Table 28.7 Lapworth yield-storage relationships

becomes sensitive to the dry weather flow (DWF) as in the first two examples above and special droughts may 'wreck' the calculations. The percentage of runoff taken for the yield and storage depends on the physical conditions of each site and on the daily requirements of each town.

Storage (cm) Average runoff from gathering ground (cm) 72.5 12.5 26.3 37.5 46.3 52.5 57.5 62.5 67.5

25 50 75 100 125 150 175 200

25

37.5

50

18.8 22.5 25.0 35.0 40.5 45.0 49.3 56.3 61.3 61.3 71.3 77.5 72.5 84.8 92.5 80.8 96.8 108.3 88.0 107.3 120.5 95.5 115.0 132.5

62.5

75

50.0 66.3 82.5 100.0 117.5 132.5 145.0

70.0 87.5 105.5 125.0 142.5 157.5

Conversions: Storage: 1 cm of water on 1 ha= 1000001 Runoff: 1 cm per annum on 1 ha = 274 I/day

The following notes may be useful in making a preliminary assessment.

Example of use of Lapworth Chart and Table 28.7. Assume a gathering ground of 1000 ha with an average annual rainfall of 175 cm. Assume average annual evaporation is 50 cm so that average annual runoff= 175 - 50= 125 cm. From Table 28.7 for average annual runoff of 125cm, consider storage of 25, 50 and 75 cm and tabulate as below. Storage related to 125 cm Corresponding yield (cm) From conversion Storage (1 per h) Therefore Ml/1000 ha Yield (Ml/day)

25 72.5

50 92.5

2.5 x 106 5 x 106 2000 5000 20.0 25.3

75 105.5 7.5 x 106 7500 28.9

26.8.4 Catchwaters

(1) The average annual rainfall. The records of the Meteorological Office should be referred to for any area in Britain. Gathering grounds in the South and Midlands, 750 to 1000mm, Wales, Lake District to Scotland 1000 to 1500mm per annum. In other parts of the world best use must be made of often sparse data. (2) The evaporation or loss of the rainfall. Penman22 has compiled a map of Britain showing the average annual losses, and similar maps can be compiled by application of the formula using local factors in other parts of the world. (3) The average annual runoff is the difference between rainfall and evaporation on the gathering ground. (4) For the yield-storage relation, reference for a first approximation should be made either to the Lapworth chart,24 from which Table 28.7 below gives extracts of the yield-storage relation for various runoffs. (The rainfall less evaporation is also known as the available yield.)

In order to augment the yield of a reservoir, catchwaters are often resorted to so that additional water is led into a reservoir from another gathering ground. As it is not economical to design catchwaters to take maximum floods, only up to 90% of the water available is taken. The proportion taken is regarded as the 'efficiency' of the catchwater and can be assessed by drawing up a flow frequency curve25 relating to the stream at the intake point if such records are available. A catchwater may be a tunnel or, more often, an open channel graded to suit the contour of the land. This is the cheapest form of structure and, provided that the length is not too great, it is cheaper than a pipe. The design of an open channel can be based on the Chezy formula.

The determination of the size of a reservoir required to balance 50% or less of the available yield is difficult as the problem

Reservoirs may be formed with earthen (or rockfill) dams, concrete dams (gravity, arch, multiple arch, cupola, buttress,

28.9 Formation of reservoirs

Table 28.8 Dam categories and design flood factors Category

Result of failure

Initial reservoir condition

Flood inflow (general standard)

Wind speed

Minimum wave surcharge (m)

A

Lives in a community endangered Extensive damage, or lives not endangered in a community Negligible risk to life. Some damage No loss of life likely. Limited damage

Spilling

PMF

1 in 10 yr Max. h

0.6

Just full

0.5 PMF or 1/10 000 yr

1 in 10 yr Max. h

0.6

Just full

0.3 PMF

Av. annual

0.4

Spilling

0.2 PMF or 150-yr flood

Av. annual Max. h

0.3

B

C D

reinforced, and prestressed), by raising existing lakes or by enclosing with artificial embankments, generally filled by pumping. In addition to the dam forming the reservoir, certain ancillary works such as drawoff tower, overflow and diversion works are normally required. 28.9.1 Valve towers A water supply reservoir usually has a valve tower to contain pairs of valves at different levels, to draw off the water as it rises and falls and to ensure the water at that level is consistent with good quality. Recording instruments are often situated within the valve tower: (1) to record the level of the water below top water level; (2) to record the level of the water when the reservoir is overflowing; (3) to record the quantity of water taken to supply; (4) to record the quantity of water discharged for compensation. A rain gauge is usually placed in the vicinity of the dam. 28.9.2 Floods in reservoir practice Earlier assessments of the size of floods were made based on valuable data recorded in the Institution of Civil Engineers (ICE) interim reports of 1933 and 1960, supplemented by local data when available and by the subjective judgement of experienced engineers. However, since 1975 these methods have been superseded by the Flood studies report1* produced by the Institute of Hydrology with the ICE and the Meteorological Office. This introduced the concepts of probable maximum precipitation (PMP) and probable maximum flood (PMF). Probable maximum precipitation is defined as the theoretical greatest depth of precipitation (i.e. rain, sleet, snow or hail) for a given duration meteorologically possible for a given basin at a particular period. The flood hydrograph resulting from PMP is called the PMF and is assessed from local topographic and land use parameters. Detailed maps have been prepared for the report and these are continuously updated for Britain, and similar work is in hand for other parts of the world. As the likelihood of a PMF occurring may be only once in 50000 years, it would only be utilized in design if a failure arising from it would endanger lives in a community. For lesser consequences proportions of PMF are recommended in the ICE guide to Floods and reservoir safety21 and these are summarized in Table 28.8 combined with associated design factors. In Britain and Hong Kong, the decision on the category of dam and design flood is made by an engineer appointed under the Reservoirs (Safety Provisions) Act, 1930, or the Reservoirs Act 1975 when implemented in full, taking into account the

