 gallons per day. Greater accuracy will be obtained if a properly-formed gauging pond is constructed independently of the manhole and a double rectangular notch, similar to Fig. 13, or a triangular or V- shaped notch, as shown in Fig. 14, used in lieu of the simpler form.
In calculating the discharge of weirs there are several formulæ to choose from, all of which will give different results, though comparative accuracy has been claimed for each. Taking first a single rectangular notch and reducing the formulae to the common form:
____ Discharge per foot in width of weir = C \/ H^3
where H = depth from the surface of still water above the weir to the level of the bottom of the notch, the value of C will be as set out in the following table:—
TABLE No. 5.
RECTANGULAR NOTCHES._____Discharge per foot in width of notch = C \/ H^3—————————————————————————————————Values of C.———————————————————+—————————————-H Measured in | Feet. | Inches.———————-+—————-+—————+—————-+———————-| Gallons | C. ft | Gallons | C. ftDischarge in | per hour. | per min | per hour. | per min———————-+—————-+—————+—————-+———————-Authority. | | | |Box | 79,895 | 213.6 | 1,922 | 5.13Cotterill | 74,296 | 198.6 | 1,787 | 4.78Francis | 74,820 | 200.0 | 1,800 | 4.81Mo'esworth | 80,057 | 214.0 | 1,926 | 5.15Santo Crimp | 72,949 | 195.0 | 1,755 | 4.69———————-+—————-+—————+—————-+———————-
In the foregoing table Francis' short formula is used, which does not take into account the end contractions and therefore gives a slightly higher result than would otherwise be the case, and in Cotterill's formula the notch is taken as being half the width of the weir, or of the stream above the weir. If a cubic foot is taken as being equal to 6-1/4 gallons instead of 6.235 gallons, then, cubic feet per minute multiplied by 9,000 equals gallons per day. This table can be applied to ascertain the flow through the notch shown in Fig. 13 in the following way. Suppose it is required to find the discharge in cubic feet per minute when the depth of water measured in the middle of the notch is 4 in Using Santo Crimp's formula the result will be
C\/H^3 = 4.69 \/4^3 = 4.69 x 8 = 37.52
cubic feet per foot in width of weir, but as the weir is only 6 in wide, we must divide this figure by 2, then
37.52/2 = 18.76 cubic feet, which is the discharge per minute.
+———+ +———+ | | FIG. 13 | | | | | | | | | | | +———+ +———+ | | | | | | | | | | | | | | +———+ | | | | | | | | | +—————————————————+
Fig. 13.-ELEVATION OF DOUBLE RECTANGULAR NOTCHED GAUGING WEIR.
+———+ +———+ | \ FIG. 13 / | | \ / | | \ / | | \ / | | \ / | | \ / | | \ / | | \ / | | \ / | | \/ | | | | | | | | | | | +—————————————————+
If it is required to find the discharge in similar terms with a depth of water of 20 in, two sets of calculations are required. First 20 in depth on the notch 6 in wide, and then 4 in depth on the notch, 28 in minus 6 in, or 1 ft wide.
____ _____(1) C\/ H^3 = 4.69/2 \/ 10^3 = 2.345 x 31.62 = 74.15____ ____(2) C\/ H^3 = 1.0 x 4.69 \/ 4^3 = 1.0 x 4.69 x 8 = 37.52
Total in c. ft per min = 111.67
The actual discharge would be slightly in excess of this.
In addition to the circumstances already enumerated which affect the accuracy of gaugings taken by means of a weir fixed in a sewer there is also the fact that the sewage approaches the weir with a velocity which varies considerably from time to time. In order to make allowance for this, the head calculated to produce the velocity must be added to the actual head. This can be embodied in the formula, as, for example, Santo Crimp's formula for discharge in cubic feet per minute, with H measured in feet, is written
__________________ 195\/(11^3 + .035V - H^2
instead of the usual form of____195\/ H^3, which is used
when there is no velocity to take into account. The V represents the velocity in feet per second.
Triangular or V notches are usually formed so that the angle between the two sides is 90°, when the breadth at any point will always be twice the vertical height measured at the centre. The discharge in this case varies as the square root of the fifth power of the height instead of the third power as with the rectangular notch. The reason for the alteration of the power is thatapproximatelythe discharge over a notch with any given head varies as the cross-sectional area of the body of water passing over it. The area of the 90° notch is half that of a circumscribing rectangular notch, so that the discharge of a V notch is approximately equal to that of a rectangular notch having a width equal to half the width of the V notch at water level, and as the total width is equal to double the depth of water passing over the notch the half width is equal to the full depth and the discharge is equal to that of a rectangular notch having a width equal to the depth of water flowing over the V notch from time to time, both being measured in the same unit, therefore ____ ____ ____ C \/ H^3 becomes C x H x \/ H^3 which equals C \/ H^5.
The constant C will, however, vary from that for the rectangular notch to give an accurate result.
TABLE No. 6.
