Drainage and Irrigation Department - Manual

PREAMBLE

This i s a r e p r i n t of t he chap te r on I r r i g a t i o n from

t h e D.I.D. Manual 1973. The purpose of t h i s r e p r i n t i s t o g ive

some background knowledge and information on i r r i g a t i o n t o

newly r e c r u i t e d s t a f f . Some of t h e m a t e r i a l i n t h i s r e p r i n t needs

t o be updated and t h e reader i s advised t o consu l t t he Planning

and Design Branches, D.I.D. Headquarters, o r r e f e r t o t h e books

l i s t e d i n the Bibliography f o r t h e l a t e s t s tandard p r a c t i c e s

i n the f i e l d of I r r i g a t i o n .

Pusat La t ihan Kakitangan, Jaba tan P a r i t dan T a l i A i r , Ampang.

February, 1981.

C O N T E N T S

Pass <'HAPIER 111 -- IRRIGATION

SH TION I O B J ~ ~ TS AND G~NERAI REQUIREMENTS o f . IRRIGATION ... 109 Irr~gatlon Requ~rements for R~ce Cultivation .. 111 General Descrlptlon of an Irrigation System . 111

Figure\ I 1 111 Irrigation System, General ... 112

SEC'TION I 1 - APPI.I( ATION o b WATER TO T H ~ LAND ... 114 Application of Water to Rice Land . . . . . .. 114 Appl~cation of Water to the Land for Crop other than Rice . . . . . . 123

Figures Il111'11I

Figures 1 IIV '111 21IV 'I11

3alIVlIII 3blIV '111 411V'III 5IIVlIII

lrr~gat~on System. Narrow Valley 9.

Distribution by Batas Irrigation System. Flat Land Irrigation System. Tan jong Karang . . . Irrigation System. Kubang Pasu Section Irrigation System. Sungai Manik and Trans lrrigation System. Pahang Paya . Irrigation System. Crops other than Rice lrrlgation System. Proposed Layout S o u ~ c c s OF I R R I G A T I ~ N WATER Direct Rainfall . . . . . . . . Surface Water Ground Water

... 116

... 117

... 117

.. I19

.. 119 , 120

. 122 Perak ... 122

1 24 126

. .. 126 . . 127 . . 127 . 127 .. 130

Trans Perak River lrrigatlon Scheme Intake at Lambor Kiri . . . . . . ... 131

14th Mile Terachl Irrigation Scheme General Plan of Headworks

Automatic Bear Trap Gate .. Bukit Merah Reservoir

9. . . . Details of Pedu Dam and Muda Dam . . . Pumping Station. Pinang Tunggal Pumping Station. Sungai Nerus Pumping Station, Pekula Control Gate. Paya Sg. Leng . DISTRIRI'TI~N Canals Canal Controls . . . .

Dimensions of Field Channels . ... Percolation Gradient .. . . . Canal With Berm . . . . . ... Canal Without Berm ... Overshot Control ... . . . ... Undershot Control . . . ... ...

Page 6,'IV;III Arrangement of Canal Control ... ... ... . 151

... 7'IV'llI Bifurcation Canal . . . . ... ... ... 152 SECTI( )N V ... DLSIGN PHINCI R.I..S I:OH W I-. IRS ( IN(-IAJIXNG BARRAGIS)

Weir Crest Level and Length .. ... . 155 ... ... The We~r Structure ... ... 157

... Surface Flow and Energy Dissipation ... ... 157 Percolation Pressures. Uplift and Exit Gradients ... 170

... ... Sediment Exclusion and Ejection ... 183 Applicability of Principles to Other Types of

... Hydraulic Strwture ... . . . 183

Bed Sediment Move Curve ... ' - 7

Hydraulic Jump . . . ... . . . ... ... ... . . . . . . Longitudinal Section o f Weir ... ...

... Diagram Connecting HL . & Ef2 ... ... Curves for Determining Ef2 ... ... ... ... Curves for Predicting Depth of Trough . . Energy of Flow Curves . ... ... ...

... Stage Discharge Curve for Weirs ... Diagram for Designing Downstream Apron and Wing W a p

... Rehbock Sill . . . ... ... Floor Block . . . . . . . . ...

... Downstream Wing Wall Splays . . . ... ... Downstream "Onion" . . .

Wing Walls ... Proportions of Inlet Flares and Upstream Floor Flow Nets . Weir with Length but no Depth . Flow Nets . Pile Line with Depth only . . ...

... Flow, Nets Weir with Floor and llpstreani Pile Line Flow Nets . Weir with Floor and Downstream Pile Line FIow Net . Weir with Floor and Pile Line at Each bnd

... Percolation Pressure Curve ... . . . . Stream Tube ...

... Exit Gradient Curve . . . ... . . . ... Percnlat~on Pressure Diagram No Surface Flow ... Percolation Pressure Diagram ' with Surface Flow ... Clay Lens Below Structure ' . . . . . ... Cause of Collapse of Rubble Weirs ... Bernam River Headworks Silt Exculder ... . .

IRRIGATION

SECTION I OBJFC T\ 4 N D GENERA1 REQUIREMENTS OF IRRIGATION

1. Irrigation 1s defined as the artificial application of water to soil for the ?urpose of supplying moisture essential to plant growth.

2. Rice is the only crop which is migated on a large scale In Malaya at present. On a small scale. vegetables are irrigated, generally by means of watering cans. Also on a small scale, maize has been irrigated at times. There is no doubt that larger scale irrigation of cr'ops other than rice would be beneficial in Malaya. ~ n d in due course such irrigation will almost certainly come.

3. Tradltlonally in Malaya, rice is sown In nurseries, and the seedlings are planted out into flooded rice fields when they are about 6 weeks old. Before planting out, the fields are prepared by successive ploughings and puddling. The rice grows In standing water. preferably about 3 inches to 6 inches in depth. till the gram is ripenmg. A1 this stage the water is drained off, and harvesting is done in the dry. 'Thus it is necessary to irrigate from one to two months before planting t a t until shortly before the harvest.

4. While it is evident that rice is tolerant to standing water and waterlogged sod. it is not proved that these conditions ate best for the growth of the plant. Standing water is. however. the best method of weed control. and until some other ratisfactory control method IS found, rice will continue to be grown in standing water.

5. Crops other than rice are not tolerant to standing water and waterlogged 9011, and will suffer or die ~f there is too much water in the soil. They will suffer also if not enough water in the soil is available to the plant. Hygroscopic water IS not available to the plant. Capillary water is available. Water in excess of what can be held by capillary action is harmful to the plant. If the soil is well drained, excess water will quickly drain away. But if it cannot drhin away, the plant will suffer. The aim of irrigation for crops other than rice is. by repeated applications of small quantities of water. to keep the moisture content of the soil within the capillary range. Care is taken to avoid over-irrigation. as this may raise the water table unduly and thus prevent free drainage of the soil. Also, over-irrigation is uneconomical of water.

6. In the process of growth, a plant uses watet by transpiration. That is to say, it draws water from the soil and releases it to the atmosphere by evaporation from its leaves. The rate of transpiration depends on the crop. the density of planting, the state of growth, the temperature, the humidity and the wind. Not very much is known about transpiration of rice. but i t should be imderstood that he amount of water c~mumed in this way is not negligible by any means. To emphasise this point, it i s noted here that at the period of maximum rate of transpiration, some crops transpire in a month a quantity of watet equivalent to a depth over the uholr cultivated area exceeding 6 inches. During the period of growth, the rate of transpiration rises from a minimum. which may be very small, to a maximum, and then falls back to a minimum. In due course, more will be learned about transp~ration from rice, as a result of work to be done at the new soils-moisture research station of the Agricultural Department at Tanjong Karang.

7. The water transpired is used by the plant. Irrigation water is also by evaporation from the ground or free water-surface, deep percolation, lateral percolation, surface losses, transpiration by weeds.

For rice, evaporation losses from the free water surface are unavoidable. For other crops, evaporation can be reduced by sutface treatments, such as mulching. Rate of evaporation depends on temperature, wind and humidity, and is variable. Deep percoiation and latetal percolation losses for rice depend on the nature of the soil and the thoroughness of preparation before planting. Normally they are small, as rice is planted usually only on the heavy and tather impermeable soils. For other crops, percolation losses are avoided by avoiding over-irrigation. Surface losses are avoided by good irrigation practice. And transpiration losses from weeds are avoided by weeding.

8. The total amount of water required in any given period is the total of the water used and the losses. Evidently it varies from month to month, and the amount of irrigation water to be supplied varies still further, conversely as the rainfall; for the more rain there is, the less irrigation water is required.

9. Maximum demand for crops other than rice must obviously depend on the crops gown; but over a whole month, the maximum demand of an area planted to crops other than rice is likely to be less than the maximum demand for an equal rice area.

10. Assuming that duting periods of maximum demand, irrigation proceeds at a steady average rate within the area, the capacity of the main supply line to the area will be fixed normally by the maximum demand. (In rare cases where it is requited to store storm water by deep flooding of the agricultural land itself, the main line may be larger than is necessary for maximum demand.) The sum of the capacities of the waterways which make up the distribution system will usually be substantially in excess of the maximum demand, however, for-as will be explained later-irrigation may not proceed at an even pace throughout the area.

11. Where the source of supply is a river without benefit of appreciable impounding, maximum demand must not exceed the normal dry weather flow of the river in the months when maximum demand can be expected. If maximum demand for the proposed scheme is greater than rivet flow, then obviously the scheme must be reduced in size. It is during the driest weather that irrigation is most required. and an irrigation scheme is useless if it fails at the crucial moment. When planning a scheme, it is necessary to study carefully the discharge hydrograph of the river from which water is to be drawn, for as long a period as possible, and to make sure that the demands of the scheme will not be excessive. Whete data for drawing the hydrograph are available for only a small number of years, these data must be considered in conjunction with longer term rainfall data. An intelligent estimate may then be made of low flow during dry years.

12. For pumping schemes, maximum demand will determine the capacity of the pumping units. It is usual nowadays to instal three identical units, and-in determining their capacity-it is assumed that they will all be serviceable at times of maximum demand, due allowance being made of course for the number of hours pet day during which there can be pumping. For example, electrically operated stations use off-peak power at a special rate, and can operate normally only 18 hours per day.

13. For a reservoir scheme, maximum demand does not fix the capacity of the reservoir. Evidently this must be fixed after oonsideration of total annual demand and losses, the hydrograph of the river supplying the reservoir and so on.

14. The term "duty" is often used to indicate the number of acres served by 1 cusec of imgation water, or alternatively the number of acres which 1 c u m should serve. Without qualiilcation, the term is rather nebulous, and it is necessary to state clearly what is meant in any given case. For example, one might state that the tequired duty for rice cultivation in a certain area during a certain month was so many acres per cusec.

15. While "duty" IS quite a useful conception in considering area irrigated In relation to size of stream. it is often more convenient to consider water requite- ments in terms of equivalent depth of water to be applied to the field in a certain period. For example. it might be stated that rice in a cettaid area required a total of 19, inches of water during a certain month of the year. of which rainfall could be relied on t c b -upply 3 inches. leaving 9 inches to be supplied by the irrigation s!stem. I t 1s convenient when thinking in these terms. to be able to convert flow is) depth per month. and it is pointed out that 1 cusec flowing for I 2 hours will give almn*t exactly. a depth of 1 ft. of water on I acre. That is to -a!. 1 acre foot of water. Thus I cusec flowing for a month of 30 days will give 60 acre feet l \ f water. So it is seen that a "duty" of 60 acres per cusec corresponds to an equivalent depth of water in one month of 12 inches. Similarly 15 inches depth per month corresponds to a duty 48 acres per cusec.

IRRICIATION REQUIREMFN'I\ FOR RlCF < 1 1-TIVATION

16. The Drainage and Irrigation Department carried out studies to assess the water requlremenb !'or rice cultivation in Malaya during 1960 1961 in two schemes. bj measuring the inflow. rainfa!!. and the outflow from padi fields throughout the growing period.

17. These studies were extended to 6 schemes and results computed by an F A 0 team comprising of Professor G. A. W. Van de Goot and G. Zijlstra during 1961 - 1963 and a report prepared for the Government in 1963. F A 0 teport Vo. 1671.

18. The recommended water duties for rice cultivation for single and double cropping are tabled below: --

I . Double Cropping System Off-season

~ncheslmonths Acres[Cusec L.!Seclha a. Presaturation period 40 days . . . 15 48 1.5 b. Normal itrigation period ... 10 71 1 .O

Main-season a. Presaturation period 40 days ... 13 55 1.3 b. Normal irrigation period .. 9 79 0.9

2. Single Cropping S.wtem a. Presaturation . . ... ... 12 60 1.2 b. Normal irrigation period ... 9 79 0.9

GENERAL DESCRIPTION OF .4N IRRIGATION SYSTEM

19. The objects and general requirements of itrigation have been considered In the foregoing. The object of the irrigation system is to meet these requirements. The system and its parts are considered in the following.

20. First let us look at an irrigation scheme as a whole and name its principal parts. Fig. 1 11111 shows a hypothetical scheme in which the area to be irrigated is shown hatched. The souice of supply (in this case) is a river. Water from the river is backed up by a weir (or by a gate-controlled weir which is called a "barrage") and diverted through a head regulator into the main canal ~f the scheme. The weir or barrage. together with the head regulator. is called the headworks.

21. Water is conveyed to the area to be irrigated by means of the main canal, which may divide into branch canals, each branch feeding a different part of the area. The branch canals may supply feeder canals, each of which supplies a number of distributary canals, or it may supply directly to the distributaries. The distributaries supply water to the field channels, and it is normally from the

FIG. I/I/DI

field channels that water is supplied to the surface of the land. There is not always a clear distinction between the various classes of canal. and some may be omitted altogether. In a very small scheme there may be one canal only, which serves all purposes. Usually it is inadvisable to pass water to the field except from a field channel. as any other practice results in maintenance difficulties and loss of water control.