effect of flood routing through the reservoir, the capacity of the spilling, available freeboard and the risk involved in the event of overtopping. To obtain an approximate estimate of PMF for preliminary assessment of dam safety, typical curves have been prepared based on an undulating impermeable catchment. For other types of terrain add 15% for mountainous areas, 5% for hilly areas, deduct 5% for flatter areas and adjust total area for permeable zones. The reservoir soil master deficit (RSMD) is an index defined as the 1-day rainfall of 5-yr return period less effective mean soil master deficit; typical values given in the Flood studies report for Britain are 25 to 35 in the Midlands and Southeast, up to 70 in the Lake District, up to 90 in the Scottish Highlands. Similar figures can be computed from first principles for other parts of the world. Some typical figures are shown in Table 28.9 for general reference. 28.9.3 Overflow weirs The function of the overflow is to carry the design peak flood safely over the dam. Overflow weirs may be formed along the crest of the dam, or as side weirs on one or both sides of the valley upstream of the dam, or as a series of siphons over the crest of the dam. Whatever form is adopted to fit in with the design of the dam, the weir should be capable of passing the design flood without overtopping of the main structure, though some limited overtopping can be accepted for lower categories of dam in the case of concrete dams or those with protected crest and downstream slope, often used in flood detention situations. 28.9.4 Drawoff and diversion culverts Tunnels or culverts preferably constructed around the ends of the dam in solid strata are used for diverting streams during the construction of reservoirs, and their dimensions during construction depend upon the best approximation of the magnitude of the flood. In the case of bell-mouth overflows the tunnel is required to carry water which flows over the bellmouth permanently. Frequently the same tunnel is used to carry water drawn off from the reservoir to supply or for regulation. 28.9.5 Earthen embankments and earth dams 28.9.5.1 General The adoption of an earthen embankment for an impounding

Table 28.9 Approximate probable maximum flood assessment Catchment area (km2)

RSMD (mm)

PM F peak flow (cumecs/km2)

Catchment area (km2)

RSMD (mm)

PMF peak flow (cumecs/km2)

1 1 1 5 5 5 5

25 40 70 25 50 75 100

9.2 14.5 26.0 6.5 13.5 20.5 27.5

10 10 10 10 50 50 50 50

25 50 75 100 25 50 75 100

5.7 11.5 17.7 23.6 4.1 8.4 12.9 17.2

reservoir is largely a matter of choice, depending on the geological factors governing the site. If clay predominates at the site of the dam, an earthen embankment in all probability will be adopted, whereas in rocky country a masonry or concrete dam would be more suitable. This generalization is not rigid, for on the granite of Cornwall in the same valley there is a concrete dam upstream with an earthen embankment downstream. The earthen embankment usually depends upon a puddled or rolled clay core for watertightness, both in the trench below ground and in the body of the embankment above ground, although a concrete-filled cut-off trench was frequently used before grouting techniques were available. It is now generally ruled out on account of cost and permeable strata are sealed by grouting with cement or chemical mixtures from ground level or occasionally from a gallery or crest level after completion of construction. Both upstream and downstream, earthen embankments depend for stability on adequate drainage by rubble and 'selected' material (known as 'filters', specially graded to certain rules) which is placed against the core or laid in layers in the ordinary filling. Broken rock, or material containing a large percentage of broken rock, permits steeper slopes to be adopted; thus for the upstream slope 1 (vertical) in 3 or 4 (horizontal) might suffice, and for that downstream 1 in 2 or 3; whereas, for the ordinary clayey materials so common in this country, 1 in 5 or 6 or more upstream, and 1 in 3 to 5 downstream are common. The guiding principle in such embankments is adequate drainage, as well as the application of the principles of soil mechanics28 such as tests for shear strength and other properties of clays, and deduced slip planes and stability diagrams.29 Bearing pressures of the subsoil below the embankment must be known if the weight of the embankment will be sustained without subsidence. 28.9.5.2 Cost of earth dams It is useful to estimate the approximate cost of a dam when carrying out a comparative study of different sites and Mitchell30 has suggested a formula, subsequently modified by Whincap as follows: Cost (£ sterling) = 3.6aLH2 + Q.6bLH2 + 2.OcLD + 66 000//+ 202 000 (1) (2) (3) (4) (5) where L is the crest length, H is the mean height from ground to crest (i.e. area of cross-section of valley divided by length of crest of chard across the valley), D is the mean depth of cutoff trench (all in metres). The numbers in brackets below the items in the formula define the different sections as follows: (1) Cost of forming the embankment, where a is mean rate per cubic metre. (2) Represents the extra cost of rolled clay core where b is the rate per cubic metre. (Typical examples indicate that b varies from zero to one-third of a.) (3) Cost of a concrete-filled cutoff trench 2 m wide, c is the rate per cubic metre; not often required. If rolled clay cutoff used, included in (2). (4) Cost of stream diversion, overflow and valve shaft. (5) Miscellaneous items such as pipework and valves, access to valve shaft, reinforcement and steelwork, recording instruments. Items (4) and (5) are based on typical 1985 values, and appropriate 1985 values for factors a, b and c would be £16.50, zero to £5.50 and £124 respectively.