TRIANGULAR OR V NOTCHES.____Discharge = C x \/ H^5.
Values of C.
———————+———————————-+———————————— H Measured in | Feet. | Inches. ———————+—————+——————+—————-+—————— Discharge in | Gallons | C. ft per | Gallons | C. ft per | per hour | min | per hour. | min ———————+—————+——————+—————-+—————— Alexander | 59,856 | 160 | 120.0 | 0.321 Cotterill | 57,013 | 152.4 | 114.3 | 0.306 Molesworth | 59,201 | 158.2 | 118.7 | 0.317 Thomson | 57,166 | 152.8 | 114.6 | 0.306 ———————+—————+——————+—————-+——————
Cotterill's formula for the discharge in cubic feet per minuteis_______16 x C x B \/ 2g H^3
when B = breadth of notch in feet and H = height of water in feet and can be applied to any proportion of notch. When B = 2H, that is, a 90° notch, C = .595 and the formula becomes ____ 152.4 \/ H^5,
and when B = 4H, that is, a notch containing an angle of 126°51' 36", C = .62 and the formula is then written____318 \/ H^5.
The measurements of the depth of the water above the notch should be taken by a hook-gauge, as when a rule or gauge-slate is used the velocity of the water causes the latter to rise as it comes in contact with the edge of the measuring instrument and an accurate reading is not easily obtainable, and, further, capillary attraction causes the water to rise up the rule above the actual surface, and thus to show a still greater depth. When using a hook-gauge the top of the weir, as well as the notch, should be fixed level and a peg or stake fixed as far back as possible on the upstream side of the weir, so that the top of the peg is level with the top of the weir, instead of with the notch, as is the case when a rule or gauge-slate is used. The hook-gauge consists of a square rod of, say, lin side, with a metal hook at the bottom, as shown in Fig. 15, and is so proportioned that the distance from the top of the hook to the top of the rod is equal to the difference in level of the top of the weir and the sill of the notch. In using it the rod of the hook-gauge is held against the side of the gauge-peg and lowered into the water until the point of the hook is submerged. The gauge is then gently raised until the point of the hook breaks the surface of the water, when the distance from the top of the gauge-peg to the top of the rod of the hook-gauge will correspond with the depth of the water flowing over the weir.
The next consideration is the amount of rain-water for which provision should be made. This depends on two factors: first, the amount of rain which may be expected to fall; and, secondly, the proportion of this rainfall which will reach the sewers. The maximum rate at which the rain-water will reach the outfall sewer will determine the size of the sewer and capacity of the pumping plant, if any, while if the sewage is to be stored during certain periods of the tide the capacity of the reservoir will depend upon the total quantity of rain-water entering it during such periods, irrespective of the rate of flow.
Some very complete and valuable investigations of the flow of rain-water in the Birmingham sewers were carried out between 1900 and 1904 by Mr. D. E. Lloyd-Davies, M.Inst.C. E., the results of which are published in Vol. CLXIV., Min Proc. Inst.C.E. He showed that the quantity reaching the sewer at any point was proportional to the time of concentration at that point and the percentage of impermeable area in the district. The time of concentration was arrived at by calculating the time which the rain-water would take to flow through the longest line of sewers from the extreme boundaries of the district to the point of observation, assuming the sewers to be flowing half full; and adding to the time so obtained the period required for the rain to get into the sewers, which varied from one minute where the roofs were connected directly with the sewers to three minutes where the rain had first to flow along the road gutters. With an average velocity of 3 ft per second the time of concentration will be thirty minutes for each mile of sewer. The total volume of rain-water passing into the sewers was found to bear the same relation to the total volume of rain falling as the maximum flow in the sewers bore to the maximum intensity of rainfall during a period equal to the time of concentration. He stated further that while the flow in the sewers was proportional to the aggregate rainfall during the time of concentration, it was also directly proportional to the impermeable area. Putting this into figures, we see that in a district where the whole area is impermeable, if a point is taken on the main sewers which is so placed that rain falling at the head of the branch sewer furthest removed takes ten minutes to reach it, then the maximum flow of storm water past that point will be approximately equal to the total quantity of rain falling over the whole drainage area during a period of ten minutes, and further, that the total quantity of rainfall reaching the sewers will approximately equal the total quantity falling. If, however, the impermeable area is 25 per cent. of the whole, then the maximum flow of storm water will be 25 per cent. of the rain falling during the time of concentration, viz., ten minutes, and the total quantity of storm water will be 25 per cent. of the total rainfall.
If the quantity of storm water is gauged throughout the year it will probably be found that, on the average, only from 70 per cent. to 80 per cent. of the rain falling on the impermeable areas will reach the sewers instead of 100 per cent., as suggested by Mr. Lloyd-Davies, the difference being accounted for by the rain which is required to wet the surfaces before any flow off can take place, in addition to the rain-water collected in tanks for domestic use, rain required to fill up gullies the water level of which has been lowered by evaporation, and rain-water absorbed in the joints of the paving.