22. Broadly speaking, canals can be divided into two classes: those concerned primarily with conveying water to the areas to be irrigated, and those concerned with distributing it over the areas to be imgated. The former do not need to flow above ground level, whereas obviously the latter must "command" the land to be irrigated, or it will be impossible for the water to pass from the canal to the land. Distributaries and field channels fall into the "distribution" category. The main, branch and feeder canals falls into the "conveyance" category .

23. At the headworks, water in the river is backed up to a level which is lower than bank level. Consequently, the hydraulic gradient of the conveyance canals must be flatter than the slope of the land, in order that command may be obtained at the head of the area to be irrigated. This, in fact, is usually the reason for the substantial length of the main or branch canals.

24. At all canal divisions and at the heads of subsidiary canals, there must be water controls of some kind. Canal controls may be provided also to fiatten steep hydraulic gradients which would result in destruction of canals on account of excessive velocities of flow, and to ensure command at critical points. Canal design and control will be discussed in much more detail later.

25. In many cases it is necessary to provide a complementary drainage system, this being shown in Fig. l/I/III in broken lines. More will be said about this later.

26. In the hypothetical example considered above, water is diverted from the river by means of a weir or barrage. In cases where the area to be supplied is small in relation to the fiver, it will be uaaconomical to build such a structure. however, and it may be better to raise water to the head of the main canal by pumping.

27. Where the total annual or seasonal flow of a river is sufacient fot irrigation of a required area, but the low flow of the river is insufficient to m a t the maximum demand (which will often occur when the river flow is lowest), then the only solution is to impound the flow. For efficient irrigation the water should be ~mpounded in the river valley above the area to this is impossible, and then some attempt is on the land itself. The Pahang Paya systems, systems which allow flush irrigation fall into valley impounding scheme in Malaya at present is the Krian Irrigation Scheme.

SECTION 11 -APPLICATION OF WATER TO THE LAND

1. As stated earlier. rlce in Malaya is generally grown in standing water, whereas other irrigated crops will not toletate standing water or even waterlogged ground. Clearly then there are methods of application of watet to rice land which we not acceptable to other crops. Therefore application of water to rice and to other crops will be considered separately: to rice first and to other crops afterwards.

APPLICAT!ON OF 'A'ATEK TO RICE LAND

2. For irrigat~on of rice, there ate two general systems; continuous dnd rotational. The traditional system is the continuous one. In this system, water is dribbled on to all parts of the area continuously and simultaneously. Theoretically. a thousand tiny suppiy trickles are simultaneously regulated to supply enough but not too much water to a thousand little fields, or "petak" as the small rice fields are called. As might be expected, this system does not work very well unless the land is very even and of very small slope, so that there can be an automatic evenning up of conditions when irregularities such as concentrations of water and local shortages tend to develop. Where the land is irregular or of modetate fall, difficulties quickly develop, for it is not easy to control simultaneously a thousand trickles.

3. The rotational system is much better and is slowly gaining favour. With this system, the area is divided into a number of parts, and each one is irrigated in turn, supply during each turn of imgation being at a higher rate of application than for continuous irrigation. A three part rotation has been found quite satisfactory on the East Coast, water being supplied to each part once each week. Conttol of water in the field is much easier with the rotational method because

(a) the overseers controlling the water do not have to be everywhere at the same time,

( b ) with larger streams of water, it is much easier to see how the spread of watef over the land is developing,

( c ) there is mcre chance of getting farmer co-operation if each farmer knows when his land is going to be supplied, and

( d ) when water supply is deficient, water can be forced on to high spots more easily when such watet as is available is concentrated into recognisable streams.

With better control, losses are lower and better use is made of the water.

4. There is, however, another very good reason for adopting the rotational method in areas where the soil is relatively permeable. as on the East Coast at the beginning of the season, before the land has been ploughed and puddled. Much of the East Coast rice land is underlain by gravel and sand beds. During the dry season, the surface clay, on which the rice is grown, cracks into a maze of smll but deep cracks which allow surface water to percolate downwards into the permeable beds below, in which the water table is many feet below ground surface. If, on such land, the continuous system of irrigation is attempted,-that is to say, if the available water is simultaneously sptead over a very wide front-the water will rapidly seep away downwards through the cracks. The land more remote from the field channels will not even be wetted until the farmets near the channels have ploughed and puddled their land, thus stopping the downward loss of water. Clearly, under these conditions, the farmers mote remote from the channels have to await the pleasure of those near the channels. On the other hand, if the tota- tional system is adopted, water is applied to sections of the area in turn, with a high rate of application. It thus sweeps over the land faster than it can leak away through the cracks, and the whole or a much greater part of the area can be wetted and thus rendered fit for ploughing and puddling.

5. The more common methods of application of water to rice land, in-

Afferent and good. as they are actually found in Malaya, are briefly described w~th reference to figures. in the following paragraphs. Every imgation scheme must have arrangements for removal of excess water of course, and typical drainage arrangements for rtce cultivat~on are shown also in the figures. Dtainage will 3e requlred to remove storm water and excess irrigation water. and to drain off at the end of the season o r possibly for some agricultural purpese during the season.

h t lg I 11 111 ~Ilustrates part of a tradittonal small narrow valley scheme. t , t util~.h I r e r e arc. hu:idrcd% dotled round the country. mainly in Malacca. Negr~ Sembtlan and Perak. A single canal. supplied from a headworks. follows tne hillfooi. and water ih tapped off to the bendang at a number of points as shown in the skeich. usuaiiy by means of uncontrolled pipes placed in the canal bank 'The ii.::dench ,I: i ::. {later 1 h u 4 tapped off is to run doun the line of steepest fall. with the result that there tends to be much too much water i n parts of the bendang and little ~v none in other parts. Eventually the concentration of water tinds the river. and goes to waste. unless prevented. The tendency is shown in the tigute The area A. B. C. D has too much water and the water runs to waste at D Area E. F. G is dry. To counteract this tendency, the farmers must so-operate rn forc~ng the water tcr move across the contours. instead of down it. This ih posstble: but farmets rarely show a sufficiently co-operative spirit, and the method does not work very well except in very small areas. A further disadvantage of the method is that rcltatton is impossible because supply from the canal to the bendang is entirely uncontrolled. *

7. Fig. 2!lI/III represents part of the area in Fig. l~I I / I I I . This time. however. field channels H I and J K are provided, supply to these being regulated t y controls at H and J . Water is supplied to the bendang only from these tield channels. and ~t is seen that a much more even distribution of water is obtained. while farmer ct)-operatton I S much less essential. If only sufficient water is passed through each tield channei head control, the water will be consumed in the field, and little w~ll go to waste. With this arrangement, rotation is possible, and this may be very dewable when water is scarce.

8. If the field channels are not too long, they may be laid strictly along the contour, and provision may be made for admitting excess water from the uphill side. Thus such excess does not run to waste, but is redistributed for use lower down the area.

9. Another method of conserving water, which has proved effectlve. is shown in Fig. 3!II/III. Here excess water delivered from H I is collected by a drain L M, just upstream from the second field channel J K. This drain is syphoned underneath J K and delivers the surplus water back to the hillfoot canal just below canal drop N. Such canal drops are very commonly used in small schemes . to control the water level at field channel head controls. The water thus tecovered IS used again lower down the area. This arrangement has much of the advantage nf the rotatmat method. as a high rate of application can be used, say from H I. without loss of water. Where the continuous system is too deeply rooted in the minds of the farmers, they may prefer this method to the rotational one.

10. Where field channels or drains are placed on the contour, as described above. it may be dewable or necessary to provide drainage end controls, as at M in Fig. 3IIIfIII.

11. Irrigation channels take up valuable space which would otherwise be productive, and therefore field channels should be built only when absolutely necessary. With a reasonable degree of farmer co-operation, it is possible to dispense with them in some cases. An example is given in Fig. 4/1IIIII. The chain dotted lines represent the contours. The full lines represent the small bunds or "batas" between the individual petak. Gaps in the full lines represent wide openings in the batas, controlled possibly by planks placed on edge and set so as to ensure a minimum depth of water of say 3 inches in the upstream petak.

CONTOURS

FIG 1/11/111

Double lines across the batas indicate small diameter pipes. Water can move freely through the openings in the batas, but flow through the small pipes is restricted. It will be seen that the strings of petak along the ridges shown shaded, act in fact as very broad and shallow field channels, from which water is tapped off, by pipes, as required, to the lower petak. In a similar way, strings of petak 3t lower elevations can be connected together for redistribution of concentrations of water.

12. With the greater farmer co-operation and awareness which may emerge under the community development drive, the method just described may often ptove to be best; it dispenses with the necessity for field channel upkeep, it provides no obstacle to movement of flood water, and no land has to be reserved for field channels. It i s unsuitable, however. for irrigation of crops other than rice.

13. With care and attention, these small valley schemes can be made to work quite well. Without attention, however, they can be complete failures. The most important point to note perhaps is that water needs no engineer effort to make it move straight down the slope. It needs endless patience and attention. however. to make it move across the slope. It stands to reason then that water should never be encouraged to move across the slope unnecessarily. Quite enough effort will be expended in making it do so when it is necessary. And when it is necessary a whole hearted effort will be essential to get results.

14. At the other end of the scale. in contrast to the small valley areas. are the broad flat areas of heavy soil such as are found in the lower parts of Krian. Such areas are, in effect, broad shallow pans. There is no difficulty about distribution, so long as there is enough watef to meet plant requirements, evaporation and percolation losses. All that is really necessary is to supply water at convenient points in ordet to keep the pan topped up. The supply points must not be too far from any part of the area, because if they are, supply at the beginning of the season to remote parts will be too long delayed. It is difficult to prescribe limits. but it is suggested that no part of a level bendang should be more than half a mile from a supply point.

15. Drainage in such flat areas may be more of a ptoblem than irrigation distribution. It must be possible to remove flood water reasonably quickly. and it must be possible to drain oii quickly at the appropriate time before harvest. Flat land is not conducive to quick drainage, so it is suggested that no land should be more tRan about a quarter of a mile from its drain. To prevent loss of water during the *igation season, drains should be provided with gates to hold up water levels in the drains.

16. A general arrangement as in Fig. 5NIIrII would be quite suitable for rice ~rrigation on heavy Bat land. The arrangement would not be adequate for other crops.

17. Intermediate in type between the small, steep. narrow valley a l a s and the broad flat areas just described, there are the extensive, even areas of gentle slope. such as are found in Tanjong Karang and Kubang Pasu. The slope of these areas may be a foot or so in the mile. They are too steep to be treated as flat. yet their slope is so small that the detailed atlention accorded to small valley areas would be quite out of place and very untconomica!.

18. Flg. 6'IIIIII shows the peneta1 pattern of imgation and drainage in Tanjong Karang. The general arrangement of lots is also indicated. The land is near]? enough level for it to be possible to irrigate across the slope. as opposed to down the slope. without special difficulty. and the arrangement works quite satisfactoril? and has the advantage of cheapness. There is of course a general tendency for water to move downhill towards the sea, but this is countered by the cross bunds. Such bunds are constructed to stop movement of watef towards the lower part of the arza. and therefore they should not be pefforated by pipes and

MAIN C A N A L

the like. except in special circumstances. As such perforations go contiary to the design intentions, such "special circumstances" should be discussed with the designer, before any action is taken. This point is most important. The purpose of

ctoss bund will be completely defeated if the bund is riddled with pipes. It will be found impractical to remove pipes again after they have been allowed, so it is imperative that pipes be not put in except with the designer's authority.

19. The general artangement in Kubang Pasu is similar in principle to that in Tanjong Karang, but is complicated by the fact that an irrigation system had 10 be superimposed on an existing, though inefficient, drainage system. The waterways running from the main canal to the sea have to setve the dual purpose of migation and drainage. They are much larger than at Tanjong Karang, and it is ndt practical to ptovide numerous offstake points in their banks. They are also too far apart to allow irrigation across the contour. Irrigation is actually "uphill" from distributaties lying immediately above the cross bunds. This uphill irrigation :s possible for two reasons. Firstly the land is nearly enough flat that a maximum dlowable depth of water at the distributaty ponds back a long way towards the next distributary on the uphill side. Secondly. the presence of contour bands of \rater. positi~ei) heid and topped up as necessary, makes it much easier to ,onserve water reaching the land in the form of rain. Rainfall in the North Kedah Plain is favoutable for wet season rice production, and conservation of rain water 1s as important as provision of irrigation water. Were it not for this fact, the method would not have much to commend it. The distributaties can be used also d\ dram. to discharge excess water. Again attention is drawn to the fact that perforation of the cross bunds would defeat the purpose of the bunds, which is to prevent water moving downhill towards the sea. Fig. 7IIIIIII shows the general irrangement on a section at right angles to the sea.

$0. The atrangements at Tanjong Karang and at Kubang Pasu are suitable for rice cultivation but would not be at all suitable for irrigation of crops not tolerant to standing water and waterlogged soil. It is doubtful whether the system a: Kubang Pasu would be adequate for off-season rice cultivation, because, as explained already. it depends largely on combination with direct rainfall.

1 The system adopted in Sungei Manik and in Trans Perak Stage I is more su~table for steeper and less even land. It is illustrated in Fig. 8/1I/III. Every lclt on the distributary frontage has a supply point, and every lot on the drain fmntage can discharge to the drain. Thus each pair of lots can work more or Ic*s independently of other pairs, as regards irrigation and as regards drainage. lu practice. individual drainage may be difficult, as water level in the drain is usually regulated for the common good.