It should be noted that the formula does not include a number of common items in comparative studies such as contingencies, engineering design and supervision, land, diversion of utilities, site clearance, access roads and bridges, excavation of unsuitable material, grouting or construction camp accommodation. The first three of these items generally amount to approximately 25% of the formula cost, but the remainder needs to be priced separately as they are particular to a site. 28.9.5.3 Freeboard For earthen dams and embankments it is important that the 'freeboard', the distance between top-water level and crest, should be adequate; it depends on the 'fetch' (the maximum distance the water is impounded at right angles to the dam), the density of trees and vegetation, the elevation of the site, intensity and direction of wind and, of course, a correct assessment of the floods.27 28.9.5.4 Design of earth dams Some of the basic factors considered in design are:31 (1) To ascertain by site investigation and laboratory testing of the materials, available for constructing the dam, preferably those nearest to the site for economy. (2) To ascertain the conditions and properties of the strata under the embankment to resist sliding or slipping. (3) To analyse the factor of safety over slipping for the particular dam section by choosing a slip plane or circle through the probable weakest line of failure in and under the dam. This can now be done by one of several computer programs. Common slopes of embankments vary between 1:1 and 1:6. 28.9.5.5 Slurry trench Certain Bentonite clays or slurries have been used for sinking dam trenches in soft strata without timbering. The clays have the effect of remaining liquid when the trench is being dug but form a gel or colloid (slightly heavier than water) when not disturbed which has the effect of keeping the walls of the trench from falling in. Recently, Bentonite cement mixtures have been used to effect a permanent cutoff seal in porous strata below embankments as at Bewl Bridge,32 in Kent. 28.9.5.6 Pore pressure under earth dams Where embankments are constructed of, or rest on, cohesive materials containing water within the pores, a pore pressure is set up within the material when loaded either during construction or on the subsequent filling of the reservoir. The increase in pore pressure leads to a reduction in the shear strength of the material and a corresponding reduction in the factor of safety which can lead to failure of the embankment by slipping. To obviate this effect, sloping layers or 'blankets' of drainage material are included in the embankments to enable the pore pressure to be dissipated before it reaches a dangerous level. A similar effect can be observed by infiltration of groundwater and if this is anticipated then relief wells or vertical sand drains should be incorporated under the embankment. Pore pressure can be measured by installing piezometers, which are ceramic pots sealed in the strata and connected to a gauge by fine tubes. 28.9.5.7 Deformation of earth dams Under this term are included the causes of the shapeless early

nineteenth-century embankments often seen before the Safety Provisions Act, 1930. These include: (1) subsidence of the crest - some have sunk nearly below the overflow top water level; (2) irregular shape of the embankment due to local slips; and (3) irregular toe lines due to slides. Although sagging of the crest can be remedied by levelling and adding additional material, and irregular shapes and toes can be regraded and re-aligned, little or nothing is known of what is going on in the strata inside a dam and, since it is dangerous to take such measures without analysing the original cause of the deformation, instruments have been developed to indicate both horizontal and vertical movements. Horizontal movements can now be measured by vibrations in an electrically stimulated wire (embedded between two concrete blocks) whose tension varies with the horizontal stress. For vertical movements, the relative displacement can be ascertained by lowering an induction coil down through a tube connecting two plates, or by various types of instruments using the U-tube principle. A fuller exposition on instrumentation in earth dams is given by Rofe and Tye.33 For schemes including recent earthen embankments, see also works by Hallas and Titford,34 Picken,35 Walters and Walton,36 Kennard and Kennard,37 Kennard and Crann.38 28.9.5.8 Compaction of earth and rockfill dams In earth dams it is important to have the impervious clay core well compacted either by heavy rollers or light vibrating rollers operating on layers of 250 to 500 m. The 'voids' percentage should be kept below 5%. The compaction of the rest of the 'fill' need not be to such a high specification. In a case such as Scammonden Dam (height about 8Om) an earth and rockfill dam near Huddersfield, a special investigation had to be made to ensure minimum settlement because a six-lane motorway goes over the crest.40 For this particular site (consisting of alternating grits, sandstones and shales of the Millstone Grit formation) extensive large-scale experiments revealed that the best compaction depended upon: (1) how the rock was quarried; (2) how the material was deposited on the embankment; and (3) the best type of compacting machines. 28.9.6 Concrete dams The adoption of concrete for constructing a dam depends on the topography and geology of the site. The trench, as in the case of an earthen embankment, may be filled either with puddled clay or concrete, depending on the hardness or softness of the strata penetrated and the cost of filling it either with clay or concrete. Below the trench there may be a necessity for extensive grouting, to reduce leakage under the dam. The surface for the broad foundation for the superstructure must also be prepared. These foundation works may cost as much as the superstructure seen above ground. For a straight gravity concrete dam 30m high, the broad foundation would be about 18m wide, and if rock or other stable formation is not found reasonably near (say 7 m below) the surface, considerable expense in foundation work may also be entailed. For a buttress or multiple-arch darn, it would be less, but the strata sustaining the buttresses would have to be stronger. Most dam failures may be attributed to faulty foundations. Above ground, the gravity dam is so-called because any section can stand by itself because of its weight. Concrete is generally vibrated, in dam construction particularly, to eliminate air pockets, prevent leakage and increase speed of setting. Shuttering must be especially well constructed to withstand vibration during construction.

Curved-on-plan gravity dams are sometimes substituted for straight dams for aesthetic reasons, but the gravity section of the dam cannot be reduced.41 28.9.6.1 Cost of concrete gravity dams31 For comparative studies the cost can be illustrated by the following formula originally devised by Mitchell. Cost = £(0.375xL/f2 + 0.675xL W2 + 0.7SyLD 1 + 2.OzLZ)2 + 152 000 (1)

(2)

(3)

(4)

(5)