The intensity of the rainfall decreases as the period over which the rainfall is taken is increased. For instance, a rainfall of lin may occur in a period of twenty minutes, being at the rate of 3 in per hour, but if a period of one hour is taken the fall during such lengthened time will be considerably less than 3 in In towns where automatic rain gauges are installed and records kept, the required data can be abstracted, but in other cases it is necessary to estimate the quantity of rain which may have to be dealt with.
It is impracticable to provide sewers to deal with the maximum quantity of rain which may possibly fall either in the form of waterspouts or abnormally heavy torrential rains, and the amount of risk which it is desirable to run must be settled after consideration of the details of each particular case. The following table, based principally upon observations taken at the Birmingham Observatory, shows the approximate rainfall which may be taken according to the time of concentration.
TABLE No. 7.
INTENSITY OF RAINFALL DURING LIMITED PERIODS. Equivalent rate in inches per hour of aggregate rainfall during Time of Concentration, period of concentration A B C D E 5 minutes …………… 1.75 2.00 3.00 — — 10 " …………… 1.25 1.50 2.00 — — 15 " …………… 1.05 1.25 1.50 — — 20 " …………… 0.95 1.05 1.30 1.20 3.00 25 " …………… 0.85 0.95 1.15 — — 30 " …………… 0.80 0.90 1.05 1.00 2.50 35 " …………… 0.75 0.85 0.95 — — 40 " …………… 0.70 0.80 0.90 — — 45 " …………… 0.65 0.75 0.85 — — 1 hour ……………… 0.50 0.60 0.70 0.75 1.80 1-1/2 " ……………… 0.40 0.50 0.60 — 1.40 2 " ……………… 0.30 0.40 0.50 0.50 1.10
The figures in column A will not probably be exceeded more than once in each year, those in column B will not probably be exceeded more than once in three years, while those in column C will rarely be exceeded at all. Columns D and E refer to the records tabulated by the Meteorological Office, the rainfall given in column D being described in their publication as "falls too numerous to require insertion," and those in column E as "extreme falls rarely exceeded." It must, however, be borne in mind that the Meteorological Office figures relate to records derived from all parts of the country, and although the falls mentioned may occur at several towns in any one year it may be many years before the same towns are again visited by storms of equal magnitude.
While it is convenient to consider the quantity of rainfall for which provision is to be made in terms of the rate of fall in inches per hour, it will be useful for the practical application of the figures to know the actual rate of flow of the storm water in the sewers at the point of concentration in cubic feet per minute per acre. This information is given in the following Table No. 8, which is prepared from the figures given in Table No. 7, and is applicable in the same manner.
TABLE No. 8.
—————————————+—————————————————| Maximum storm water flow in| cubic feet per min per acre| of impervious area.Time of Concentration. +———+———+———+———+———| A | B | C | D | E—————————————+———+———+———+———+———5 minutes | 106 | 121 | 181 | — | —10 " | 75 | 91 | 121 | — | —15 " | 64 | 75 | 91 | — | —20 " | 57 | 64 | 79 | 73 | 18125 " | 51 | 57 | 70 | — | —30 " | 48 | 54 | 64 | 61 | 15135 " | 45 | 51 | 57 | — | —40 " | 42 | 48 | 54 | — | —45 " | 39 | 45 | 51 | — | —1 hour | 30 | 36 | 42 | 45 | 1091-1/2 " | 24 | 30 | 36 | — | 852 " | 18 | 24 | 30 | 30 | 67—————————————+———+———+———+———+———-l inch of rain = 3,630 cub. feet per acre.
The amount of rainfall for which storage has to be provided is a difficult matter to determine; it depends on the frequency and efficiency of the overflows and the length of time during which the storm water has to be held up for tidal reasons. It is found that on the average the whole of the rain on a rainy day falls within a period of 2-1/2 hours; therefore, ignoring the relief which may be afforded by overflows, if the sewers are tide-locked for a period of 2-1/2 hours or over it would appear to be necessary to provide storage for the rainfall of a whole day; but in this case again it is permissible to run a certain amount of risk, varying with the length of time the sewers are tide-locked, because, first of all, it only rains on the average on about 160 days in the year, and, secondly, when it does rain, it may not be at the time when the sewers are tide-locked, although it is frequently found that the heaviest storms occur just at the most inconvenient time, namely, about high water. Table No. 9 shows the frequency of heavy rain recorded during a period of ten years at the Birmingham Observatory, which, being in the centre of England, may be taken as an approximate average of the country.
TABLE No. 9.
FREQUENCY OF HEAVY RAIN ———————————————————————————-
Total Daily Rainfall. Average Frequency of Rainfall
———————————————————————————-
0.4 inches and over 155 times each year 0.5 " 93 " 0.6 " 68 " 0.7 " 50 " 0.8 " 33 " 0.9 " 22 " 1.0 " 17 " 1.1 " Once each year 1.2 " Once in 17 months 1.25 " " 2 years 1.3 " " 2-1/2 1.4 " " 3-1/3 1.5 " " 5 years 1.6 " " 5 years 1.7 " " 5 years 1.8 " " 10 years 1.9 " " 10 years 2.0 " " 10 years
—————————————————————————
It will be interesting and useful to consider the records for the year 1903, which was one of the wettest years on record, and to compare those taken in Birmingham with the mean of those given in "Symons' Rainfall," taken at thirty-seven different stations distributed over the rest of the country.