22. The arrangement just described is quite a good one. It could not be improved on much except by making individual lots natrower and longer, so that tach lot ran the whole way from the supply canal to the drain. Each farmer would then be independent of other farmets. This is not so important where only rice is grown. But it IS important where other crops are to be idigated, as will be explained later.

23. With reasonably favourable rainfall, a great deal can be done on flat. heavy land. merely by careful conservation of rain water. Such schemes are called ' controlled dtainage schemes". because they conserve water by controlling the drainage within them and from them, to the best advantage. Where there is no .:deqcate sources of irrigation water. a controlled drainage scheme is the best that can be done. Some schemes, destined finally for irrigation, start off, however, as controlled drainage schemes. The reason is that in the tirst few years of opening up an area from swamp jungle. a great measure of success can be obtained without irrigation. and it is cheaper and quicker to open up initially for controlled drainage. Tanjong Karang and Kubang Pasu are examples of areas developed in this way. The layout of such schemes is as described fot Tanjong Karang or Kubang Pasu, but without any irrigation supply.

FIG. 7/11/111

ARRANGEMENT I I I I I OF LOTS

h - I - 7 - r

DRAIN

24. The Pahang Paya system of irrigation is peculiar to Pahang; and it is likely to remain so, for few ateas outside Pahang would be suited by the system. The Fahang Paya is generally a shallow basin-shaped depression lying near the Pahang River, with a narrow outlet to the river. When the river rises in flood, the payas are inundated. Consequently the rice crop hasato; be harvested before the annual tise of the river. It follows that planting must.start in the dry season, and that little reliance for irrigation can be placed od small streams passing through the payas. The system is illustrated in Fig. 9/II/Iq. The paya outlet is blocked by an earth dam, and a conttol gate and culvert &re-provided for drainage. Before the planting season. water is impounded in the pya and allowed to thoroughly saturate the soil to a great depth. At the appropriate time, planting starts round the edge of the paya, and follows the receding water as this is slowly lowered. In the higher part of the paya, short stemmed v rieties are planted; and in the lower parts. long stemmed varieties. When plantin!% complete, the water level in the paya is allowed to rise, but at a rate not gfeater than the rate of growth of the plants in the lower parts of the paya. Thus the crop depends for its growth on water held in the soil and in the petak, on such rain as may fall and on such water as may be safely impounded in the paya as growth proceeds. A detailed description of Pahang paya cultivation is at Enclosute No. (1) in DDI. 215'55.

APPLIC.4TION OF WATER TO THE LAND FOR CROPS OTHER THAN RICE

35. At present. no serious attempt is made to irrigate crops other than rice. except by a thousand Chinese vegetable gatdeners who hump the water in watering cans. There is little doubt, however. that there will be a call, sooner or later. for irrigation of other crops. Therefore a little is said about the subject here, if only to avoid rice irrigation layouts which will be unsuitable when the demand for irrigation of other ctops comes.

26. The aim of such irrigation will be to keep the moisture content of the soil. in the root zone of the crop being gtown, within the limits of moisture content acceptable to the plant. If too little water is supplied the plant will wilt, If too much is supplied, so that the water table rises into the root zone, the plant will drown. If the soil is free dtaining, so that the water table does not rise in spite of excessive irrigation, water will be lost by deep percolation and this water may reappeat elsewhere in a manner which may do damage to other agriculture. Also. plant foods will be carried away by the percolating water. Thus it is necessarv to apply the right amount of water at the ptoper intervals of time, and this is impossible unless the irrigation layout is correct.

27. The subject is amply covered in literature, including "Irrigation Principles and Practices", by Israelson, Hyd. 4211 in the D.I.D. HQ. library. This is recommended reading for layout designers. A quick appreciation, however, is given in the following.

28. For irrigation of field crops, the land is usually divided into long narrow beds with a gentle slope. The beds are commonly from about 30 ft. to 100 ft. wide. and from 300 ft. to 2,000 ft. or 3,000 ft. in length. The method of irrigation is to supply water at the top end of the strip, and ta let it run down to the bottom end, wetting the gtound as it flows. Usually a shallow drain is provided at the bottom end of the strip. but this is not meant to remove excess irrigation water. There should not be a@ excess irrigation watet. The general arrangement is shown in Fig. lO'II/III.

29. In Fig. lOIIIIIII, assume that the irrigation supply is turned into the bed shown, and that after a certain time the sheet of watet flowing between the batas has reached line B. Of the water being supplied at the head of the bed, some is now percolating into the soil over the area A1. which has been covered, and only the balance is available f a ' moving on. When the water has advanced to line C, water

SCALE: OF MILES '/r o II~UILC

is percolating into the soil oyel;@e g Az, and still less is available for moving on beyond line C. Thus the rate of 86 a vance of the water down the bed is steadily getting less. Eventually if the bed is long enough, when the rate of percolation into the ground is equal to the rate of supply at the head of the bed, the water will stop advancing. Thus it is seen that the longer the bed which is to be irrigated by a stream of a certain size, the greater will be the average depth of water supplied before the advancing water reaches the end of the field. If the bed is too long fot the size of the irrigation stream, the average depth of water supplied is too great, and there wilI be waterlogging or loss of irrigation water and plant nutrients by deep petcolation. For a bed of a given width and length, the average depth of water can be reduced by increasing the rate of application. Evidently light soils to which water infiltrates freely require higher rates of application or shorter beds. Beds can be somewhat longer if they slope more steeply, as the water then advances mote rapidly and irrigation is completed more quickly. But ~f the slope is too steep. the irrigation water may erode the surface. On the other hand, for heavy soils, longet beds, flatter slopes and lower rates of application are appropriate. The science and art of application of irrigation water to land for ctops other than rice, aims to get the right combination of bed size, dope and rate of application. for the crop being grown. It is not normally feasible to change the layout for different ctops. The layout will usually be determined once and for all on consideration of the nature of the soil, the crops likely to be grown and the maximum size of irrigation stteam which may be available. Latitude in operation can be obtained by varying the size of stream for the purpose required (e.g. a relatively small rate of application would be used for a deep rooted crop). or by controlling the rate of movement and percolation of water by the use of furrows or by checks, etc.

30. We have little knowledge in Malaya at present on the subject of irrigation of other crops. It is suggested, howevet, that if we lay out our new irrigation schemes on the lines shown in Fig. 11/II/III, where thisl is possible, we shall have schemes which are suitable for rice and for most other field crops, and which will not be unduly expensive.

31. The suggested length of field channel is not irrelevant. It is based on an assumed channel capacity of 2 cusecs, it being assumed that, when crops other than rice being grown, the whole of this is applied in turn to each bed, and that this is as much water as can be handled comfortably by a single farmer. If the channel capacity is 2 cusecs, and the duty is 60 acres per cusec, then the total area served by the field channel must not exceed 120 acres. Hence the length of the channel.

32. The arrangement suggested would give a rate of application of 0.4 acres per cusec, if each lot were tteated as a single bed. The lots could be longitudinally subdivided, however, thus increasing the rate of application to any figure which proved to be desirable.

33. The lot length proposed, i.e. 30 chains. is a comptomise with economy and practicability. To reduce the length to 20 chains would result in an expensive layout and one in which too much land was used for canals and drains. Lots 40 chains long, on the other hand, might be difficult on account of irregulatities in the land surface.

FIELD CHANNEL

---- DRAIN

SECTION Ill - SOURCES OF IRRIGATION WATER

1. There are three sources from 'which water can be obtained for the irrigation of padi. namely direct rainfall ovef the padi fields, from rivers, streams or lakes, and from underground water.

DIRECT RAINFALL

2. The traditional padi fields in Kedah, Perlis, Kelantan, Perak, Negeri Sembilan and Malacca otiginally depended on rainfall for padi cultivation, and the fanners planned their ploughing and planting of nurseries at the beginning of the wet season. The planting dates and other agricultural activities would therefore be dependent on the start of the rains, and one of the most important features in these areas is the existence of well formed batasses between lots laid out on contours. These batasses are absolutely necessary for the retention of rain water within each lot for the successful cultivation of padi, and even to-day there are many small sized padi areas scattered throughout Malaya where the farmers use this method for padi cultivation. The success or failure of a crop depends on the availability of rainfall.

SURFACE WATER

3. The second source of irrigation supply. i.e.. from rivers and streams has been most extensively used in Malaya.

4. In Malaya. the existence of a mountain range running roughly along the centre of the peninsula has resulted in large numbers of rivers or streams flowing eastwards and westwards to the sea. This feature coupled with the heavy rainfall and thick forest cover has ptoduced conditions where streams seldom run dry and provides excellent sources for irrigation and other water supply purposes.

5. The most common methods of abstracting water for ifrigation from surface water are listed below: -

(a) Direct tap off from a river by means of a side intake. (b ) Barrage or headworks across a river or stream. (c) Dam across a valley to form a sforage reservoir. (dl Pumping from a wide river. (e) Low bunds across a valley to inundate the padi fields within the valley.

or control structures across rivers to regulate the water levels and inundate the surrounding padi fields.

6. Where physical conditions permit it is most economical to abstract wate- from a river by a side intake suitably located along a river bank. The important factors to be taken into consideration are :-

(a) Flow in the river during dry periods and quantity to be abstracted. ( b ) Water levels in the river at abstraction point. and whether these are

stable. (c) Amount of silt load and suitability of water for irrigation purposes. fd) Flood levels and exclusion of flood flows from t h e iniake canal. . -

This can be catered for by judicious bundin? and provision of gates at the intake structure.

7. It is generally advisable to carry out detailed surveys of the river 1 mile upstream and f mile downstream of tne proposed abstraction point and collect hydrological data covering discharges and corresponding water levels to enable the conduct of a model study of the intake to verify the hydraulic performance of the structure.

8. A typical side intake structure used for extraction of 600 cusecs from the Sg. Perak for irrigating the Trans Perak Stage IV scheme is shown as Figure 1 /III/III.

9. The most extensively used method of extraction of water from a river is by means of a batrage across the river with an intake of suitable size to meet irrigation requirements. The structure usually consists of a reinforced concrete Boot slab with side walls and a steel segmental gate.

10. During irrigation seasons, the gate is lowered in order to raise the river level so that this is high enough to command the irrigation area. Normally the structure is large enough to pass the flood waters when the gate is lifted. Sometimes for reason of economy, a headworks structure is built with a limited flood discharge capacity and an earth flood spillway provided along side to cater for excess flood waters.

11, The selection of sites for headworks structures should be based on the following considerations: water level command; suitable foundation soils; ease of access to facilitate construction and also the minimum compensation for loss of agricultural or village land as a tesult of inundation due to the raised water levels upstream.

12. A typical headworks structure is shown in Figure 2/III/III, where two numbers of 3'-0" x 12'4" segmental gates are provided.

13. For small structures where it is not economical to station a gate keeper to operate the headworks and off-take gates, an automatic type of gate of the trip type or a bear trap gate can be provided. Fig. 3/III/III shows a 20 feet wide Bear Trap Gate.

14. It is usual to incorporate a drop in the headworks structure to cater for the tetrogression of the downstream river bed, especially in steep riverine valleys. Fig. 2/1II 1111 shows the 14th mile Terachi Headworks with a 9 feet drop.

15. Where the dry weather flow of a rivet is insufficient to meet irrigation requirements, particulatly in the case of double cropping, then the solution will be to study the possibility of constructing a dam across the valley of sufficient storage capacity to meet the demand, based on water requirements for padi cultivation.

16. In this case, studies will need be made from existing top0 maps, aerial photographs and actual surveys including geological studies to locate a suitable site for a dam. The site should be across a narrow valley with high ground on both flanks and with a reasonably large reservoir capacity upstream.

17. Detailed soil investigations should be carried out at the proposed site to determine the most suitable type of dam and its height, and in most cases of low dams up to 50 feet high, earth filled dams would be most suitable, in which event the source of fill material will also require investigation.

18. Hydrological data including rainfall, mean fiver discharges and maximum flood discharges will be required for studies on irrigation storage and flood spillway designs.

19. An earth dam with the intake and flood spillway at Bukit Merah in Krian is shown as Fig. 4a/III/III & 4b/III/III.

20. A reinforced concrete buttress dam of 100 feet high on the Sg. Muda and a rolled rock fill dam 200 feet high on the Sg. Pedu for the, Muda Irtigation Project, with a concrete lined canal 44 mile long connecting the two reservoifs is shown in Fig. S/III/III.

21. Should a river from which water is to be abstiacted for irrigation be very large, or where a considetable rise in water level will be necessary to enable gravity irrigation, then pumping will be the best method of abstraction. For this

reason, pumping stations are mainly located at the lower reaches of larger rivers in the country.

22. Generally the choice of a pumping station is based on the following considerations: sufficiency of flow at all times to meet maximum pumping requirements, easy access. freedom from silt and saline intrusion at pump inlet and non-submersion of the station during floods.

23. Information required for the design of a pumping station includes a survey of the tiver for a distance of 4 mile upstream and ) mile downstream of the proposed site, and maximum and minimum water levels together with discharges. It is advantageous to locate a pumping station at the outside of a bend to lessen the problem of siltation.

24. Irrigation pumps are generally large-capacity. low-head pumps. and a \ ~ a l flow or mixed flow pump are most suitable for this purpose. The pumps are operated by eithet electrical motors or diesel engines dejxnding on the availability of electrrcity supply and the economics of using either diesel or electric drive taking into consideration of both capital and operation and maintenance costs.

25 As a general guide. diesel engines are more costly than electric motors. but cheaper to run. Electric motors are smaller, less noisy and easier to keep clean and do not requrre fuel storage tanks with no problems connected with fuel suppl!.

