Cover Image
close this book Local Experience With Micro-Hydro Technology
View the document Preface and Acknowledgment
View the document Foreword to the 3rd edition
View the document Abstract
close this folder A. Introduction
close this folder B. Development of hydropower resources
View the document 4. BIG OR SMALL HYDRO?
close this folder C. Small hydropower in the rural situation
View the document 1. PAST AND RECENT HISTORY
close this folder D. A practicable approach
View the document 2. TECHNOLOGY
close this folder E. PROJECT EXAMPLES
View the document 1. BASIC APPROACH a) Cost-Benefit-Approach for Socio-Economic Selection
View the document 2. FINANCE
close this folder ANNEXES


a) Water Turbines

b) Other Equipment

c) Survey and Civil Engineering

A discussion of implementation and operation-specific questions remains theoretical as long as no approach of a workable technology is developed that has scope for cost-reduction and self-reliance. Realising this, the Nepal Industrial Development Corporation, together with the Swiss Technical Cooperation program, undertook the development of locally made water turbines in their joint venture, a medium-sized metal workshop, Balaju Yantra Shala (BYS). In the initial stage, a number of Propeller (low-head) turbines were manufactured and installed, mainly for direct power drives. An assessement of the performance of these machines after somer years, led to the conclusion that a more versatile turbine was needed in terms of output capacity and head range. Consequently, a Cross-Flow (Michell-Banki) turbine was developed that combined ease of manufacture with considerable adaptability to different situations. This turbine met with a growing interest and a few other workshops started manufacturing small numbers of turbines, adopting the

Cross-Flow principle. Almost simultaneously a number of other countries began to concentrate on the Cross-Flow turbine for small projects, notably Thailand, Indonesia, Pakistan and Peru. Basic principles and state of the art of the technology involved including other components that are required -as well as how it compares to existing conventional technology -shall be discussed here.

a) Water Turbines

In water turbines the kinetic energy of flowing or falling water is converted into mechanical rotary motion. As noted earlier, theoretical power is determined by head and mass flow rate. To calculate available power, head losses due to friction of flow in conduits and the conversion efficiency of machines employed must also be considered. The formula, thus, is the following:

P(kW) = Hn × Q × g × p × ntot = Hn×Q×ntot×9.81 where: P = Output power in kilo Watts (10³ W)

Hn = Net head = Gross head -losses (m)

Q = Flow in m³/second g = Specific gravity ~ 9.81 m/s² p = Density (for water ~ 1000 kg/m³) ntot = Overall efficiency =n1×n2×...×nn

For small outputs of interest here, and as a first approximation, the formula can be simplified:

P(kw)=(Hn(m)*Q (l/s.))/200 where Q is in liters per second and an overall efficiency of 51 % is implied. The "rule of thumb" calculation is therefore on the conservative side.

The oldest form of "water turbine" is the water-wheel. The natural head -difference in water level -of a stream is utilised to drive it. In its conventional form the water-wheel is made of wood and is provided with buckets or vanes round the periphery. The water thrusts against these, causing the wheel to rotate.


The principle of the old water-wheel is embodied in the modern Pelton wheel(There are water-wheels working on different principles. The statement above applies only to the impingement-type of water-wheel) which consists of a wheel provided with spoon-shaped buckets round the periphery (fig. 8). A high-velocity jet of water emerging from a nozzle impinges on the buckets and sets the wheel in motion (fig. 9). The speed of rotation is determined by the flow rate and the velocity of the water; it is controlled by means of a needle in the nozzle (the turbine operates most efficiently when the wheel rotates at half the velocity of the jet). If the load on the wheel suddenly decreases, the jet deflectors partially divert the jet issuing from the nozzle until the jet needle has appropriately reduced the flow (fig. 10). This arrangement is necessary because if in the event of sudden load decrease the jet needle were closed suddenly, the flow of water would be reduced too abruptly, causing harmful "water hammer" phenomena in the water system. In most cases the control of the deflector is linked to an electric generator. A.Pelton wheel is used in cases where large heads of water are available (Section on Pelton turbines: From How Things Work, The Universal Encyclopedia of Machines, by arrangement with Bibliographisches Institut AG, Mannheim)