TABLE No. 10.RAINFALL FOR 1903.
Mean of 37stations inBirmingham England andWales.Daily Rainfall of 2 in and over …… None 1 dayDaily Rainfall of 1 in and over …… 3 days 6 daysDaily Rainfall of 1/2 in and over …. 17 days 25 daysNumber of rainy days……………… 177 days 211 daysTotal rainfall …………………. 33.86 in 44.89 inAmount per rainy day ……………. 0.19 in 0.21 in
The year 1903 was an exceptional one, but the difference existing between the figures in the above table and the average figures in Table 9 are very marked, and serve to emphasise the necessity for close investigation in each individual case. It must be further remembered that the wettest year is not necessarily the year of the heaviest rainfalls, and it is the heavy rainfalls only which affect the design of sewerage works.
If the whole area of the district is not impermeable the percentage which is so must be carefully estimated, and will naturally vary in each case. The means of arriving at an estimate will also probably vary considerably according to circumstances, but the following figures, which relate to investigations recently made by the writer, may be of interest. In the town, which has a population of 10,000 and an area of 2,037 acres, the total length of roads constructed was 74,550 lineal feet, and their average width was 36 ft, including two footpaths. The average density of the population was 4.9 people per acre. Houses were erected adjoining a length of 43,784 lineal feet of roads, leaving 30,766 lineal feet, which for distinction may be called "undeveloped"—that is, the land adjoining them was not built over. Dividing the length of road occupied by houses by the total number of the inhabitants of the town, the average length of road per head was 4.37 ft, and assuming five people per house and one house on each side of the road we get ten people per two houses opposite each other. Then 10 x 4.37 = 43.7 lineal feet of road frontage to each pair of opposite houses. After a very careful inspection of the whole town, the average area of the impermeable surfaces appertaining to each house was estimated at 675 sq. ft, of which 300 sq. ft was apportioned to the front roof and garden paths and 375 sq. ft to the back roof and paved yards. Dividing these figures by 43.71 in ft of road frontage per house, we find that the effective width of the impermeable roadway is increased by 6 ft 10 in for the front portions of each house, and by a width of 8 ft 7 in, for the back portions, making a total width of 36 ft + 2(6 ft 10 in) + 2(8 ft 7 in) = 66 ft 10 in, say 67 ft On this basis the impermeable area in the town therefore equals: 43,7841 in ft x 67 ft =2,933,528; and 30,766 lin ft x 36 ft = 1,107,576.
Total, 4,041,104 sq. ft, or 92.77 acres. As the population is 10,000 the impermeable area equals 404, say, 400 sq. ft per head, or ~ (92.77 x 100) / 2037 = 4.5 per cent, of the whole area of the town.
It must be remembered that when rain continues for long periods, ground which in the ordinary way would generally be considered permeable becomes soaked and eventually becomes more or less impermeable. Mr. D. E. Lloyd-Davies, M.Inst.C.E., gives two very interesting diagrams in the paper previously referred to, which show the average percentage of effective impermeable area according to the population per acre. This information, which is applicable more to large towns, has been embodied in Fig. 16, from which it will be seen that, for storms of short duration, the proportion of impervious areas equals 5 per cent. with a population of 4.9 per acre, which is a very close approximation to the 4.5 per cent. obtained in the example just described.
Where the houses are scattered at long intervals along a road the better way to arrive at an estimate of the quantity of storm water which may be expected is to ascertain the average impervious area of, or appertaining to, each house, and divide it by five, so as to get the area per head. Then the flow off from any section of road is directly obtained from the sum of the impervious area due to the length of the road, and that due to the population distributed along it.
[Illustration: FIG. 16.—VARIATION IN AVERAGE PERCENTAGE OFEFFECTIVE IMPERMEABLE AREA ACCORDING TO DENSITY OF POPULATION.]