26. A typxal pump house with 3 units of 150 cusec axial flow pumps operated by electric motors for Pinang Tunggal Pumping Scheme is shown as Fig. 61111 111. while a pump house for the Nerus Irrigation Scheme with 3 units of 90 cusec axial flow pumps using diesel engine drive is shown as Fig. 7IIIIIIII and a submersible pump for Pekula Pump House is shown as Fig. 8/III/III.

17. As a general rule the operation costs of pumped irrigation is higher than that for other types mainly because of additional pumping costs, and should only be adopted when other methods are not acceptable.

18. When a padi area IS located in a flat valley and there is no reliable water supply either from a stream or from a suitable source of water nearb,c. then an inundation scheme consisting of a low bund across the valley provided with a control gate i s the most suitable method. This type of schemes are exclu- sively located in Pahang within the Pahang River Basin.

29. The irrigat~on areas are usually small ranging from 50 acres to 500 acres and the stream catchments are of I to 10 square miles in area. Earth bunds are formed to heishts of 4 to I2 feet with a control gate to discharge flood waters. Sometimes an earth spillwa) IS also provided alongside the gate to facilitate the passage of flood flows

30. These schemes are generally considered as "sub-standard" schemes and are not ~dzal. During the monsoon periods the gate is closed and water is retained behind the bund to soak the field. As planting takes place, the stored water is gradually released and from then on the padi plants depend only on rainfall for growth. as the stream generally runs dry. Yields for this type of scheme are very low and only one crop can be planted out a year.

3 1 . A typrcal layout of an inundation scheme is shown on Fig. 9;III '111.

32. Similar to these. but embracing larger areas and located on coastal plains are the so called controlled drainage schemes. Noteable among these are the Panchang Bedina Scheme in Selangor. the Wan Mat Saman Area and the Kubang Pasu Areas in Kedah.

33. A main feature of these areas is that the latge tract of padi land has relatively small fall. One or more rivers may traverse the area, part of which may be subject to saline intrusion. Structutes are built across suitable sections of the river with gates to regulate the water level and cause backing up thus inun- dating the adjacent fields to varying depths. During floods the gates are opened to let out the flood waters. Gates are also opened during harvesting and ofheason period to drain out water from the fields.

34. These schemes lack assured water supply, as during dry years river flaws may not be sufficient for flooding the higher fields. Watet depth in the fields is not uniform, with the lower fields flooded to greater depths than higher fields. Drainage during rainy seasons often presents problems.

However these schemes are cheap and during years with well distributed rainfall reasonably good yields have been obtained.

GROUND WATER

35. Hitherto irrigation using ground water has not yet been done in Malaya, because of the availability of surface water in the majority of cases. and the inadequate knowledge of ground water resources potential.

36. In places where the exploitation of available surface water is approaching its limit, there will be a need to look for possible ground water resources for fuither irrigation development. Much work remains to be done in this connection.

SECTION A - A

OPSRATING MAR PPLRATING PLATFORM

U ~ S ELEVATION SECTION A - A

D ~ S ELEVATION

14 TH. MILE TERACHI I RR1GATION SCHEME

GENERAL PLAN OF HEADWORKS

F I G . 2/%/m

SECTIONAL ELEVATION

PLAN -

QPSTREAM ELEVATION

AUTOMATIC BEAR - TRAP GATE

-.

CROSS SECTION OF WE18

W A T E R LEVEL V

m r w IMTAUE ,

IRRIGATION RESERVOIR

BUKIT M E R A H RESERVOIR

GENERAL P L A ~ OF INTAKE \

1 R E M O V A B L E

TYPICAL SECTION THROUGH PENSTOCK OF I N T A K E

bUKlT MERAH RESERVOIR F 1 G . 4 b/111/111.

T Y PlCAL SECTION OF RAILWY EMBANKMENT

AIUlMJiNT POI-SCILLWIYI SPILLWAY 8 U T T R t B S t S pm. SPRLWAV JAIUTMENT BLOCK @ L u r k T B U T T I E J S E S

C U T W

UP

OFF* A L L

Q SPILLWAY CREST

DOWNSTREAM SLAB

' S T R E A M S L A B T A l N l N G W A L L

Sf

CREST

MUDA IRRIGATION PROJECT 105 C T . UlCY 60,000 CU YO). C O M C R E T E

0

CAST !N S I T 0 R t l N F 0 R C € O C O N C R E T E A R T ~ C U L A T E D SCABS OM TOP OF Y E U ~ ~ A U

,/ 2 2' cn

FOR S C U C C U RAKE R A I L S . FOR ORTAILS S E E DRAWING No 2 0 9 1

2

C U T OFF WALL-\

ILL

REGULATOR VALVE B L O C K F O R D E T A I L S

V A L V 8 IVLCAIL AUO V A L ~ C I ~ U T I - S ~ L V A C U U M O ~ ~ W I ~ C A N D nu. 4111 C o b SECTION THROUGH DAM SEE D R A W ~ H C NO 212 J

D E T A I L S OF MUDA DAM DETAILS OF PEDU DAM

GLASS PANELS R.B-PANELS WTM~GLAMROCK'FINISH EXTERNALLY

CRANE RAIL 6 GANTRY. COLUMNS FOR CRANE RAIL 6 GANTRY SPACED

RECORD€ R H O U S E O I 2 ' - 0 " CR5. 12"s 18" A T THE TOP, 9"r ! * * A T THE BOTTOM.

~ / l ' E x ? ~ w s ~ o u J o w l .

1 / + 18-20 MAS. I A- F.S.L. + 1 7 . 0 0 ..... I

1 DOWNSTREAM ELEVATION

UPANSWN JOIN?

WA?CR STOP

+IUEOF M S. SHEET PILES

CA? L A W =

360 H P. DlESLL ENGINL, DRYSDALE , M 5. SHLCT PILES 42" BORE 90 CU3LC CAPACITY VCRT ICAL SPINDLE AXIAL F L O W

T Y P E PUMPS TO BL SUPPLIED BY DRYSDALE. L CO. LTD

GENERAL SECTION ALONG &

CoUC. GI)lLLCS,, n I i

/

~"THICK COW APRON ON 4"THICK +RDCORC t

ENTRANCE

SG. NERUS IRRIGATION SCHEME

GENERAL I AYOUTOF PUMPING STATON

NECO LOUVRE G L A Z E D PANELS

CONCRETE

L .

PLAN -

PEKULA I.RRIGATION SCHE

PUMPING STATION

SECTION A - A

P L A N PAYA SG. LENG INUNDATION SCHEME

CONTROL GATE S ITE PLAN

F I G g/m/IIZ

DIAGRAM CONNECTING

HL AND E f t FOR GIVEN DISCHARGES

PER FOOT RUN

(TROUQM EUPM)

FIG. 4 / ~ / l l l

SLOPE I : I C-

13:O"

1 SLOPE I : (

S E C T I O N A - A .

04 ...... .,.. I.... L i .. . - ,.:.-.i .......... i.... T O M A I N C A N A L

+ 22.00 .. ., .. :..L , . : ...,: .- .. <. ., .-,. . . . - --.- . - .. : .. -.. - . , / 'd ' "1

OFFTAKE C U L V E R T

EXISTING R.C S L A B 8 6 : 0 "

A

S E C T I O N 8 - 6

TANJONG KARANG IRRIGATION SCHEME

BERNAM RIVER HEADWORKS SILT EXCLUDER

SECTION 1V.-DISTRIBUTION

1. As noted in Section I, the distribution system of an irrigation scheme may include a main canal, branch canals, feeders and distributaries, and will certainly include field channels. In fact, the whole system aims at getting the water into the field channels, from which it is passed to the field to do its work. Hence, all channels down to and including distributaries are conveyance channels and should be as etlicient as possible as conveyors. They should be as small as possible (unless they have an additional storage function) and they should go down lines of ample fall. Field channels, on the other hand, have to distribute water to the land, over which is convenient for it to follow the line of steepest fall. Hence, to be efficient (efficiency can be measured by area commanded per foot run of channel), field channels must run as nearly as possible along the contom's. A field channel on the cont:>ur has also the advantage of having a very flat grade: thus the head over offtake pipes will remain more ot less uniform and will make uniformity c-f distribution easier. Where a field channel has to be on a substantial slope, water lcvel controls at frequent intervals become necessary.

2. Field channels should be designed for at least 6 inches command over the land to be irrigated. Nine inches is better and should be provided where possible. More command would be better, but channel banks become too high for easy maintenance. If the command is too small, irrigation from an offtake pipe can be completely stopped by failure of the cultivator beside the canal to allow passage of water through his batas.

3. Because a field channel lies across the fall of the land, it constitutes an obstruction to drainage. It will cause flooding if too long and unbroken. Where possible it should be given a slight fall, and it should not be too long. For other reasons, a maximum length of about 3 mile has been suggested already. If channels must be much longer than this, arrangements must be made for cross drainage.

4. Field channel size may be determined by the discharge to be cafried. If a rotational system is to be adopted in which rotation between field channels (as distinct from between lots on one field channel) is to be practised, then due allowance must be made. Frequently, however, the size of the channel is decided by the amount of eatth which must be found to build the banks, and the size is greater than hydraulic considerations demand. Where the D.I.D. is to maintain the channel, the banks must be initially sufficiently robust, and hydraulic requirements are unlikely to affect the issue. It is better, and more in keeping with the spirit of "community development" that the farmers should build and maintain their own field channels. Farmer maintained channels can be much smaller, and hydraulic requirements may dominate. A further reason for preferring farmer built field channeh is that small channels of hydraulic section make a scheme more positive in operation. Where channels are of unnecessarily large cubic capacity, the system becomes sluggish on account of reservoiring of water within the canals.

5. Whete a field channel is built by the D.I.D. for Government maintenance. minimum dimensions must be as in Fig. I/N/III. If built to a lesser standard. maintenance becomes impractical. By contrast is shown, also in Fig. l/IVIIII, a field channel to dimensions which have proved adequate, when the farmers them- selves build and maintain their field channels. A small channel like this needs to be repaired immediately a weak point develops. So long as this is done, it is perfectly serviceable.

6. Water must be taken off from field channels by means of pipes; not by making cuts in the canal bank. Pipes are orifices of a sort, and the discharge through them is proportional to the square root of the head. Therefore they ate not unduly sensitive to small changes in water level in the channel. Pipes may be hollow bamboos, glazed earthenwate pipes or asbestos cement. They may be uncontrolled, In which case supply is stopped by inserting a clod of earth or the like. Or they

may be fitted with plugs or rudimentary buttetfly valves. There is no point in using anything more complicated and expensive than circumstances dictate. Otltalces through D.I.D. maintained banks should be of a permanent material. howevet.

7. Where the land to be irrigated is uneven, water will tend to gravitate to the valleys. Offtake pipes should then be placed to delivet onto the ridges. Delivery, if any. direct to the valleys. should be very restricted.

8. Field channel headgates will be discussed together with other controls of the distribution system. Sufficient to note here that they must be big enough to supply the requirements of the Held channel, taking into account where applicable rotational irrigation as between differen1 lield channels.

9. The purpose of the m a n branch. feeder and distributary canals is to convey water to the field channels. Normally, they should not be used to supply directly ti) the field. but the occasional field offtake is difficult to avoid.

10. As a conveyance canal does not normally have to have direct command of the land. it should be placed in cut whenever possible. as there can then be no breached banks.

I I. In hydraulically designing a conveyance canal it is necessary to ensure: (a) that its capacity is sufficient. and ( h ) that i t will be stable.

By stability is meant that there is no progressive change of slope or of cross section. In fact. the channel should be in regime. For a given discharge and sediment charge. there will be stability only . i f width. depth and slope are acceptable. taking into account the material of which the channel banks arc formed. An appreciation of this fact is important. Two examples will make the principles clear.

12. Consider a channel which is in regime with a certain discharge and sediment charge. Let the discharge remain the same but increase the sediment charge. The additional sediment will deposit at the head of the channel and a steeper slope will build up progressively till the higher velocity over the new slope can carry forward the whole of the new charge. With increase in velocity. the drag of the water on the wetted perimeter increases, and this may now exceed the permissible intens~ty of tractive force on the banks of the channel. The banks will then erode and fall. thus causing the channel to become wider and shaIlower. until--as a result of decrease in velocity and in depth in the wider and shallower section-the attack on the banks ceases and stability is once more achieved. A new regime has been attained in which all dimensions are different. Thus it is seen that discharge. sediment charge. bank material, slope and cross section are all interrelated.

13. Again. consider an excessively wide channel carrying suspended sediment. Drag at the sides will be so low that some of the sediment is deposited. causing the channel to become narrower and deeper.

14. It is interesting to note too that where the sediment consists of coarse particles only, the channel may be able to silt its bed, but it will not silt its banks because the coarse sediment is travelling along or close to the bottom. Similarly a channel carrying fine sediment only may grow narrower by deposition on its banks. but will not silt its bed unless velocities are very low. Thus a channe! carrying turbid water but no sand tends to be narrow and deep. and vice versa.

15. Canal regime theory, developed in pre-partitioned India, expresses these principles quantitatively. Blench's exposition in "Civil Engineering Reference Book". Han. 62 1 in D.I.D. HQ. library. is brief and readable. and is recommended for study.

16. The formulae given by Blench, cast in convenient form are:

where W = mean width of channel, taken as trapezoidal in design. D = depth of bed below water level, assumed level in cross section. S = slope, b = bed factor, r is the kinematic viscosity.

In the absence of a bettet estimate, the bed factor can be taken as a first approximation as

b = 1.9mt where m is the mean diameter in mm. of the sand exposed on a canal bed in regime.

s, the side factor. vanes from about 0.1 for rather cohesionless banks to about 0.25 for particularly tough banks.