where H is the mean height from broad foundation to crest in metres, i.e. area of cross-section of valley divided by length of crest or chord across valley, L is the length of crest, in metres, W the width of dam at crest in metres, Z)1 the mean depth of broad foundation below ground level in metres and D2 the mean depth of cutoff trench below broad foundation in metres. Item (1) Cost of concrete in dam where x is rate per cubic metre (£82.50 at 1985 prices). Item (2) Cost of crest road or footpath. Item (3) Cost of excavating broad foundation where y is rate per cubic metre (£8 at 1985 prices). Item (4) Cost of concrete-filled cutoff trench 2 m wide where z is rate per cubic metre (£99 at 1985 prices). Item (5) Ancillary works such as pipework and valves, reinforcement and steelwork and recording instruments for water level, overflow and discharge below dam. Note: Like the corresponding formula for earth dams, this formula is intended for comparing a number of sites when making a preliminary survey of alternative sources, and does not include the other factors described in that context. 28.9.6.2 Buttress and multiple-arch dams The buttresses of a buttress dam form part of two adjacent halves of two arches (thick) which act as cantilevers and hence if the bearing pressure of one buttress differs relatively from the other, movement may occur at the centre of the arch, where an expansion joint is (or should be) inserted. The internal buttresses of the multiple-arch dam are merely blocks of concrete acting as abutments for supporting the two halves of two rigid (thin) arches. Hence, the foundations for the buttresses of the multiple-arch dam must have equal bearing capacity to avoid fracture of the true thin arches; whereas in those for the buttress dam, some inequality is taken care of by the expansion joint between the two cantilever arms of the buttresses. Thin-arch dam. The true thin-arch concrete dam is suitable for the valley which has a good foundation and where the width at the level of the dam crest is not more than 3 times the proposed maximum height of the dam. The volume of concrete in an arch dam is about half that in a comparable gravity dam. Preliminary calculations are directed to finding the thickness / in metres of the dam at any depth in terms of the water pressure P (i.e. on a metre strip of dam), and the radius R of the upstream face in metres and the compressive strength S. If the abutment pressure is greater than the compressive strength of the strata on which the abutment rests, the concrete should be increased in width. Thick-arch dam. The thick-arch dam lies between the thicknesses of the gravity and arch dams. It has been adopted in valleys with chord:height ratios up to 5 or 6. The theory of design involves doubtful assumptions, but nevertheless tests on

models seem to confirm that these assumptions are reasonable. The saving in concrete and cost for all arch dams is well worthwhile but the supporting foundation strata must be above suspicion. Some notable failures have been attributed to inadequate treatment of the foundations. Cupola, dome, or double curvature arch dam. This type of dam is generally suitable for valleys with a chord:height ratio of under 3. It is economical in concrete and its strength, for this thickness, is like an eggshell. Calculations are complex but models for testing to destruction are used with success. Foundations must be above reproach. 28.9.6.3 Prestressed concrete dam Prestressed concrete dams have been developed and adopted in recent years for economy of concrete where good foundations are available. The thin concrete structure is anchored by steel cables embedded vertically in the concrete of the structure and with grout inserted in boreholes in hard strata below. Existing dams have been raised successfully and for this purpose the use of prestressed steel enables the existing structure to be little interfered with beyond drilling vertical boreholes for the prestressed cables to be inserted.42 28.9.6.4 Special problems concerning concrete dams32 Floods over dams. The precise estimate of floods is not so important as those for earth dams but nevertheless overtopping the crest should not be permitted not only because of the extra weight of water pressure on the dam but particularly the risk of scouring the strata under the toe, unless these factors are taken into account in the design of the dam. It is true that an overflow of 100 m over the Vaiont dam caused no damage to the dam but the abutments were against hard dolomite limestone. Rock testing by seismic methods. The velocity of sound through rock may give a valuable indication of its state below the surface, i.e. whether it is faulted, dry, wet, disturbed, open or revealing unsuspected faults or whether the density of concrete foundations is sufficient and the efficiency of grout curtains and contact grouting particularly for concrete dams which are on rock. Pore pressure and uplift. Pore pressure is dangerous under a concrete dam as the pressure is upward and 'lightens' the dam tending to turn it over and make it slide. In some cases pore pressure leads to leakage at the toe of the dam. In other cases it appears after a few years possibly from some kind of clogging and the pressure has to be reduced either by grouting or alternatively by drilling drains under the dam which, although it increases the leakage, reduces dangerous uplift. Deformation of dam and strata. Strain gauges embedded in the dam give a measure of any untoward trouble going on in a solid concrete dam. They consist of electrically stimulated vibrating wires in which any change in vibration speed indicates change of stress in the dam, indicating degree of movement. Other indicators of deformation are surface effects due to weather and temperature for which thermometers are inserted in the dam. Although movement of a dam can be ascertained by elaborate surveying equipment, pendulums and inverted pendulums inserted in boreholes inside the dam measure deformation more exactly. Pendulums for high dams are used especially for showing the movement when the reservoir is filled and empty and if these

values are the same or whether they change over the years. The relative movement of the strata with the dam can also be found. 28.9.7 Examples of raised lakes There are two or three instances of the utilization of natural lakes (other than Thirlmere and Haweswater which have been developed by high dams), the chief of which is Loch Katrine, for Glasgow, where the natural surface of the lake has been raised 4.5 m to draw off 1 m below the original natural lake level, the total supply available being about 320 million I/day. Similarly, the Crummock Lake for Workington has been raised 600 mm; the drawoff pipe is 2.4 m below this level. This is estimated to give a gross supply of 59 million I/day. The utilization of existing lakes raises special methods of tunnelling to draw water from lower existing levels as well as raising the level of water and at the same time coping with storm water. 28.9.8 Pumped storage reservoirs The largest examples of these reservoirs, constructed in Britain on clay and with clay cores and supported by any suitable material nearest the site, are those of the London metropolitan area. These reservoirs are of the order of 20 m in height and store water pumped from the Thames during periods of high flow.

28.10 Desalination43 Desalination should be regarded as a method of treatment to remove impurities, particularly salts, from a saline water. It has come into use at an increasing rate during the last 30 yr; and in certain circumstances can compete with orthodox sources which depend on conventional treatment of water from boreholes, impounding reservoirs and river waters. However, generally the cost of desalination is at least double the cost of fresh water sources, and frequently a factor of 10 greater. The variation in cost arises from: (1) The degree of salinity to be treated, e.g. sea water (chlorides 35 000 mg/1), brackish water (5000 down to 500 mg/1, which is tasteless to most palates). (2) The location and availability and cost of power, heat, transport. (3) The selection of plant, i.e. (a) multi-stage flash (MSF) and other variations of this distillation plant; (b) electrodialysis and the somewhat similar reverse osmosis plant; and (c) several other types used on ships and in factories and other special purposes. 28.10.1 Multi-stage flash distillation (MSF) (vacuum separation) If sea water is evaporated, steam is condensed as pure water leaving solid salt behind, as in the Dead Sea region. In the mechanical MSF process the sea water is pumped through a pipe (sufficiently long to ensure that sand is not drawn in during rough weather), heated, and passed into a tank under partial or reduced vacuum. As water boils at a lower temperature than normal when at a lower pressure (as on a mountain), fresh water 'flashes off' as steam which is cooled by incoming pipes conveying the sea water and condenses to fresh water. This process is repeated in several stages to increase the efficiency. Many problems arise, apart from the multiplicity of pipes, particularly the elimination of alkaline and calcium sulphate scale which can be controlled by the addition of polysulphide, acid, and lime to