In addition to being undesirable from a sanitary point of view, it is rarely economical to construct special storm water drains, but in all cases where they exist, allowance must be made for any rain that may be intercepted by them. Short branch sewers constructed for the conveyance of foul water alone are usually 9in or 12 in in diameter, not because those sizes are necessary to convey the quantity of liquid which may be expected, but because it is frequently undesirable to provide smaller public sewers, and there is generally sufficient room for the storm water without increasing the size of the sewer. If this storm water were conveyed in separate sewers the cost would be double, as two sewers would be required in the place of one. In the main sewers the difference is not so great, but generally one large sewer will be more economical than two smaller ones. Where duplicate sewers are provided and arranged, so that the storm water sewer takes the rain-water from the roads, front roofs and gardens of the houses, and the foul water sewer takes the rain-water from the back roofs and paved yards,
it was found in the case previously worked out in detail that in built-up roads a width of 36 ft + 2 (8 ft 7 in) = 53 ft 2 in, or, say, 160 sq. ft per lineal yard of road would drain to the storm water sewer, and a width of 2 (6 ft 10 in) = 13 ft 8 in, or, say, 41 sq. ft per lineal yard of road to the foul water sewer. This shows that even if the whole of the rain which falls on the impervious areas flows off, only just under 80 per cent. of it would be intercepted by the special storm water sewers. Taking an average annual rainfall of 30 in, of which 75 per cent. flows off, the quantity reaching the storm water sewer in the course of a year from each lineal
30 75yard of road would be —- x 160 x —- = 300 cubic12 100feet = 1,875 gallons.
[Illustration: FIG. 17.—SECTION OF "LEAP WEIR" OVERFLOW]
The cost of constructing a separate surface water system will vary, but may be taken at an average of, approximately, l5s. 0d. per lineal yard of road. To repay this amount in thirty years at 4 per cent, would require a sum of 10.42d., say 10-1/2d. per annum; that is to say, the cost of taking the surface water into special
10-1/2 d. x 1000sewers is ———————— = 5.6, say 6d. per 1,0001875gallons.
If the sewage has to be pumped, the extra cost of pumping by reason of the increased quantity of surface water can be looked at from two different points of view:—
1. The net cost of the gas or other fuel or electric current consumed in lifting the water.
2. The cost of the fuel consumed plus wages, stores, etc., and a proportion of the sum required to repay the capital cost of the pumping station and machinery.
The extra cost of the sewers to carry the additional quantity of storm water might also be taken into account by working out and preparing estimates for the alternative schemes.
The actual cost of the fuel may be taken at approximately 1/4 d. per 1,000 gallons. The annual works and capital charges, exclusive of fuel, should be divided by the normal quantity of sewage pumped per annum, rather than by the maximum quantity which the pumps would lift if they were able to run continuously during the whole time. For a town of about 10,000 inhabitants these charges may be taken at 1-1/4 d. per 1,000 gallons, which makes the total cost of pumping, inclusive of capital charges, 1-1/2 d. per 1,000 gallons. Even if the extra cost of enlarging the sewers is added to this sum it will still be considerably below the sum of 6 d., which represents the cost of providing a separate system for the surface water.
Unless it is permissible for the sewage to have a free outlet to the sea at all states of the tide, the provision of effective storm overflows is a matter of supreme importance. Not only is it necessary for them to be constructed in well- considered positions, but they must be effective in action. A weir constructed along one side of a manhole and parallel to the sewer is rarely efficient, as in times of storm the liquid in the sewer travels at a considerable velocity, and the greater portion of it, which should be diverted, rushes past the weir and continues to flow in the sewer; and if, as is frequently the case, it is desirable that the overflowing liquid should be screened, and vertical bars are fixed on the weir for the purpose, they block the outlet and render the overflow practically useless.
Leap weir overflows are theoretically most suitable for separating the excess flow during times of storm, but in practice they rarely prove satisfactory. This is not the fault of the system, but is, in the majority of the cases, if not all, due to defective designing. The general arrangement of a leap weir overflow is shown in Fig. 17. In normal circumstances the sewage flowing along the pipe A falls down the ramp, and thence along the sewer B; when the flow is increased during storms the sewage from A shoots out from the end of the pipe into the trough C, and thence along the storm-water sewer D. In order that it should be effective the first step is to ascertain accurately the gradient of the sewer above the proposed overflow, then, the size being known, it is easy to calculate the velocity of flow for the varying depths of sewage corresponding with minimum flow, average dry weather flow, maximum dry weather flow, and six times the dry weather flow. The natural curve which the sewage would follow in its downward path as it flowed out from the end of the sewer can then be drawn out for the various depths, taking into account the fact that the velocity at the invert and sides of the sewer is less than the average velocity of flow. The ramp should be built in accordance with the calculated curves so as to avoid splashing as far as possible, and the level of the trough C fixed so that when it is placed sufficiently far from A to allow the dry weather flow to pass down the ramp it will at the same time catch the storm water when the required dilution has taken place. Due regard must be had to the altered circumstances which will arise when the growth of population occurs, for which provision is made in the scheme, so that the overflow will remain efficient. The trough C is movable, so that the width of the leap weir may be adjusted from time to time as required. The overflow should be frequently inspected, and the accumulated rubbish removed from the trough, because sticks and similar matters brought down by the sewer will probably leap the weir instead of flowing down the ramp with the sewage. It is undesirable to fix a screen in conjunction with this overflow, but if screening is essential the operation should be carried out in a special manhole built lower down the course of the storm-water sewer. Considerable wear takes place on the ramp, which should, therefore, be constructed of blue Staffordshire or other hard bricks. The ramp should terminate in a stone block to resist the impact of the falling water, and the stones which may be brought with it, which would crack stoneware pipes if such were used.