The value of r is round about unity. The formulae give width and depth in feet. Slope. of course. is dimensionless.

17 The formulae are quoted by way of illustration Their use in practice should not be attempted without studying the theory and discovering the limita- tions. methods of estimatrng b and s. and so on.

18. By way of example. the table below has been prepared from design curves, given by Blench. for channels with a small charge of bed sediment, such as might be expected In a normal well regulated canal system. A bed factor of 1.0. corresponding to a sand of mean diameter about 0.3 mm. has been assumea. And side factors of 0.25 and 0.15 corresponding to tough banks and average banks. have been assumed.

Discharge Cusecs W~dth ft. Depth ft Slope Width /Depth Bed factor= 4.0 Side jactor=O 2.7

5 4 45 1.08 00033 4.13 10 6.3 1.36 .OOO3 4.65

100 20 2.92 .. . .00020 6.85 1 .OOO 63 6.30 .OOO15 10.0

Bed factor= 1 .O Side factor=U I S 5 5.75 0.91 . .00032 6.32

10 8.15 1.14 . BOO29 7.15 100 25.8 2.46 .00019 10.5

1,000 81 5 5.30 . .00014 15.4

The points to note from the formulae are the following.

(al Large canals In reglme have a greater width to depth ratio than small ones.

rh) Large canals In regime have a flatter slope than small ones.

(c) The coarser the bed material, the greater the regime widthldepth ratio. (d) The more erodible the bank material, the gteater the regime width/

depth ratio. (e) The coarser the bed material, the steeper the slope. ( f ) The erodibility of the banks does not affect the tegime slope very much.

The most important point to note from the table is that regime sections are far removed from the "most efficient" channel sections of conventional rigid boundary hydraulics.

19. To what extent is regime theory applicable in Malaya? The question is worth asking, for we have generally managed quite well without it. At any rate, this is probably the belief of many. who have just regarded the misbehaviour of a canal as "one of those things". There is no doubt that the theory is applicable to canal systems taking off from sediment laden streams. Also there is no doubt that in the past a numbet of our canals have been constructed to non-regime dimensions and slopes. Probably the most common fault has been to make sections too deep and too narrow, so that there has been a tendency to silt and to erode banks. It is suggested that regime theory is worth applying in the case of canal systems supplied from sandy tiver headworks, at least in the main canal and in all canals to which bed sediment has access.

20. For pumping schemes and systems supplied from impounding reservoirs, there will be no bed sediment in the canals unless it is brought in by streams entering the canal system or by bank or bed erosion. If sediment does become present, the canal will attempt to acquire a regime section and slope. To avoid erosion of banks in small canals of this kind, the following are the mean velocities. beyond which trouble will almost certainly occur:

Sandy loams ... ... 2 ftlsec. Clay loams ... ... 3 ftlsec. Stiff clays ... ... 4 ftlsec.

For larger canals, a regime theory check is advisable.

21. It is safer and better to err on the side of excessive width, if in doubt. With a little care, a canal which is initially of too wide a section will build itself tough berms, if there is a reasonable amount of sediment in suspension, and these berms will be beneficial. If, on the other hand, the canal is made too narrow, it will scour its banks and silt its bed, and permanent repairs will be impossible unless the faulty bank alignment is rectified by setting back. This may be difficult and expensive.

22. Whether or not regime theory is applied in any pafflcular case, there is much less possibility of blundering if the principles have been digested.

23. It is probably appropriate to mention here that except when a canal is in cut, the initial cross section of waterway is not usually determined by hydraulic considerations at all. Usually the banks of a canal which is requited to command the land will require more spoil than can be obtained from a hydraulic section of waterway. Therefore the section is excavated large enough to obtain the required amount of bank spoil. It is better in such cases to excavate wide and shallow rather than deep and narrow. In developed areas, however, land acquisition will be expensive if this principle is pressed too far.

24. The capacity of any particular canal must depend on its place in the system and the type of system. The main canal obviously has to carry the maximum demand of the scheme, taking into account any losses which are likely to occur.

25. It does not follow, however, that the sum of the capacities of branch

canals should be equal to the capacity of the parent canal. If the system is a rotational one, in which rotation is practised as between different parts of the system, then the canals supplying the different parts must have sufficient capacity to supply the maximum demand of the parts served. For example, if a feeder supplies three distributaries, and rotation is to be practised between the three distributaries, then each of the distributaries must be of the same capacity as the feeder.

26. Losses from canals may be by evaporation and transpiration, by percolation, or by plain leakage. Loss by leakage should not be allowed for. Tho correct course is to stop the leaks. Transpiration and evaporation losses from canals in Malaya do not seem to be at all serious, and can be ignored. Per- colation losses from canals in heavy clay are negligible. It is only percolation losses from conveyance canals on sandy or sidelong ground that may be serious Each case will have to be examined on its own merits. Rule of thumb cannot be applied. Where loss is serious, lining may be necessary. Canals may be lined with a skin of concrete or mortar, or bitumastic or plastic membranes may be placed in the soil, around or under the section. Usually in Malaya, canal losses are not serious, and in this we are fortunate.

27. Where possible, canals should be in cut, in order to reduce recurrent maintenance costs.

28. Where a canal cannot be put in cut, the water must run between banks. Fundamentally, the banks of a canal are for retaining water in the canal, and must be stable when regarded as dams. They must also be reason, ably watertight.

29. Rather arbitrarily, a percolation gradient of 4 to 1 is generally accepted in Malaya as good enough, for average good clayey bank material. See Fi8. 2/IV/III. Each case should be considered on its merits, however. If a bank shows seepage generally at its outside toe, then the percolation gradient is too steep for the bank material, and in all probability bank failures by piping wit1 be frequent.

33. The dimensions obtained by considerations of percolation may not bo the governing ones, however. Experience has shown that a bank which is to b maintained departmentally must have certain minimum dimensions. Down to and including distributaries, the following are the minimum requirements, which must not be set aside on any consideration of expediency.

Top width must never be less than 4 feet. where only pedestrian traffic is to be allowed. Where cycle traffic is to be allowed, this shouid be increased to 6 feet. For large canals and where special maintenance methods are to be used, (e.g. tractor weedin,.) the top width should be further increased.

Freeboard for distributaries and small canals must be maintained at not less than 18 inches. 2 feet or 3 feet of freeboard is appropriate for larger canals.

Side slopes for banks in the best materials should never be steeper than I f to 1. For less good materials, side slopes of 2 to 1 or evaq 3 to 1 should be adopted.

31. It is debateable whether or not berms should be provided. Fig. 3aiIVIIII shows, in full lines, a canal bank set back from the channel, leaving a berm between. The dotted line shows an alternative bank slope on a flatter gradient. On the face of it, it appears that the dotted line would be better than the arrangement with a benn. In practice, however, a flat bank slope below water level seems to be peculiarly susceptible to attack; and very aeqrly always, the section finishes up as in Fig. 3bINIIII. This, clearly, is less stable

CANAL BANK

WATER LEVf L

GROUND, - - -- -- - -

FIG. 2/1V/ll1

FIG . 3 b/lVlllI

than the berm arrangement, and bank repairs will be more difficult and expensive. Berms are very useful as walkways for excavators on desilting and repair work. At a pinch. berm material can also be used for emergency bank repair work. On the whole then, it seems preferable to provide berms, if space permits. Unfortunately it does not always permit. Where provided, berms should preferably be wide enough to run an excavator, appropriate in size to the maintenance work which may be required on the canal.

CANAL CONTROLS

32. Canal controls are required for the following purposes:

(a) To flatten a hydraulic gradient in a canal, which would otherwise be too steep.

(b) To provide for proper distribution of water as between different parts of a system.

(c) To obtain command at specific points.

33. Where the only purpose of a control is to maintain a gradient in a conveyance canal, a notched fall is probably the best type, as this can be designed to maintain parallel flow at all stagas, and it does not obstruct the natural movement of sediment down the canal. Reference may be made to "Irrigatton Pocket Book" by Buckley, or to other standard works on irrigation.

34. There has not been much application in Malaya, however. for notched falls. This is because canal systems are short, and usually a canal fall is required to command some other canal on the land, as well as to reduce the hydraulic gradient. Where command is required. the best type of control is one which maintains water levels at the control within close limits, without constant adjustment by the irrigation staff. The head above an overshot type of control will vary as the discharge raised to the ower 213, whereas the head above an undershot type of control will vary as & square of the discharge. Clearly then an overshot type of control is indicated where it is desired to maintain a constant command level at a point in spite of some variation in canal discharge. That is to say, the control should take the form of a weir, and not of a gated culvert. The greater the length of the weir, the less sensitive will be the u stream water P level to changes in discharge. Fig. 4/IV/III illustrates a typica small canal command control. as commonly used in the D.I.D. The long crest of the circular spill-well assures a practically constant upstream water level, over a wide variation in canal discharge.

35. It may be desirable in certaia cases to provide for periodic alteration to the command level, in which case the control should take the form of an adjustable weir.

36. Usually the object in holding a command level steady is to maintain a constant head on the control supplying a distribu&. in which a constant discharge is required. The discharge will be more or less constant provided the control is insensitive to changes in upstream water level. Thus the distributary control should be of undershot type. It generally takes the form of a gated culvert. A typical distributary head control is shown in Fig. 5/IV/III.

37. The general arrangement of a major conveyance canal command control and a distributary discharge control is shown in Fig. 6/IV/III.

38. Cpnsider what would happen if the controls in Fig. 6/IV/IIL were reversed. If more water came down the conveyance canal as a result of increased intake at the headworks, or as a result of closing down a branch canal hiiher up, the distributary would get almost the whole increase and would be hope- lessly overloaded.

TOP OF BUN0 -

1 1 . - -. PLAN F I G .

SECTION

PLAN FIG. 5/1~ / l l l

OVERSHOT CONTROL

(USUALLY A LONG WEIR)

UNDERSHOT CONTROL USUALLY A GATED R MAJOR CONVEYANCE CANAL

(VARIABLE DISCHARGE)

/ DISTRIBUTARY (CONSTANT DISCHARGE)

SECTION A-A. SECTION B - 8.

39. Similarly, command and gradient controls along the length of a distributary should be weir type controls, and field channel head controls should be undershot types.

40. A system in which main canal discharge varies will work smoothly and nearly independentIy with overshot main and branch controls and undershot distributary and field channel controls. With types reversed, it will need constant fiddling, with great sensitivity in one part as a result of fiddling in another.

41. In some schemes, a constant proportionate division of water is required, as between two or more branch canals. This can be achieved by setting weir type controls with equal crest levels at the head of the branches. (Orifice types are not suitable for this purpose, as they are sensitive to downstream water levels, which may vary in the branches). So long as the bifurcation control is not also a command control, the matched weirs can be narrow and deep, for economy, provided they are of streamlined section, as shown in Fig. 7/IV/III. Weirs of this type are not affected by downstream water level, provided the depth of submergence does not exceed 80 per cent. That is to say, if the head over the weir was 1.0 ft. the discharge would not be affected by tailwater level, so long as this did not stand more than 0.8 ft. above crest level. In fact the discharge is not seriously affected till over 90 per cent submergence. So long as a clear hydraulic jump forms, it can be taken that the discharge is not appreciably affected.

42. The undershot distributary head control already described is a very rudimentary form of "module". a module being a control aiming to give constant discharge notwithstanding differences in head. There are a number of more efficient types, but economic circumstances have not yet favoured their use in Malaya. A very simple type is the NEYRPIC module, developed by the NEYRPIC Research Laboratories at Grenoble. Details of this patent type are available at D.I.D. HQ. NEYRPIC has also developed a number of automatic control gates, which are in use in systems in other countries where water is very scarce. These might be useful in Malaya when more value is placed on water than at present. These types include automatic gates designed to hold constant upstream or constant downstream water levels.

43. The control regulating entry of water from the source 'of supply to the canal is called the canal head regulator. Where supply is from a sandy river, the location of the head regulator requires special attention. It must be placed at a point on the river bank where the surface movement of the water is towards the bank and the bottom water is moving away from the bank. This, of course. implies a corkscrew motion of the water as it travels down the river. and this motion is found on bends. The surface water moves towards the outside of the bend, and is clean, apart possibly from floating trash. The bottom water is moving towards the inside of the bend, and is sweeping the bed sediment towards the inside of the bend. The sediment is deposited on the inside of the bend, as we all know. If the canal head were on the inside of the bend, this sediment would be drawn into the canal. Hence, the canal head regulator should be put on the outside of a bend and must in no circumstances be put on the inside. Where possible, the bend should be gentle, as the bank may be unstable on a sharp bend. Also, the current may not follow closely round a sharp bend, but may bounce off the side, creating a reverse eddy which might lift sand, at certain parts of the bend. (This has happened, in fact, at the Muda River intake. The main current bounces off the outside bank a short way upstream of the intake, and a reverse corkscrew current lifts sand into the intake mouth). It is recognised nowadays that correct curvature of flow at intake is the most important factor in prevention of entry of bed sediment to a canal.

44. The corkscrew motion of water round a bend results from an unequal

tendency of top water and bottom water to back up on the outside of the bend as a result of centrifugal force. The top water is going faster and tends to back up more than the bottom water which is going more slowly as it is subjected to bed drag. Hence the top water tends to back up more than the bottom water. The unbalanced head cannot be sustained however, so at the outside of the bed there is a downward flow, and at the inside there is an upward flow. Consequently there must also be a flow at the surface from h i d e to outside, and from outside to inside at the bottom.

45. A further point about the canal head regulator is that it will draw a greater proportion of clean top water if its crest is high and its length is substantial.