increase the pH from 5 to 7. If temperatures could be used above 12O0C the cost could be reduced. 28.10.2 Electrodialysis (membrane-electrode separation) If brackish water is pumped through a tank between two membranes on each side of which is a positive electrode and a negative electrode, the electropositive sodium will go through one membrane to the negative electrode and the electronegative chlorine will go through the other membrane to the positive electrode. The water between the membranes, thus denuded of sodium chloride and other salts, is fresh. The method is only economic for water containing up to about 10 000 mg/1, and for reductions down to 500 mg/1, a cost of 20 to 5Op per 10001 is incurred (1983). Operational plants with outputs of up to 22 million I/day have been installed in the Middle East, and the method is now well established. 28.10.3 Reverse osmosis (membrane pressure separation) 'Osmosis' may be envisaged as a natural flow of fresh water into sea water when in contact with each other; whereas 'reverse osmosis' acts when pressure is applied to the brine which, when pushed through a special membrane such as cellulose acetate, causes the fresh water to flow out of the brine for separate use. Several plants are now in operation and act in the same range as electrodialysis plant. The main disadvantages are the high operating pressure and the fine limits involved in production of the membrane, but most of these problems have been overcome. A further disadvantage is the need for prefiltration and treatment to remove excess solids and biological impurities before the influent can be accepted through the membranes without early clogging. 28.10.4 Freezing Two freezing processes are being evaluated - vacuum and secondary refrigerant - but neither are yet proven in a full-scale operation.

high cost of trunk main laying has led to a reappraisal and has indicated that in many cases a greater reliable yield can be established by using a reservoir for regulation of the river flow and abstracting direct from the river in its lower reaches, e.g. Clywedog.46 Similarly, lowland pumped storage reservoirs (e.g. Grafham Water, Draycote and Empingham) are filled from low-quality river waters. Therefore, increasing use is made of lowland waters taken from the lower reaches of comparatively slow-moving and turbid rivers. They present far greater problems from the point of view of organic and industrial pollution and the treatment aspect becomes far more complex. Increasing sophistication in the equipment available for automatic control is leading to consideration of continuous monitoring of raw-water qualities for automatic control of the treatment process, and this now becomes established practice. Underground supplies from aquifers such as Limestone, Chalk, Sandstone and Greensand are normally biologically pure, but are often very hard and can contain objectionable levels of iron, manganese, sulphates and chlorides, as well as excess carbon dioxide and hydrogen sulphide. In some circumstances, river gravel can also be used as an underground source, although this is not necessarily of the same degree of organic or biological purity. Although there are large quantities of water in old mine workings, it tends to be very high in dissolved solids, particularly sulphates and chlorides. Boreholes near the coast can also suffer from an infiltration of salinity with resultant brackish water. 28.11.1 Water characteristics The main characteristics of a raw water which affect treatment processes are summarized in Table 28.10 follows:

Table 28.10 Chemical characteristics of raw water Group

Main constituents affecting treatment

Gases

Oxygen, carbon dioxide, hydrogen sulphide, ammonia Clays, minerals, siliceous matter, vegetable debris Organic acids, humus, peat, algae, faecal matter (a) Hardness salts: Permanent - calcium and magnesium sulphates, nitrates and chlorides Temporary - calcium and magnesium bicarbonates (b) Nonhardness salts: sodium sulphates, chlorides, nitrates or bicarbonates

Suspended solids Organic matter

28.11 Treatment of water for potable supply4445 The type of treatment required varies considerably according to the source of supply of the raw water, whether it be a surface water or from underground sources. Surface waters may be divided into upland and lowland sources. Except in the case of very small supplies, upland waters are usually impounded in the catchment area and are goodquality waters, low in dissolved solids and with little organic contamination from the biological point of view, although they are frequently high in organic colour due to deposits of peat and can also contain, particularly in the Pennine area, iron, manganese, and aluminium in solution. Although impounded water is generally of good quality it can be subject to disturbance due to stratification, thermal turnover if the water is deep, or by surface winds and flash runoff if the water is shallow. In these circumstances there is a marked and often very sudden deterioration in the quality of the water and, although this may be for only a short duration, it must be given full consideration when a treatment plant is being considered. The increasing demands being made on upland sources, the difficulty of finding suitable reservoir sites and the extremely

Dissolved solids

A range of possible treatments for different characteristics of the raw water is summarized in Table 28.11. The processes given in Table 28.11 are associated with the appropriate sedimentation and filtration plant and techniques. The full range of treatment for a poor-quality lowland river water may include storage, algal control, aeration, pH control, coagulation, precipitation softening, flocculation, sedimentation, filtration, chlorination, dechlorination, pH adjustment, and taste control. The range of chemicals commonly used in treatment processes is summarized in Table 28.12. The basic principles adopted in each stage of these treatment processes are briefly described in the following, with typical examples.