In cases where it is not convenient to arrange a sudden drop in the invert of the sewer as is required for a leap weir overflow, the excess flow of storm-water may be diverted by an arrangement similar to that shown in Fig. 18. [Footnote: PLATE IV] In this case calculations must be made to ascertain the depth at which the sewage will flow in the pipes at the time it is diluted to the required extent; this gives the level of the lip of the diverting plate. The ordinary sewage flow will pass steadily along the invert of the sewer under the plate until it rises up to that height, when the opening becomes a submerged orifice, and its discharging capacity becomes less than when the sewage was flowing freely. This restricts the flow of the sewage, and causes it to head up on the upper side of the overflow in an endeavour to force through the orifice the same quantity as is flowing in the sewer, but as it rises the velocity carries the upper layer of the water forward up the diverting plate and thence into the storm overflow drain A deep channel is desirable, so as to govern the direction of flow at the time the overflow is in action. The diverting trough is movable, and its height above the invert can be increased easily, as may be necessary from time to time. With this arrangement the storm-water can easily be screened before it is allowed to pass out by fixing an inclined screen in the position shown in Fig. 18. [Footnote: PLATE IV] It is loose, as is the trough, and both can be lifted out when it is desired to have access to the invert of the sewer. The screen is self- cleansing, as any floating matter which may be washed against it does not stop on it and reduce its discharging capacity, but is gradually drawn down by the flow of the sewage towards the diverting plate under which it will be carried. The heavier matter in the sewage which flows along the invert will pass under the plate and be carried through to the outfall works, instead of escaping by the overflow, and perhaps creating a nuisance at that point.
In small sewerage schemes where pumping is necessary the amount expended in the wages of an attendant who must give his whole attention to the pumping station is so much in excess of the cost of power and the sum required for the repayment of the loan for the plant and buildings that it is desirable for the economical working of the scheme to curtail the wages bill as far as possible. If oil or gas engines are employed the man cannot be absent for many minutes together while the machinery is running, and when it is not running, as for instance during the night, he must be prepared to start the pumps at very short notice, should a heavy rain storm increase the flow in the sewers to such an extent that the pump well or storage tank becomes filled up. It is a simple matter to arrange floats whereby the pump may be connected to or disconnected from a running engine by means of a friction clutch, so that when the level of the sewage in the pump well reaches the highest point desired the pump may be started, and when it is lowered to a predetermined low water level the pump will stop; but it is impracticable to control the engine in the same way, so that although the floats are a useful accessory to the plant during the temporary absence of the man in charge they will not obviate his more or less constant attendance. An electric motor may be controlled by a float, but in many cases trouble is experienced with the switch gear, probably caused by its exposure to the damp air. In all cases an alarm float should be fixed, which would rise as the depth of the sewage in the pump well increased, until the top water level was reached, when the float would make an electrical contact and start a continuous ringing warning bell, which could be placed either at the pumping station or at the man's residence. On hearing the bell the man would know the pump well was full, and that he must immediately repair to the pumping-station and start the pumps, otherwise the building would be flooded. If compressed air is available a hooter could be fixed, which would be heard for a considerable distance from the station.
[Illustration: PLATE IV.
To face page 66.]
It is apparent, therefore, that a pumping machine is wanted which will work continuously without attention, and will not waste money when there is nothing to pump. There are two sources of power in nature which might be harnessed to give this result—water and wind. The use of water on such a small scale is rarely economically practicable, as even if the water is available in the vicinity of the pumping-station, considerable work has generally to be executed at the point of supply, not only to store the water in sufficient bulk at such a level that it can be usefully employed, but also to lead it to the power-house, and then to provide for its escape after it has done its work. The power-house, with its turbines and other machinery, involves a comparatively large outlay, but if the pump can be directly driven from the turbines, so that the cost of attendance is reduced to a minimum, the system should certainly receive consideration.
Although the wind is always available in every district, it is more frequent and powerful on the coast than inland. The velocity of the wind is ever varying within wide limits, and although the records usually give the average hourly velocity, it is not constant even for one minute. Windmills of the modern type, consisting of a wheel composed of a number of short sails fixed to a steel framework upon a braced steel tower, have been used for many years for driving machinery on farms, and less frequently for pumping water for domestic use. In a very few cases it has been utilised for pumping sewage, but there is no reason why, under proper conditions, it should not be employed to a greater extent. The reliability of the wind for pumping purposes may be gauged from the figures in the following table, No. 11, which were observed in Birmingham, and comprise a period of ten years; they are arranged in order corresponding with the magnitude of the annual rainfall:—
TABLE No. 11.