46. As intakes are usually put immediately above weirs or barrages (these being of course river command controls), it foIlows that weirs and banages should be sited if possible on bends of correct curvature. Straight reaches of rivers are not so good, as the current inay tend to swing, thus setting up corkscrew currents with the wrong direction of rotation. The same principles apply to location of pump intakes.

47. Miscellaneous canal structures include syphons, flumes and bridges. No special comment is called for here. Approved types can be seen by arrangement with the A.D. Planning.

48. It is to be noted that most canal controls are energy dissipating structures and that they are also subjected to percolation pressures. They should be designed on the principles explained later in comexion with weirs and barrages.

SECTION V.-DESIGN PRINCIPLES FOR WEIRS (INCLUDING BARRAGES)

WEIR CREST L&EL AND LENGTH

1. irriiation. is most likely to be required in dry weather. h Malaya. where rivers are short.. this means that irrigation will be required when river discharges yu@ low and therefore when water level is low. Occasionally, with favourable' topography. low flow level in the river may be high enough to command tbe head of the irrigation canal. If this is the case, then the headworks consists simply of a canal head intake and regulator. Usually. however, it will be necessary to back up the water level in the river in order to command the canal head. If only a small amount of backing up is necessary, a plain weir may suffice. If more backing up is required a weir surmounted by crest gates is used. This is called a barrage. The weir or barrage and the canal intake and regulator are known collectively in Malaya as the "headworks".

2. The reason for using a barrage instead of a weir, where much backing up is required, is that a high fixed crest would result in raising of upstream water levels (afflux), at high stages of flow as well as at low stages. This would result in deterioration of the .river upstream, by aggradation, aild possibly in flooding upstream.

3. How high is it permissible to place a fixed crest? Consider Fig. I/V/III. On the left hand side of the diagram, frequency of flow at various stages is plotted, and also bed sediment load at various stages. By multiplying together frequency and sediment load at various stages, another curve--drawn on the right-is obtained, which shows the importance of various stages of flow, so far as shifting bed sediment is concerned. The most important stage is sometimes called the "formative stage". Different streams will have different characteristics. But clearly the most important stages from this point of view are rather above the stage of greatest frequency. If upstream aggradation is to be avoided. discharge corresponding to these important stages must not be backed up. but must be allowed to pass unimpeded. It would probably be safe to raise the fixed weir crest to a level which allowed the most frequently occurring discharge to pass without afflux.

4. Where there is to be irrigation. however, it will probably be the case that stages higher than the most frequently occurring one will be held anyway. If this is so, then upstream aggradation must be expected, but no additional harm will result in raisiag the fixed crest slightly higher than otherwise, but this must not be overdone.

5. Fixed weir proportions are not determined by level alone. If, as normal lye the idea is to avoid d u x at formative and high discharges, then the weir must be sufTiciently wide that side constriction does not bottle up the flow. Obviously the higher the crest level, the wider must be the structure in order to avoid afnux at higher discharges. A wide structure will be expensive to build though there will be some saving in the cost of crest gates, as shallow gates will go with a high crest.

6. Flood levels on the average Malayan plain rise 2 or 3 feet above general plain level. If a weir is so proportional that there is any afflux at bank full stage, then the river flood plaii must be bunded, for it is unsafe to dissipate energy over the ground surface on the flanks of the weir. Buading is a nuisance and afflux means upstream flooding. Therefore Malayan weirs and barrages are usually so proportioned that afnux has disappeared entirely before the river rises to bank full stage. This consideration also affects the weir crest proportions.

7. Weirs which are too narrow will result in undue concentrations of

BED SEDtMENT LOAD

DlSCn ARGL

FREQUPNCY i

0 g 0 SEDIMENT LOAD

AMOUNT OF BED SEDIMENT MOVED.

FIG. I/Y/III

flow, which may cause troublesome scour or erosion. And weirs mu& in excess of normal river width cause local bed instability.

8. So it is apparent that the level and length of the fixed weir are arrived at by compromise. Usually an acceptable solution is found with a crest length approximately equal to the bed width of the river.

THE WEIR STRUCTURE

9. When fixed weir crest level has been decided, the height of the etes will follow automatically if the desired F.S.L. is known. Thus we arrive at the basic proportions of the bar which is to be put in the river, and we now have to consider the structure which will surround and sustain this bar. It is not the intention to go into details of design, but only to outline the principles involved.

10. As already seen, the function of the weir (barrages will be included generally in the following) is to raise water levels upstream. Evidently then there must be a fall of water through the structure at certain stages of flow. The fall results in conversion of potential energy to kinetic energy, and this energy must be destroyed within the solid boundaries of the structure or it will destroy the river channel downstream of the structure. If the river channel IS destroyed. the structure is likely to fail also. structurally, or as a result of undermining by piping.

11. If some upstream bed aggradation takes place as a result of the construction of the weir, there will be corresponding downstream retrogressioh. at least until normal sediment movement is once more restored. Thus, for a while, the downstream stage discharge curve of the river will be lower than previously, and this will result in release of still more energy which must be killed within the structure. Allowance may have to be made also for permanent downstream bed retrogression. resulting from removal of old brush- wood dams or other obstructions.

12. The weir is also a dam, and is subject to overturning, sliding and uplift forces. It must be stable against these under all conditions. Most river dwersion structures. in Malaya at any rate, are low, and are built on permeable foundations requirbg substantial length. There is usually no danger of sliding or overturning, and in design we do not generally have to consider these. Uplift is important. however, for unless proper provision is made, the structure, or part of it, may be lifted bodily by the pressure underneath it.

13. Provision must be made also to preveilt failure by progressive wash- out of the soil beneath the structure. This is known as "piping".

14. The surface flow aspect of design will be considered first, and then uplift pressures and piping.

SURFACE FLOW AND ENERGY DISSIPATION

15 The standard method of dissipating energy in a weir structure is by means of the hydraulic jump. If a jet of water with a hypercritical velocity strikes water of sufficient depth. a hydraulic jump is formed, with dissipatioh of energy. Assuming that all unusual turbulence is dissipated in the jump, the amount of energy dissipated (H, ) is the difference in level between total energy level upstream of the jump and total energy level downstream of it.

16. Without going into the mathematics of it. it can be stated that the basic proportions of the jump are fixed for any given set of conditions, and can be expressed by a series of curves of the form shown in Fig. 2/V/III.

17. For the jump to form properly, it is necessary that the required Er. snould be available above the floor level. It cannot be provided by way of a hole below floor level. The high velocity jet would simply persist and no proper jump would be formed.

18. In practice. the correct downstream depth for all conditions of flow cannot be obtained 011 a horiz~ntal floor. and the expedient of sloping the floor

gentlr is adopted. for reasons which become plain on consideration of Fig. 3 V 11 which illustrates the longitudinal shape of the floor of a typical low head weir, as used in Malay@.

19. Consider a small flow of intensity q, running over the weir and down the slope (or glacis) to meet tailwater at a level which will be determined by the downstream stage-discharge curve of the river. At some point on the slope (if the slope is long enough) the downstream depth will be equal to dz, the correct downstream depth for formation of the jump. The jump will therefore form. Similarly for a large discharge q,. at some point the depth will be correct. so the jump will form. Thus. provided the downstream floor is low enough. there will always be a point on the glacis at which the depth is right for formation of the jump. for any value of q and for any value of H ~ . and a jump will form.

20. The dissipation of energy in a jump formed on a slope will be satisfactory. prov~ded the slope is not too steep. though not so satisfactory as on a level floor. In the D.I.D.. the standard slope is 3 to 1. and on this slope there IS quite good energy dissipation. The slope is steep enough that the structure doer not become unduly long. yet not so steep that shuttering must be used to place the concrete in the slope.

21. The first part of the design problem is to find the highest level, and therefore the most economical level at which the downstream floor can be put. The curves in Fig. 4'V'III provide a convenient method. The following example illustrates the method of use.

12. Suppose that a discharge of intensity 50 Cusecs per foot width passes over a weir of level ~uch that the total energy level upstream of the weir is +52.2 (Upstream to!al energy level can be found of course if the weir crest level and the wecr coefficient of discharge arc known). Suppose also that the corresponding downstream total energy level. derived from the stageldischarge curve of the ncer. I \ c4S.b. Then H, the energy head lost. is 52.2- 45.6=6.6 feet. From the curve\. He see that when q is 50 and H, is 6.6. E,, is 9.4 fee: The downstream ffoor !etel must be set at least 9.4 feet below $45.6 then: tha: is to say. it must not be aboce +36.2. If the floor is not as low as this. the iump uill not form. the high velocity jet coming down the glacis will merelq shoot off the end of the floor and tear up the river bed downstream.

23. In an actual example. maximum permissible downstream floor levei h i l l be worked out for various conditions of flow. and the level selected wil: be the one appropriate to the worst case.

24. From the above. it i s evident that correct design depends on a correct tailwater discharge curve. and this is why a stage'discharge curie should always be established at a site where it is proposed to build a ueir. This curve should be adjusted to allow for swh retrogression as may occur after construction of the weir or after removal of obstruction in the river downstream. Flood levels are not likely to be affccted greatly. but low flow levels may drop by 2 or 3 or more feet. N o rule for estimation of retrogression can be given. The curve is easier to use if i t is still further modified to give total energy level against discharge. instead of stage against discharge. If there has been no measurement of stage and dischurge at the sits, the designer has to use his judgement and assume a curve. This i s not satisfactory for a structure of any size and importance.

25. Also in determining downstream floor levels, some allowance should be made for the possibility of concentrations of flow over parts of the weir. Intensities of flow should be increased by 10 per cent to 20 per cent, according to approach conditions, without benefit of correspondingly increased tailwater levels.

26. As previously explained, the D.I.D. normally proportions its plains rivers weirs in such a way that d l u x disappears before bank full stage is teached. Hence, the tailwater stageldischarge curve will rise above the weit crest stagel discharge curve at some discharge less than bank full discharge. Thus there will be no afflux for high discharge, and the most unfavourable conditions will be found at medium or even at low discharges.

27. In reality, the jump does not take place instantaneously : it is spread over a substant~al length. The water surface rises gradually from the point of formation of the jump; and while it continues to rise. there continues to be substantial turbulence. The whole of the turbulence, and hence the whole "length" of the jump. must be contained within the solid structure. Therefore ~t is necessary to know where the jump is going to form, and how long it is. Unfortunately rt does not form at the theoretical point, but at a point further up the glacis. The point of formation is determined from the curves shown in Figs. SIVIIII and 61VIIII which were experimentally derived by the D.I.D. The method of their use is most easily explained by means of an example. Suppose that 30 cusecs per foot width flows down a glacis with a slope of 3 to. 1. and meets tailwater at a level of +20.0, downstream horizontal floor level bemg \ a at -I- 11.0. Suppose also that H,, the difference between upstream tota! energy level and downstream water level, is 3.0 feet. First, from Fig. 5 'V/III. ~t is seen that for 30 cusecs per foot width, E, is 4.56 feet, where Et, is the total energy at critical depth. Now refer to Fig. 6'VIIII. It will be seen that before the curves can be used. the value of E,, H L I must be found. In this case it is 4.56 3.0 = 1 52. Also D2 H L, must be found. and in this case it is 9.0 3.0 = 3.0 Usmg the curves then. it is seen that D, H,, = 0.56. and therefore D, = 1.68. D, is the depth of the trough of the jump below downstream water level. Hence the level of the bottom of the trough is 20.0 - 1.68= + 18.32. The bottom of the trough also lies on the surface of the high velocity jet coming down the glaas, so its location is at the point where this surface cuts a line at level + 18.32. The surface profile of the jet running down the glacis. is determined by the use of standard "energy of flow" curves, as in Fig. 7lVIIII.

28. In the interests of accuracy, the best possible estimate should be made of H,, Thic mvolvec accurate determination of upstream total energy level. For the type of crest normally used in D.I.D. structures, the value of h, the height of total energy level above weir crest. can be read off for any required intensit!, of discharge. for crests of various widths. from Fig. 81VIIII.

29. The length of the jump can be taken as 5.6 times the depth from downstream water surface to downstream horizontal floor level. There is substantial turbulence over th~s length. which must therefore be contained within the structure and must not be allowed to extend over the unprotected downstream river bed. The horizontal floor of the structure must extend at least to the end of the jump Furthermore it must be impermeable up to this . If it is not impermeable, a pressure gradient will be induced in the foun d- atlon material. as a result of the varying static head at different parts of the jump: foundation material may then be lifted through the floor, and failure may occur.

30. The rule here given for limit of downstream floor is illustrated in Fig. 9/V/III. which also illustrates a rule relating to location of wing walls. This rule will be discussed presehtly.

31. For practical purposes, the shape of the jump profile can be taken as a straight line joining the point of formation of the jump to the end of the jump on the downstream water surface. This also is shown in Fig. 9/V/IlI.

ENERGY OF FLOW CURVES

1 2 3 4 5 10 25

F IG 8

MINIMUM DISTANCE AT WHICH SLOIIWG WOE W4LL MAV CUT DCWUSTRCAM WATER SURFACE LEVEL - 4.5 0 2

PIAGRAM I L L U I T R A T I N ~ THE CONSTRUCTION FOR THE PROFILE FOR THE JUMP A N 0 FOR STANDARD LEHGTN OF IMPQRMEABLE FLOOR AND WSlTlON OF WINO WALL.

F I G . 9/~/11l

32. As will be seen later, the point of formation of the jump and the jump profile are essential also in determining unbalanced uplift pressures.