Table 28.11 Range of treatments for different characteristics

Table 28.12 Chemicals commonly used in water treatment

Characteristics

Possible treatment

Substance

Formula

Purpose

Gases Dissolved impurities Suspended matter Colloidal matter Colour

Aeration Precipitation - (oxidation) Coagulation, settlement Coagulation Coagulation, activated carbon, ozone, chlorine Aeration, activated carbon Activated carbon, chlorine, chlorine dioxide, ozone Aeration, control by alkali

Activated carbon

C

Taste and odour control

Aluminium sulphate (alum) Ammonia

Al2(SOJ3 NH3

Ammonium sulphate

(NHJ2SO4

Coagulant With chlorine for sterilization A source of ammonia for chloramine

Calcium carbonate (chalk)

CaCO3

A source of bicarbonate alkalinity

Calcium hydroxide (slaked lime)

Ca(OH)2

Softening and pH control

Ca(OCl)Cl

Disinfection

CaO

Softening and pH control Disinfection Disinfection, taste and odour removal

Odour Taste Acidity /free carbon dioxide Hardness Iron and manganese Other metals Salinity (brackish) Oil Algae

Biological impurities Industrial pollution

Lime and/or soda precipitation, or ion exchange methods Aeration, precipitation and filtration with iron-removal media Coagulation and precipitation Distillation, demineralization, reverse osmosis, freezing Flotation, coagulation Straining, copper sulphate, chlorine, cuprichloramine Storage, chlorine, chloramine, ozone, ultraviolet light Combination of above as required

28.11.2 Storage For a scheme using direct river abstraction a raw-water storage of 7 days is recommended to allow for settlement of heavy silt load to even out any rapid changes in water quality and to allow for rejection of water containing accidental and heavy pollution (in, for example, Oxford47 and Nottingham48). 28.11.3 Algae The growth of severe blooms of algae which would interfere with the treatment process can be inhibited by the use of an algicide (e.g. copper sulphate) and the exclusion of direct sunlight. 28.11.4 Aeration The level of dissolved gases can be reduced substantially by an aeration system which in the order of ascending efficiency takes the form of cascades, sprays, and induced draft towers. If the water is particularly 'flat' in appearance the level of oxygen can be increased and the appearance of the water enhanced by similar means (in, for example, Ardleigh49 and Oxford). 28.11.5 Coagulation Any raw water containing colour or finely divided suspended solids needs the addition of a coagulant to neutralize the electrical charges causing dispersion and induce the impurities to coalesce and flocculate. This process is often assisted by slowspeed agitation to increase the collision between the particles. Normally, the reagents are delivered by road vehicle and taken into bulk storage at the treatment works. The method of adding the coagulant most usually adopted is the use of positive displacement ram-type metering pumps which can be controlled easily to vary the dose according to the treatment flow and, if required, to the water quality (as, for example, in Colchester and Swansea).

Calcium hypochlorite (bleach powder) Calcium oxide (quick or burnt lime) Chlorine Chlorine dioxide Copper sulphate (bluestone) Ferrous sulphate (copperas) Ferric chloride Ozone Potassium permanganate Sodium aluminate Sodium carbonate (soda ash) Sodium chloride (common salt) Sodium hypochlorite (Chloros or Voxsan) Sulphur dioxide

Cl2 ClO2

CuS044H20 Algal control FeSO4TH2O Coagulant Coagulant FeCl3 Disinfection, taste, O3 odour and colour removal KMnO4 Na2Al2O4

Removal of iron, manganese, algal control Coagulant

Na2CO3

Removal of permanent hardness and pH control

NaCl

Regeneration of zeolites

NaOCl SO2

Disinfection Dechlorination

28.11.6 pH control For efficient coagulation the pH value is critical and as the pH of the water with the added coagulant is unlikely to be at the required level it is necessary to correct this by the addition of acid or alkali, preferably under automatic control. Variation of pH from the level necessary for optimum coagulation can produce light fluffy and fragile floes and high residual dissolved coagulant (as, for example, in Bradford and Londonderry50). 28.11.7 Precipitation In any treatment process whch includes coagulation and sedimentation it is possible to precipitate the hardness salts by the addition of lime and/or soda and the precipitated salts are effectively removed in the general system and can in some circumstances increase the efficiency of the treatment although

inevitably producing an increase in the volume of sludge to be discharged from the works (in, for example, northeast Lincolnshire51 and Sheffield52). 28.11.8 Mixing As the volume of reagent is small compared with the volume of water being treated it is critical to ensure that the chemical is fully dispersed into the body of the water and also that the reagents are added in the correct sequence according to the chemical requirements. Reagents can be diluted to ease the mixing problem, this is really two-stage mixing, and are then added in an area of turbulence induced either hydraulically or mechanically: hydraulically in the nappe of a weir or the standing wave of a Venturi flume or mechanically by high-speed mixing and sometimes by a pumped recirculating system (as, for example, in Bristol53). 28.11.9 Flocculation After the reagents have been added and fully mixed it is normal to induce flocculation by passing the water through an area of slow agitation which, again, can be induced either hydraulically or mechanically (as, for example, in north Derbyshire). 28.11.10 Sedimentation In its simplest form, sedimentation is the use of tanks giving a retention time that is long enough for the floe particles to settle and compact into sludge on the bottom of the tank, from which point the solids are discharged for disposal and the settled water is decanted to the following filters. However carefully such tanks are designed, the physical retention seldom exceeds 40% of the nominal retention and this has led to the development of the upflow type of treatment unit. After a flocculation zone the water is induced to flow vertically upwards through an area of suspended sludge where the floe particles have a large area of contact which greatly assists in the formation of denser agglomerates. Such tanks are designed so that the sludge can be withdrawn at the rate at which it is forming and provide a stable process that can be controlled over a varying range of duties. Construction can be in either steel or concrete, with the units either square or circular in plan, and as the capacities of treatment works increase it is generally more economical to consider a smaller number of circular-type tanks (as, for example, in Bradford in the Derwent Valley). 28.11.11 Filtration After coagulation and sedimentation, the water still retains a quantity of suspended matter which is removed by filtration. It should be noted that the filtration process associated with the treatments being described is that known as 'rapid' as distinct from slow-sand filtration which is a biological process completely in itself. The settled water passes through a layer of comparatively fine and specially graded sand supported on underbeds of graded pebbles with a piped header and lateral under-drain system, or supported on a flat floor with a system of closely spaced nozzles. In either design the clean water is collected from the base of the filter and as the resistance to flow increases, in proportion to the quantity of intercepted matter building up, the filter bed is cleaned, first by expanding the compacted bed, usually by the application of an air scour, which effectively loosens the intercepted impurities which are then flushed out to waste by a reverse flow of water. The filter bed can be contained equally well in a steel pressure vessel or an open-topped gravity tank and the siting of the plant