Reference | Rainfall |Number of days in year during which the mean |Number | for |hourly velocity of the wind was below || year | 6 m.p.h. | 10 m.p.h. | 15 m.p.h. | 20 m.p.h. |—————+—————+—————+—————-+—————-+—————-+1… 33·86 16 88 220 3142… 29·12 15 120 260 3343… 28·86 39 133 263 3364… 26·56 36 126 247 3235… 26·51 34 149 258 3306… 26·02 34 132 262 3337… 25·16 33 151 276 3328… 22·67 46 155 272 3299… 22·30 26 130 253 33710… 21·94 37 133 276 330—————+—————+—————+—————-+—————-+—————-+Average 31·4 131·7 250·7 330·8
It may be of interest to examine the monthly figures for the two years included in the foregoing table, which had the least and the most wind respectively, such figures being set out in the following table:
TABLE No. 12
Number of days in each month during which the mean velocity of the wind was respectively below the value mentioned hereunder.
Month | Year of least wind (No. 8) | Year of most wind (No. *8*) || 5 10 15 20 | 5 10 15 20 || m.p.h. m.p.h. m.p.h. m.p.h. | m.p.h. m.p.h. m.p.h. m.p.h. |———+———-+——-+———-+———-+———-+———+———+———-+Jan. 5 11 23 27 3 6 15 23Feb. 5 19 23 28 0 2 8 16Mar. 5 10 20 23 0 1 11 18April 6 16 23 28 1 7 16 26May 1 14 24 30 3 11 24 31June 1 12 22 26 1 10 21 27July 8 18 29 31 1 12 25 29Aug. 2 9 23 30 1 9 18 30Sept. 1 13 25 30 1 12 24 28Oct. 5 17 21 26 0 4 16 29Nov. 6 11 20 26 3 7 19 28Dec. 1 5 19 24 2 7 23 29———+———-+——-+———-+———-+———-+———+———+———-+Total 46 155 272 329 16 88 220 314
During the year of least wind there were only eight separate occasions upon which the average hourly velocity of the wind was less than six miles per hour for two consecutive days, and on two occasions only was it less than six miles per hour on three consecutive days. It must be remembered, however, that this does not by any means imply that during such days the wind did not rise above six miles per hour, and the probability is that a mill which could be actuated by a six-mile wind would have been at work during part of the time. It will further be observed that the greatest differences between these two years occur in the figures relating to the light winds. The number of days upon which the mean hourly velocity of the wind exceeds twenty miles per hour remains fairly constant year after year.
As the greatest difficulty in connection with pumping sewage is the influx of storm water in times of rain, it will be useful to notice the rainfall at those times when the wind is at a minimum. From the following figures (Table No. 13) it will be seen that, generally speaking, when there is very little wind there is very little rain Taking the ten years enumerated in Table No. 11, we find that out of the 314 days on which the wind averaged less than six miles per hour only forty-eight of them were wet, and then the rainfall only averaged .l3 in on those days.
TABLE No. 13.
—————-+——————-+——————+————+—————————————————Ref. No. | Total No. | Days on | | Rainfall on eachfrom Table | of days in | which no | Rainy | rainy day inNo. 11. | each year. | rain fell. | days. | inches.—————-+——————-+——————+————+—————————————————1 | 16 | 14 | 2 | .63 and .2452 | 15 | 13 | 2 | .02 and .023 | 39 | 34 | 5 | .025, .01, .26, .02 and .034 | 36 | 29 | 7 | / .02, .08, .135, .10, .345, .18| | | | \ and .025 | 34 | 28 | 6 | .10, .43, .01, .07, .175 and .076 | 32 | 27 | 5 | .10, .11, .085, .04 and .1357 | 33 | 21 | 2 | .415 and .708 | 46 | 40 | 6 | .07, .035, .02, .06, .13 and .029 | 26 | 20 | 6 | .145, .20, .33, .125, .015 & .07510 | 37 | 30 | 7 | / .03, .23, .165, .02, .095| | | | \ .045 and .02—————-+——————-+——————+————+—————————————————Total | 314 | 266 | 48 | Average rainfall on each of| | | | the 48 days = .13 in
The greater the height of the tower which carries the mill the greater will be the amount of effective wind obtained to drive the mill, but at the same time there are practical considerations which limit the height. In America many towers are as much as 100 ft high, but ordinary workmen do not voluntarily climb to such a height, with the result that the mill is not properly oiled. About 40 ft is the usual height in this country, and 60 ft should be used as a maximum.
Mr. George Phelps, in a paper read by him in 1906 before the Association of Water Engineers, stated that it was safe to assume that on an average a fifteen miles per hour wind was available for eight hours per day, and from this he gave the following figures as representing the approximate average duty with, a lift of l00 ft, including friction:—
Diameter of Wheel.
10
12
14
16
18
20
25
30
35
40
The following table gives the result of tests carried out by the United States Department of Agriculture at Cheyenne, Wyo., with a l4 ft diameter windmill under differing wind velocities:—
TABLE No. 15.
POWER or l4-rx WINDMILL IN VARYING WINDS.
Velocity of Wind (miles per hour).