33. It might be expected that small values of H, (or H,,) would result in a small tendeacy to scour. In fact this is not so. For a given intensity of flow, the greater is HL, the greater must be the tailwater depth to accommodate the jump. Consequently the lower the velocity of the tailwater off the end of the structure: and vice versa. D.I.D. research has shown that undesirable scour will not take place at the end of the structure if the Froude number. F at the end of the structure does not exceed about 0.3. That is to say v does not exceed about 0.3. Therefore, not only must the down-

4 3 stream floor level be low enough to a~~~rnmoda te the jump for all conditions of flow: it should also be low enough to ensure that the safe value of F io not exceeded at any stage of flow. If for any reason the floor canaot be put low enough to give the safe value of F, then spacial bed protection must be provided ia the form of a loose rubble apron. This problem docs not often arise in Malaya. In cases where it does, however, reference should be made to the Chapter on "Canals, Channels and Rivers", by Blench. in "Civil Engineering Reference Book", Butterworth, or to other standard worka which treat this subject.

34. Even if the dowhstream floor is low enough, there will be a tendency to formation of undesirable vortices, where the water passes off the end of the downstream floor. For this reason, a "Rehbock sill" is used at the end of D.I.D. structures. This is illustrated in Fig. 10/V/lII. Note the double height tooth. adjacent to the wall. Note also that the vertical face of the teeth is upstream, not downstream. The tooth height is not critical and may be about one tenth of the downstream depth of water at the stage at which there is maximum energy dissipation. That is to say, when qXH, is a maximum.

35. Downstream apron floor blocks are used sometimes as aa aid ta shortening the jump. Their size and position is empirical, but the proportions shown in Fig. I I /V/III are known to work quite well.

36. The shape and location of side walls at the downstream end of weir structures has come ib for a good deal of attention in the D.I.D. Research has shown that straight side walls. sloping down at 23 to 1 or 3 to 1 are satisfactory and economical. Further. a splay of 1 in 10 on each side is beneficial, as this stops the play of secondary vortices over erodible bank materials adjacent to the side walls. The general arrangement is shown in Fig. 12/V/In. The location of the side wall slope in relation to the hydraulic jumps profile is shown in Fig. 9/V/III. It will be seen that the surface slope is allowed to cut the wing wall at a distance 4.5 4 downstream from the point of formation of the jump. This results in a lazy vortex, which is not objectionable and may be beneficial to the sides of the walls. Various conditions of flow will be checked, of course, and the worst will be taka.

37. The form of outlet is non-scouring provided that an "onion" of adequate size is formed in the downstream channel at its junction with the structure. This is shown in Fig. 13/V/m. It is emphasised that this onion is a vital part of the design. In spite of repeated reminders, engineers go on putting in weirs without onions and then go to endless trouble in attempt- ing to stop erosion which must and will occur if the onions are omitted. ln this comexion, see also D.I.D. Technical Circular No. 1. "Outlets from Structures".

38. To produce a hydraulically satisfactory job, the sloping side walls must continue till they cut the river bed. Structurally this is not satisfactory.

F I G . 10/~/111

SECTION

WALL ,

PART PLAN

F I G . 11/V/111

PART PLAN

however, unless the side wall extension beyond the end of the floor slab is structurally separate from the more upstream point of the wall. Consequently it is the practice at present to use sheet piling with a sloping capping beam to continue the falling side wall line beyond the end of the floor slab. The side walls are returned at right angles, and are continuous with the drop wall or beam at the end of the apron, thus forming a U beam which gives rigidity to the end of the structuie. The arrangement is illustrated in - ~ i ~ . 14/V/III.

39. Before leaving the subject of downstream end design for structures of this type, it might be as well to point out that there are many "trick" devices for shortening and cheapening structures containing hydraulic jumps. Careful study shows, however, that none of these devices is effective for

I \

I v ' jumps occurring at low Froude numbers. .( ! . such as are normal in D.I.D.

I Jgd I work. This comment goes for all manner of cunningly shaped buckets and teeth. Their place is in structures containing jumps occurring at high Froude Numbers, where they can be very effective. For detailed and comprehensive information on these types. refer to Papers Nos. 1401 to 1406 (October 1957) of the American Society of Civil Engineers.

40. Upstream end design requires consideration also. If the upstream end of the floor of the structure is at weir crest level, there will be drawdown of the water surface and increase in normal velocity over the river bed upstream of the structure This will result in bed scour and exposure of the upstream end of the floor. This exposure will cause additional turbulence. and a considerable scour hole will develop immediately upstream of the structure.

41. Again, unless the side walls at entry are streamlined, turbulence. which results in bed and bank scour, will be generated. Also the effective width of the structure will be reduced as a result of end contractions.

42. Therefore. the upstream floor is ramped up to the weir crest, and the side wails at entry are flared. The proportions shown in Fig. IS/V/III have been found, by experiment, to be satisfactory. It has been found hydraulically unnecessary to catry the curve of the flare beyond the point shown; and to do so would only result in additional cost and structural complication.

PERCOLATION P R E S S U R E S . UPLIFT AND EXIT GRADIENTS

43. The normal D.I.D. weir is built in an alluvial valley and is constructed on sediments which have been deposited' by water and which are more or less permeable. Hence water percolates underneath the structure from the upstream end to the downstream end. driven by the difference in head of water at the two ends. As the water passes beneath the structure, it loses pressure. The pressure at entry to the soil beneath the structure is equal to the upstream head of water, and the pressure at exit is equal to the downstream head of water. Between these two points (or areas) the percolation pressure becomes less at a rate which is not constant, but depends on the shape of the bottom of the structure and on variations in the permeability of the soil. The present object is to explain the principles involved, so it will be assumed for the time being that the permeability of the soil is uniform.

44. The percolation pressure in the soil in contact with the base of the structure acts upwards on the structure and tends to float it off. This tendency has to be resisted by the weight of the structure plus the weight of any water contained inside the structure. Sufficient weight must be built into the structure to hold i t down against the uplift forces. Hence the importance of knowing the percolation pressure distribution.

\ \ - RIVER BED

R.C. PILE LINE AND GAPPINQ -

RECOMMENDED PROPORTIONS

FOR INLET FLARES AND

UPSTREAM FLOOR.

I - SLOPE 3 TO I .

45. Flow of water through permeable soils is according to Darcy's Law

L where V = velocity

H = head L = length of path of flow. K is the "transmission constant"

Thus the flow pattern and the pressure beneath a structure on a permeable foundation can be represented by a flow net, consisting of streamlines and equipotential lines.

46. There are a number of ways of preparing flow nets, e.g. sketching. mathematical, electrical analogy and viscous fluid. There is plenty of standard literature on this subject, and method of preparation of nets will not be discussed further here. The D.I.D. has equipment for preparation of nets by the electrical analogy method, using Teledeltos paper.

47. Sketching and study of flow nets brings out the influence of arrangement and dimensions of the common elements of the ' undersides of weirs. That is to say, floors and pile lines or drop walls. A few cases are examined in the following paragraphs.

0 48. Fig. 16/V/III shows a hypothetical structure which has length,

n o depth. From the flow net it will be seen that there is a concentratioMof lines in the immediate vicinity of the ends of the floor, indicating a igh pressure gradient in these regions. Except near the ends the stream-lines $ld equipotential lines are fairly widely and evenly spaced, indicating a relati*ely low pressure gradient and one which is not changing rapidly. The preslure distribution is as sketched. It will be noted that the percolation pressure diagram starts at upstream water level and finishes at downstream water level.

49. The pressure gradient of the percolating water emerging at the downstream end of the structure is called the "exit gradient". With a high con- ceiltration of flow-net lines at "exit". as in this case, there must be a high exit gradient. High exit gradients are very objectionable because they lift the unsupported surface soil, thus uncovering more soil at a lower level, and this in turn lifts. Thus a "pipe" develops under the structure by progressive action, and the structure will eventually collapse when the pipe gets big enough. "Safe" exit gradients will be considered a little later.

50. The exit gradient at the end of a structure .with no depth, as in the case now being considered, is very high (in theory it is infinite) and the chances of such a structure standing for long are negligible. Nobody with any appreciation of the problem will design a structure like this.

51. Fig. 17!V 111 sketches a structure with depth but no length, consisting of a single pile line. It will be seen that the concentration of flow net lines is round the base of the pile line and the net is well spread at entry and exit. Thus the pressure gradient is high where the soil is well supported, and the exit gradient is low. As the structure has no length, an uplift diagram cannot be drawn. A pile line is thus a satisfactory structure so far as exit gradient and uplift is concerned. But unless well supported it has unsatis- factory resistance t o overturning. And if water passed over its top, the bed on the downstream side would be eroded away rapidly. Thus a single pile line would not make a satisfactory weir structure. As a water stop in a structure such as a coffer dam, it is of course excellent.

52. Fig. 18/VIIII shows a structure with length and with an upstream pile line. It will be seen that the pile line has the effect of reducing the

TOP OF OATL OR W E I R CRfSt X = UPLIFT PRESSURE

A T POINT A. P E R C O L A T I O N PRESSURE LINE

pOWNSTREkU . W.L.

WEIR WlTH LENGTH BUT NO DEPTH

U.S.W.L. 3 - --- - --

PILE LINE WITH DEPTH ONLY.

F I G . 1 7 / ~ / 1 ~

J 5 . W L PERCOLATIOM PRESSURE LINE +-, - ,

,WEIR WITH F L O O R AND UPSTREAM P I L E L I N E

WElR WlTH FLOOR AND DOWNSTREAM PlLE L l N E

stvlrtr PERCOLATION PRESSURE L I N E

Y' , , j-- I - ' \X 1. / I

WElR WlTH FLOOR AND PlLE LlNE A T E A C H END

F I 6 . 2 O/v/lJJ

percolation pressures beneath the structure and thus it is beneficial. The structure can be made lighter and cheaper. There is. however, a concentration of flow lines round the downstream end of the floor, indicating a high and dangerous exit gradient.

53. Fig. I9/V/III shows a structure with length and with a downstream pile line. The exif gradient is now satisfactory, but the uplift has been greatly increased.

54. Fig. 20/V/III shows a structure with a pile line at each end. The downstream pile line reduces the exit gradient but increases the uplift. The upstream pile line reduces the uplift. This is the usual compromise solution.

55. By similar reasoning, it will be seen that a pile placed under the centre of the structure will not have a very marked effect, unless it is very deep.

56. A further point to note and which is important, is that vertical surfaces (terminal ones at any rate) are far more important in changing the flow pattern and pressure distribution than horizontal surfaces.

57. Summarising then, it can be said that in permeable soils of indefinite depth, a downstream drop wall is essential to ensure a reasonably low exit gradient. If it is too long, however, uplift pressures will be greatly increased. Upstream pile lines reduce uplift pressures, but do not have much effect oh exit gradients. The downstream pile line should not be deeper then than is necessary to ensure a satisfactory exit gradient, and the upstream pile line should be as deep as economically possibIe in order to reduce percolation pressures. and thus to lighten and cheapen the main part of the structure.

58. While percolation pressure distribution is best studied by means of flow nets. it will be more convenient in practice, in many cases, to use the curves given in Fig. 21 /V/III to determine the percolation pressures inside the pile lines at the ends of the structure. The curves are self explanatory. Between the two points at which the pressures are thus determined, the pressure distribution on the underside of the structure can be represented, with sufficient accuracy. by a straight line. These curves are based on Khosla's "Method of Ihdependent Variables" and apply to the case where there are pile lines of equal depth at each end of the structure, Khosla's method embraces more complicated cases also, and curves for solution of these may be found in modern standard reference books.

59. A favourable condition arises if the upstream pile line can pierce an impermeable stratum; say clay. Subsoil flow is then cut off, and theoretically there is neither uplift pressure nor exit gradieat.

60. If on the other hand the downstream pile line pierces clay and the upstream does not, full upstream head pressure will develop under the whole of the underside of the structure, and this may easily float off, unless it is very heavy indeed, or unless the pressure under the structure is artificially relieved. This is not a recommended procedure.

61. The necessity for a low exit gradient has been mentioned. Let us examine this a little more closely. Consider the equilibrium of the bed material in a length L at the end of a "stream tube" emerging at the tail end of a structure.-Fig. 22/V/III. Let the cross sectional area of this short length L bed A. Also, %r simplicity, aBsume that the bed is horizontal so that the flow emerges vertically. Then the force pushing out the plug of bed material in the stream tube is dp XLXdA

7

dL

LENGTH OF FLOOR OF Oc = DEPTH OF PILE

c RIVER BED

ST REAM TUBE

where dp is the pressure gradient at exit. --

dL The force preventing the plug from blowing out is the weight in water of the solid material in the plug. For normal bed sands with normal pore space, the submerged weight of the sand in the plug is about equal to the weight of an equal volume of water. It follows that for the plug to be in equilibrium. the head of water on its underside must be about equal to the length of the plug. That is to say. the pressure gradient. dp, must be about unity.

dL

62. Clearly a factor of safety must be applied, for conditions are not entirely stable. For example, pressure gradients can be temporarily upset by local surging in the river or by sudden operation of gates. The following factors of safetv have been found appropriate.

Shingle . . . . . . . . . 4 t o 5 . . . . . . Coarse sand . . . S t 0 6

Fine sand . . . . . . 6 t o 7 Clayey soil . . . . . . . 3 t o 5

Submerged densities do vary considerably, however. according to the material and the pore space: so in unusual cases, the factor of safety should be applied. not to an exit gradient of unity. but to the properly determined "flntation gradient" for the material of which the bed is composed.

63. So far as exit gradients are concerned, the practical design problem will be to ensure that the exit gradient is kept sufficiently low. The length of structure will have been determined from surface flow considerations. The exit gradient will be reduced then to the required value. by provision of a downstream pile line or drop wall of sufficient depth. This is a matter of trial and error. Some depth will be assumed, and exit gradient for that depth will be determined from the flow net. or perhaps more conveniently from the standard curves shown in Fig. 23/VIITI.