relative to the hydraulic gradient can determine which method is preferable. Where large flows are being considered the gravitytype filter does not have the same restriction on the size of individual units and it is unusual to follow sedimentation, requiring open-topped type tankwork, by pressure filters (as, for example, in the Lune Valley, West Glamorgan53). Filtration technique is currently going through a period of change with advocates for downward, upward and sideways flow, for deep beds and shallow beds, for single media, multimedia, and multi-layer media. All these variations have some application, however limited, and provide filtration in the depth of the bed rather than on the surface, with a greatly increased efficiency. Although there is very little, if any, long-term operating experience on some of the designs, the use of a two-layer downflow filter with a top stratum of graded anthracite resting on a shallower layer of conventional sand is a system that has been in use for a number of years in the UK and in gaining support as it has been proved that filter ratings can be increased and the length of filter runs extended. 28.11.12 Backwashing Air for expanding and scouring the filter bed is normally provided by electrically driven blowers delivering large volumes of air at relatively low pressure direct to the induction system which is usually the underdrain system used for collecting the filtrate. Wash water is most often provided by direct pumping to the same common induction system (as, for example, in Wakefield54). 28.11.13 Chlorination After the water has been clarified satisfactorily it is still necessary, if it is to become potable, for it to be fully disinfected, and chlorine is the most usual reagent for this duty. It is normally delivered as a liquid under pressure in either cylinders or drums, depending on the quantity required, and in some of the largest installations it is being delivered by bulk tanker and transferred into storage vessels at the treatment plant. As it is considered prudent to carry a chlorine residual into the reticulation system as a measure of safety this has to be controlled at a level low enough to be unobjectionable to the consumer. Current practice is often to dose above the chlorine demand of the water and to control the residual passing into supply by adding sulphur dioxide to neutralize any excess. Sulphur dioxide is handled as a liquid under pressure in exactly the same way as chlorine and the dosing is normally under automatic control to ensure that the final chlorine residual is maintained at the correct level (as, for example, in West Surrey). 28.11.14 pH adjustment Depending on the treatment adopted, the final water is unlikely to be at the pH required for distribution purposes and will need correction by the addition of acid or alkali. At the same time it is important to correct any undue corrosive tendencies which may be inherent in the treated water (see, for example, descriptions of works at Bedford55 and in north Devon56). 28.11.15 Taste control Taste which is objectionable as far as the consumer is concerned can be present in the raw water or can develop during the treatment process and activated carbon is often used to absorb the elements that are responsible. It can be added as a powder before the sedimentation process and removed with the sludge, directly as a powder or granule on to the filter beds and removed

with the washwater, or as a granular filter medium in a separate filtration stage added to the end of the clarification treatment. In the first two applications the carbon is not recovered but if it is used as an additional filtration unit it can either be regenerated on site or returned to the manufacturer for this purpose (as, for example, in east Surrey57 and Oxford47). 28.11.16 Waste products Whatever the process used for clarification, or precipitation softening, the impurities removed are concentrated in the form of a sludge which under the best operating conditions is unlikely to be less than 95% water. In this form it can be fed to a filter press, or possibly a centrifuge for a softening sludge, to produce a dry solid suitable for mechanical handling and disposal. The filtrate, or centrate, has to be disposed of as a liquor. In a few cases, attention is being given to the possibility of treating the sludge with acid to recover the coagulant but this process is not yet proven as commercially viable (as, for example, in the Fylde, mid Northamptonshire). Sludge disposal in waterworks is not such a problem as sewage sludge. Local conditions can normally cope with the smaller quantities of waterworks sludge by distributing it on land, quarries, pits, river or sea which may be, and generally are, available. Transport of sludge should be in closed containers in hilly districts, otherwise there is loss of sludge from well-filled opentop vehicles!

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14

15 16 17

Institution of Water Engineers and Scientists (1979) The structure and management of the British water industry.' Water practice manuals, Book 1, p. 19. Phillips, J. H. (1983) 'Water usage and the quantification of unaccounted water in a universally metered supply area.' J. Instn W. Engrg, 37, 4. Sharp, R. G. (1967) 'Estimation of future demand on water resources in Britain.' /. Instn W. Engrg 21, 232. Ministry of Agriculture, Fisheries and Food (1967) Watercress growing. Ministry of Agriculture Bulletin No. 136. HMSO. Klein, R. L. (1959) The use of asbestos-cement pressure pipes in water supply practice.' W. & W. Engrg, 63, 356. National Water Council (1980) Leakage control policy and practice. NWC. Hydraulics Research Station, Hydraulic Design Charts Nos 1, 2 and 4. Hayton, J. G. (1964) The use of polythene sleeving as a form of protection to spun iron water mains against external corrosion.' J. Instn W. Engrg, 18, 465. Young, O. C. and Smith, J. H. (1970) Simplified tables of external loads on buried pipelines. Building Research Establishment, Garston. Hydraulics Research Station (1983) Charts for the design channels and pipes. HRS. Lea, F. C. (1938) Hydraulics. Edward Arnold. Lapworth. C. F. (1944) Surge control in pipelines. /. Instn W. Engrg, 49, 29. Lupton, H. R. (1953) 'Graphical analysis of pressure surge in pumping systems.' /. Instn W. Engrg, 7, 87. Thorley, A. R. D. and Enever, K. J. (1979) Control and suppression of pressure surges in pipes and tunnels. Construction Industry Research and Information Association Report No. 84, CIRIA, London. Crump, E. S. (1952) 'A new method of gauging flow.' Proc. Instn. Civ. Engrs, 1, 749. Water Resources Board (1970) Crump weir design. WRB Technical Note No. 8, British Geological Survey (n.d.) Geological survey maps and memoirs. British Geological Survey, London.