0—5 6-10 11-15 16-20 21-25 26-30 31-35
It will be apparent from the foregoing figures that practically the whole of the pumping for a small sewerage works may be done by means of a windmill, but it is undesirable to rely entirely upon such a system, even if two mills are erected so that the plant will be in duplicate, because there is always the possibility, although it may be remote, of a lengthened period of calm, when the sewage would accumulate; and, further, the Local Government Board would not approve the scheme unless it included an engine, driven by gas, oil, or other mechanical power, for emergencies. In the case of water supply the difficulty may be overcome by providing large storage capacity, but this cannot be done for sewage without creating an intolerable nuisance. In the latter case the storage should not be less than twelve hours dry weather flow, nor more than twenty-four. With a well-designed mill, as has already been indicated, the wind will, for the greater part of the year, be sufficient to lift the whole of the sewage and storm-water, but, if it is allowed to do so, the standby engine will deteriorate for want of use to such an extent that when urgently needed it will not be effective. It is, therefore, desirable that the attendant should run the engine at least once in every three days to keep it in working order. If it can be conveniently arranged, it is a good plan for the attendant to run the engine for a few minutes to entirely empty the pump well about six o'clock each evening. The bulk of the day's sewage will then have been delivered, and can be disposed of when it is fresh, while at the same time the whole storage capacity is available for the night flow, and any rainfall which may occur, thus reducing the chances of the man being called up during the night. About 22 per cent, of the total daily dry weather flow of sewage is delivered between 7 p.m. and 7 a.m.
The first cost of installing a small windmill is practically the same as for an equivalent gas or oil engine plant, so that the only advantage to be looked for will be in the maintenance, which in the case of a windmill is a very small matter, and the saving which may be obtained by the reduction of the amount of attendance necessary. Generally speaking, a mill 20 ft in diameter is the largest which should be used, as when this size is exceeded it will be found that the capital cost involved is incompatible with the value of the work done by the mill, as compared with that done by a modern internal combustion engine.
Mills smaller than 8 ft in diameter are rarely employed, and then only for small work, such as a 2 1/2 in pump and a 3-ft lift The efficiency of a windmill, measured by the number of square feet of annular sail area, decreases with the size of the mill, the 8 ft, 10 ft, and l2 ft mills being the most efficient sizes. When the diameter exceeds l2 ft, the efficiency rapidly falls off, because the peripheral velocity remains constant for any particular velocity or pressure of the wind, and as every foot increase in the diameter of the wheel makes an increase of over 3 ft in the length of the circumference, the greater the diameter the less the number of revolutions in any given time; and consequently the kinetic flywheel action which is so valuable in the smaller sizes is to a great extent lost in the larger mills.
Any type of pump can be used, but the greatest efficiency will be obtained by adopting a single acting pump with a short stroke, thus avoiding the liability, inherent in a long pump rod, to buckle under compression, and obviating the use of a large number of guides which absorb a large part of the power given out by the mill. Although some of the older mills in this country are of foreign origin, there are several British manufacturers turning out well-designed and strongly-built machines in large numbers. Fig. 19 represents the general appearance and Fig. 20 the details of the type of mill made by the well-known firm of Duke and Ockenden, of Ferry Wharf, Littlehampton, Sussex. This firm has erected over 400 windmills, which, after the test of time, have proved thoroughly efficient. From Fig. 20 it will be seen that the power applied by the wheel is transmitted through spur and pinion gearing of 2 1/2 ratio to a crank shaft, the gear wheel having internal annular teeth of the involute type, giving a greater number of teeth always in contact than is the case with external gears. This minimises wear, which is an important matter, as it is difficult to properly lubricate these appliances, and they are exposed to and have to work in all sorts of weather.
[Illustration: Fig. l9.—General View of Modern Windmill.]
[Illustration: Fig. 20.—Details of Windmill Manufactured by Messrs. Duke andOckenden, Littlehampton.]
It will be seen that the strain on the crank shaft is taken by a bent crank which disposes the load centrally on the casting, and avoids an overhanging crank disc, which has been an objectionable feature in some other types. The position of the crank shaft relative to the rocker pin holes is studied to give a slow upward motion to the rocker with a more rapid downward stroke, the difference in speed being most marked in the longest stroke, where it is most required.
In order to transmit the circular internal motion a vertical connecting rod in compression is used, which permits of a simple method of changing the length of stroke by merely altering the pin in the rocking lever, the result being that the pump rod travels in a vertical line.
The governing is entirely automatic. If the pressure on the wind wheel, which it will be seen is set off the centre line of the mill and tower, exceeds that found desirable—and this can be regulated by means of a spring on the fantail—the windmill automatically turns on the turn-table and presents an ellipse to the wind instead of a circular face, thus decreasing the area exposed to the wind gradually until the wheel reaches its final position, or is hauled out of gear, when the edges only are opposed to the full force of the wind. The whole weight of the mill is taken upon a ball-bearing turn-table to facilitate instant "hunting" of the mill to the wind to enable it to take advantage of all changes of direction. The pump rod in the windmill tower is provided with a swivel coupling, enabling the mill head to turn completely round without altering the position of the rod.