64. As has been shown, the water percolating under the structure is under pressure and this pressure exerts an upward force on the bottom of the structure. Thus there is a force trying to lift up the structure. This upward force is resisted by the weight of the structure and the water inside the structure, which try to push it down. Obviously the downward forces must exceed the upward ones or the structure will float away. Usually. however. the percolation forces under some parts of the structure exceed the weight of the water in the structure together with the weight of concrete necessary for purely structural purposes. It is necessary then to add weight to the structure to produce a favourahle balance. As a weir is not a particularly rigid form of structure. ~t i s not assured to act as a rigid whole: the additional weight required is put where it is needed.

65. The tirst step is to draw out. for the various conditions of flow. the water profiles and the corresponding percolation pressure diagrams. From previous considerations. we shall already know the longitudinal section of the floor surface and the depth of the terminal pile lines. so all this can be done. The unbalanced hydraulic uplift pressure then at any point is the difference in level between the percolation pressure at that point and the water surface level In the structure This difference has to be balailced by the weight of concrete In the floor of the structure. taken at its submerged weight, assisted by the weight of adjnccnt piers and walls taken at submerged or unsubmerged weight as appropriate.

66. By wav of example, consider the weir in Fig. 24alV:III. which is retaining water, without thrnu~h flow. by means of a crest gate. The un-

25

CALCULATED FOR

20

15

\O

5

0 10 20 30 40 50

V A L U E S O F OC

?s LENGTH OF FLOOR

DEPTH OF PILE

NO SURFACE FLOW

FIG. 2 4 d / ~ / n l

PERCOLATION PRLSWRE LINE

WITH SURFACE FLOW

blanced hydraulic uplift 1s shown vertically hatched. and this must be balanced by the submerged weight of concrete. Thus, at section A-A, the unbalanced upMt is h ft. of water. That is to say 624Xh lbs.lft.2. If the weight of concrete in air is taken as 144 Ibs./ft.? its submerged weight is 81.6 Ibs./ft.'

,The thickness of concrete at taction A-A must therefore be 62.4 h feet. Put - 81.6

m a slightly different way, it needs 314 foot thickness of submerged concrete to balance 1 foot of hydraulic uplift.

67. Fig. 24b/V/III illustrates a case where water is flowing over the weir. Thc water surface profile will have been obtained as previously described. It will be seen that the uplift pressure diagram is different from the case for no flow. According to circumstances. either the flow condition. or the no flow condition, can be more severe, so both should be checked. There is of course a variety of cases for the flow condition, and enough cases should be checked to ensure that the worst for every cross section has been allowed for.

68. The thickness of the floor should not slavishly follow the requirements of uplift. Provided there is adequate stiffness in the floor slab. the mass can bo dispersed within reason in a structurally satisfactory manner. As noted. might of walls and piers can also be taken into account: and it is worth nnting that concrete above water level can and should be taken at its full weight. It is therefore more economical of material than additional weisht In the floor

69. Adjustment of floor thickness will of course affect the flow net slightly. and hence the pressure distribution. It is usual therefore to assume a reasonable floor thickness. based on experience. when drawing the flow net. Minor changes in floor thickness will then have no practical effect on the flow net.

70. The factor of safety assumed for uplift is generally small or unity. €or the following reasons :

(a) Gravity is not in doubt. (h ) The tendency of the structure to overturn throws additional weight

where weight is required. i.e. at the downstream end. (c) There is considerable friction between side walls and backfilling.

and this would come into play if the structure tried to lift. fd) The soil at the upstream end of the structure tends to become

less permeable as a result of inwash and lodging of fine sediment carried in the river water. The downstream pressures thus tend to be lower than estimated.

71. So far we have considered the weir to be on a theoretically perfect and pcrmetrble foundation. Usually foundations are not perfect. At least they are likely to be more permeable horizontally than vertically. This can be dowad for by distortion of the flow net. Reference on this point should be made to standard literature. Vcry ofien the ground is stratified. and then good judgement. based on appreciation of the principles involved. is necessary Fig. 2SIVIIII illustrates a simple example. The uplift pressure at B will be as high as that at A.

72. In a case like that illustrated in Fig. 25/V/II1. pressure to B could be cut off by an intermediate pile line near A. and to guard against leakage through this pile line. a "reversed filter" could be provided at B for pressure (if any) relief. The use of filters under impermeable structures on permeable foundations is not really recommended. however They should not be used except in consultation with the A.D. Planning. If filters must be used. the following points must be noted

F I G . 2 5 / ~ / l l l

REVERSE FLOW WILL EMERGE IN THIS REGION. L I F T I N G SAND TWOUGH THE INTERSTICES IN THE RUBBLE. THE RUBBLE WILL

3uBslDE AND COLLAPSE OF T ~ E WEIR WILL HAVE STARTED.

BBLE LARGE I WEAD

- ..u R E V E R ~ E FLOW IN SUBSOIL S E T OF BY DIFFERENTIAL - PRESSURES.

F I G . 26/~/111

( a ) The filter must be capable of passlng, wlthout appreciable loss of head. the water reaching it Otherwise there will be a pressure build up

( k ) Filters placed under an area over whrch water surface level is not constant. should have only one t d e t Otherw~se there may develop a circulation through the filter. and the filter my become choked w~th fine sediment

fc, Filter materials should be graded according t o D 1.D Technical Circular No. 2. which is based on standard workb on the subject

73 Wem may also be constructed on lmpermeable sods. such as clays I'he determ~nat~on of uplift 1s then vet! pmblemat~c in fact estimates are yulte arbitrary It 15 suggested that 11 is safe to assume that uplift pressures will not exceed 5 0 per cent of the uplift pressures which would be expected < I n permeable soils. prcwded end cwt offs are good and prov~ded there is an adequate filter for lnterceptlon of water percolating along the junct~on between the concrete floor and the ground below ~t Somet~mes the subso11 draws slightly away from the slab. leaving a path along wh~ch water can move relatrvelq freely Thls IS known as "roofing" Theoretically. cuch a filter cannot he designed, as the clay part~cles wh~ch 11 would be expected to retam are too fine The filter should really be regarded as a temporary safeguard and as a "tell-tale" of impending trouble If. one day. the filter were found t t ) be d~scharging. a welr grouting operation would be rnd~cated

74 It has been explained In a previous Sect~on that the most important fac~or in keepmg sediment out of a canal IS to slte the canal Intake correctly on the outs~de of a bend In the rlver This 1s agaln stressed here It is hetter t o take clean water stralght away than ~t 1s to take dirtv water and to attempt to clean ~t

75 There are wcaslons. however, where rntake of some sediment 1s un- abo~dable In such cases specla1 devices for getting rid of 11 may k worth while 4 n "excluder" may be incorporated m the Intake structure w h ~ h reiects sed~ment before 11 enter\ the canal OT an "elector" may be placed In the canal to eject .;ediment hndlng ~ t s way past the head regulator A typ~cal tvpe of silt excluder used In the Bernam River Headworks Intake is \hewn In Fig 27 C 111 The Intake can extract up to 857, of the total dry weather floa without drawing rn bed sediment by operatmg the excluder dev~ce wlth the remarnine 15r,, of rlver flow

7h The workmg principle of excluders and ejectors 1s that the sediment rs persuaded to become concentrated In a very small part t t f the flow, and that pan of the flow 1s then rejected together with the sed~ment ('oncen- tratlon may be achieved bv settlement In smoothed flow. or by curvature of flow (the same pnnciple as rn good river Intakes) Some devlces are highly refined They all have the d~sadvantage that substant~al head must be expended In rejectmg the sediment laden water. and this 1s not always possible

77 Another way of dealing w~th excess sediment 1s to prov~de an over- \ize head reach for the canal. and to dredge out the deposits as required.

78 The popular raised c~ll 1s. In itself, almost useless as a sediment control device About the only thlng 11 wlll keep out is stones and gravel. I t is repeated that the real solutlon lies in correct curvature of flow in the rnver. at intake.

APPLICABILITY OF PRINCIPLES TO OTHER TYPES OF HYDRACII.I( STRIKTURF

79 Finally, ~t IS noted that the principles which have been explained apply to all types of structure m which there is surface or sub-soil flow.

Consider for example a gently sloping rubble weir, which. in essence, is a permeable heir. I f the surface is made too smooth. a hydraulic jump will form and a reverse flow will be set up In the subsoil. Water will enter the subsoil, through the rubble of the weir. under the crest of the jump wave &here pressure is high, and it will merge again in the trough of the jump. Probablj it will bring up sand from underneath the rubble. and the rubble wftl collapse into the hole thus formed. See Fig. 26/V 111. If the rubble of the weir has been roughly bound with a surface layer of mortar. to "hold the stones". it is quits: likely that sufficient pressure will be developed under the troush of the wave to lift out ;1 whole ~ection of paving. All proposals, whether for concrete. rubble. wood. or any other sort of hydraulic structure. should be examined on these lines.

4ff lux . . . . . . . . . 4ggradation . . . . . . Bank dimensions . . . . . .

. . . . . . .. full Stage Batas . . . . . . ....... E M factor . . . . . . . . Berms . . . . . . . . . Bifurcation Control . . Canal Branch and Main ... . Capacity . . . . . .

. . . . . . .. Controls . . . . .

. . . . . . .. Design . . . . . . . Linings . . . . . .

Collapse of Weits . . . . . . . . . . . . . Command . . . . . . . . ..

Controls . . . . . . . . . . . . . . . . . . I

Bifurcation .. ... . Overshot ... . Undershot ...

Controlled Drainage Schemes . . . Crest . . . . . . . . . . . . Crops other than rice. Irrigation of ... Cross Bunds . . . . . . . . . .. Section . . . . . . .

. . . . . . . . Darcy's Law Demand. maximum . . . . . . Design of weim . . . . . . .. Canals . . . . . . . . . Dimensions of Banks . . . . . Discharge . . . . . . . . . . . . . Hydrograph . . . . . . Distributaries . . . . . . . . . Distribution Systems Narrow Valley

n ,. Flat Plains .. Gentle Slopes Drainage Systems ... Dry weather flow ...

. . . . . . Duty Energy dissipation Evaporation ...

... n

Exit Gradients ... Falls, notched ... Field Channels ...

... n n

Filters rrtrerse ... noor blocks ... .. levels ... .. t h i c w ... Flow naets . . . . . . Formative Stgp ... Froude nember ... Gradicnt . . . . . Head Regulator ...

Headworks

Hydraulic Jump Intakes lrrigatlon Continuous ...

.. Crops other than rice

.. Definition ot .. Systems .. Flat Pla~ns . .. Gentle slopes .. Narrow Valley .. Pahang Paya

.. Rotational

Jump Hydraulic .. Length of

Leakage Length of Jump Maximum Demand Method of independent variables Module Notched falls Nurser~es Offtakes "Onions" ... Overshot Controls Pahang Paya System Percolation

.. Gradients

.. Pressures .. diagrams

Piping Pumping Schemes Regime Rehbock sills Reservoir Schemes Retrogression Reverse filters R ~ c e cultivation Roofing Rubble weirs Sediment

.. Extraction and Exclusion Seepage Side factor .. walls downstream . . . . upstream

Slope . channel Stage . bank full StageDischarge Surface flow Transpiration .

Undershot Controls . Uplift Velocity . Water requirement ...

Weirs Crest . Design of .. Structures ..

W~dth . channel .

Section

1 111 111 I1 I I I I 1 I I I I I1 111 111 I I 111 I

111 I1 11 I

I I 111 I l I I I1 11 IT1 111 111 1 11 111 I

Ill 111 I

111 111 I1 111 111 I1 I1 IT1 11 1 11 111 111 111 1 11 I1 111 I I C

111 111 111 I1

Paragraphs

18 . 21 1-79 16-39 43-46 27 50.58 1 17-25 39-41 4247 31-38 49 27-29 25 16-39 27 . 29 26 27 . 29 10-14 58 42 34 3 6 37 34. 40 49 7-9 26 29 43-73 65 . 66 13 13 . 24 11-22 34 I4 . 25 I1 . 2 4 72 . 73 2.4 . 2 6 73 79 12-2 1 3 74-78 29 16 36-39 40-42 11.16 . 3 3 6 . 26 24-26 15-42 6 . 7 . 9 26 36 . 40 . 42 12 . 43-73 12.16 . 2 0 8

3 1-19 7-14 12~21

"Calibration of Junction Boxes", D. 1 .D. Research klemorandum No. 18:

Chow, V.1. 119691, "Open Channel Hydraulics".

"Des~gn Manual for Water Conveyance Svstems" 11980). Division of Drainage and Irrigation, Kuala l umpur.

"Des~gn ut \ma1 1 Canal Structure" ( 1976) USBR

"Desrgn of i.owHead Hydraulic structures" (1973) United Nations, Water Rescl~r(.es Series No. 45.

D.L.1). Information Paper No. 2 "Presaturation of Padi Fieds".

Feasibility Report on Tertiary Lrrigation Facilities for Lntensive Agricultural Development in the Muda Irrigat~on Scheme (1977) MADA.

Henderson, F . M . (1966) "Open Channel Flow".

"Irrigat~on Development and Management" (1980) Proceeding of ADB Regional Seminar on Irrigat~on Development and %anagement .

Michel, A.M.,(1978) "Irrigation Theory and Practice".

Thavaraj , S.H. - "The Necessity of Terminal Facilities for Water Management at the Farm Level" - Bulletin No. 139. Proceedings of the National Seminar on Water Management and Control at Farm Level.

Drainage and Irrigation Department - Manual

Documents