18 Walters, R. C. S. (1929) 'Hydro-geology of Chalk.' J. Instn. W. Engrg, 34, 79. 19 Walters, R. C. S. (1936) 'Oolites.' /. Instn. W. Engrg, 41, 134. 20 Ineson, J. (1970) 'Development of groundwater resources in England and Wales.' J. Instn. W. Engrg, 24, 155. 21 Rofe, B. H., Durrant, P. S. and Egerton, R. H. L. (1977) 'Some aspects of the use and management of groundwater resources.' J. Instn. W. Engrs and Scientists, 31. 22 Penman, H. L. (1954) Evaporation over the British Isles. Institution of Water Engineering, London. 23 Armstrong, R. B. and Clarke, K. F. (1972) 'Water resource planning in SE England.' J. Instn W. Engrg, 26, 11. 24 Lapworth, C. F. (1949) 'Reservoir storage and yield.' /. Instn W. Engrg, 3, 269. 25 Mansell-Moulin, M. (1966) Flow frequency curves for design of catchwaters. Institution of Water Engineering, London. 26 National Environmental Research Council (1975) Flood studies report. 27 Institution of Civil Engineers (1978) Floods and reservoir safety - an engineering guide. Thomas Telford, London. 28 Terzaghi, K. and Peck, R. B. (1967) Soil mechanics in engineering practice. New York. 29 Bishop, W. A. (1955) The use of the slipcircle in stability analysis of slopes.' Geotechnique, 5, 7. 30 Mitchell, P. B. (1951) 'Reservoir site investigation and economics.' J. Instn W. Engrg, 5, 445. 31 Walters, R. C. S. (1973) Dam geology. (Appendices by J. L. Knill). Butterworths, London. 32 Kennard, M. F. and Eden, W. H. (1978) Bewl Bridge Reservoir: design and construction. Thomas Telford, London. 33 Rofe, B. H. and Tye, P. F. (1971) 'Application of instrumentation to earth dams.' J. Instn W. Engrg, 25, 157. 34 Hallas, P. S. and Titford, A. R. (1971) 'Design and construction of Bough Beech Reservoir.' J. Instn W. Engrg, 25, 293. 35 Picken, J. A. (1957) The Chew Stoke Reservoir scheme.' J. Instn W. Engrg, 11, 33. 36 Walters, R. C. S. and Walton, R. J. C. (1957) 'Water supply for the Yeovil District (Sutton Bingham Scheme).' Proc. Instn Civ. Engrs, 8, 71. 37 Kennard, J. and Kennard, M. F. (1962) 'Selset Reservoir.' Proc. Instn Civ. Engrs, 21, 277. 38 Kennard, M. F. (1963) 'Balderhead Reservoir.' Civ. Engng. & Publ. Wks Rev., 58, 633. 39 Crann, H. H. (1968) 'Design and construction of Llyn Celyn.' /. Instn W. Engrg, 22, 13. 40 Williams, H. and Stothard, J. N. (1967) 'Rock excavation and specification trials for the Lancashire-Yorkshire motorway - Yorkshire (West Riding) section.' Proc. Instn Civ. Engrs, 37, 607 and Discussion, 38, 135. 41 Farrar, R. E. S. (1972) 'Meldon Reservoir.' Civ. Engng & Publ. Wks Rev., 67, 895. 42 Water and Water Engineering (1961) The heightening of Argal Dam for the Falmouth Corporation Water Undertaking.' W. & W. Engrg, 65, 537. 43 Silver, R. S. (1967) Desalination. HMSO, London. 44 Skeat, W. O. (ed.) (1969) Manual of British water engineering practice, 4th edn, vol. 3, Institution of Water Engineers. 45 Holden, W. S. (ed.) (1970) Water treatment and examination. Churchill. 46 Fordham, A. E., Cochrane, N. J., Kretschmer, J. M. and Baxter, R. S. (1970) The Clywedog Reservoir project.' /. Instn W. Engrg, 24, 17. 47 Cartwright, F. (1964) 'Design of Farmoor Treatment Works, Oxford Corporation Water Department.' /. Instn W. Engrg, 18, 381. 48 Adams, R. W., Robinson, R. D. and Kennett, C. A. (1973) The River Derwent Scheme of the Nottingham Corporation.' J. Instn W. Engrg, 27, 15. 49 Water and Water Engineering (1972) The Ardleigh Reservoir Scheme in North-East Essex.' W. & W. Engrg, 76, 3. 50 Wilcock, E. J. and Sard, B. A. (1964) 'Design and operation of the Carmoney Water Treatment Works: Faughan River scheme - Londonderry RDC.' J. Instn. W. Engrg, 18, 477. 51 Ashe, R. V. (1966) The Great Eau scheme: North-east Lincolnshire Water Board.' /. Instn. W. Engrg, 20, 435. 52 Earnshaw, F. (1962) 'Design of the Yorkshire Derwent head works.' J. Instn. W. Engrg, 16, 139.

53 54 55 56 57

Water and Water Engineering (1955) 'The Usk Reservoir Scheme of the Swansea Corporation.' W. & W. Engrg, 59, 377. Collins, P. G. M. and Gibb, O. (1964) 'Design and construction of the Fixby Treatment Works of the Wakefield and District Water Board.' J. Instn. W. Engrg, 18, 491. Water and Water Engineering (1959) 'New water treatment works of the Borough of Bedford Water Undertaking.' W. & W. Engrg, 63,61. Water and Water Engineering (1972) The Meldon Reservoir scheme of the North Devon Water Board.' W. & W. Engrg, 76, 353. Shinner, J. S. and Davison, A. S. (1971) The development of Bough Beech as a source of supply (The East Surrey Water Company).' /. Instn. W. Engrg, 25, 243.

Bibliography General Water Supply The journals of the Institution of Water Engineers and Scientists (IWES). The Proceedings of the Institution of Civil Engineers (ICE). Institution of Water Engineering (1969) Manual of British water engineering practice. IWE, London. Twort, A. C., Hoather, R. C. and Law, F. M. (1974) Water supply. Arnold, London. Institution of Water Engineers and Scientists (1983) Water practice manual, Book 3: 'Water supply and sanitation in developing countries.' IWES, London.