Fig. 9: Jet Impinging on Bucket

Fig. 10: Operation of Jet Deflector and Needle

Pelton turbines belong to the group of the impulse (or free-jet) turbines, where the available head is converted to kinetic energy at atmospheric pressure. Power is extracted from the high velocity jet of water when it strikes the cups of the rotor. This turbine type is normally applied in the high head range (>40 m). From the design point of view, adaptability exists for different flow and head. Pelton turbines can be equipped with one, two, or more nozzles for higher output (see fig. 11). In manufacture, casting is commonly used for the rotor, materials being brass or steel. This necessitates an appropriate industrial infrastructure.

Fig. 11: Schematic of 2 Nozzle Pelton-Wheel

Source: (Fig. 9, 10, 11) How Things Work, p 49


In the great majority of cases (large and small water flow rates and heads) the type of turbine employed is the Francis or radialflow turbine. The significant difference in relation to the Pelton wheel is that Francis (and Kaplan) turbines are of the reaction type, where the runner is completely submerged in water, and both the pressure and the velocity of water decrease from inlet to outlet. The water first enters the volute, which is an annular channel surrounding the runner, and then flows between the fixed guide vanes, which give the water the optimum direction of flow. It then enters the runner and flows radially through the latter, i.e., towards the centre. The runner is provided with curved vanes upon which the water impinges. The guide vanes are so arranged that the energy of the water is largely converted into rotary motion and is not consumed by eddies and other undesirable flow phenomena causing energy losses. The guide vanes are usually adjustable so as to provide a degree of adaptability to variations in the water flow rate and in the load of the turbine.

The guide vanes in the Francis turbine are the elements that direct the flow of the water, just as the nozzle of the Pelton wheel does. The water is discharged through an outlet from the centre of the turbine. A typical Francis runner is shown in fig. 12. The volute, guide vanes and runner are shown schematically in fig. 13 and the diversion of the water at right-angles to its direction of entry is clearly indicated in fig. 14, which is a cross-section through the turbine.

Fig. 12 Francisc Runner

Fig. 13: Schematic of Flow in Francis Turbine

Fig. 14: Cross Section through Francis Turbine

In design and manufacture, Francis turbines are much more complex than Pelton turbines, requiring a specific design for each head/flow condition to obtain optimum efficiency. Runner and housing are usually cast, on large units welded housings, or cast in concrete at site, are common. With a big variety of designs, a large head range from about 30 m up to 700 m of head can be covered.

For very low heads and high flow rates -e.g. at barrages in rivers -a different type of turbine, the Kaplan or Propeller turbine is usually employed. In the Kaplan turbine the water flows through the propeller and sets the latter in rotation. The water enters the turbine laterally (fig. 15), is deflected by the guide vanes, and flows axially through the propeller. For this reason, these machines are referred to as axial-flow turbines. The flow rate of the water through the turbine can be controlled by varying the distance between the guide vanes; the pitch of the propeller blades must then also be appropriately adjusted (fig. 16). Each setting of the guide vanes corresponds to one particular setting of the propeller blades in order to obtain high efficiency.

Fig. 15: Kaplan Turbine (schematic)

Source: (fig. 13, 14, 15+16) How Things Work, p 51

Fig. 16: Propeller of Kaplan Turbine

Specially in smaller units, either only vane adjustment or runner blade adjustment is common to reduce sophistication but this affects part load efficiency. Kaplan and Propeller turbines also come in a variety of designs. Their application is limited to heads from 1 m to about 30 m. Under such conditions, a relatively larger flow as compared to high head turbines is required for a given output. These turbines therefore are comparatively larger. Manufacture of small Propeller turbines is possible in welded construction without the need for casting facilities.


The concept of the Cross-Flow turbine -although much less well-known than the three big names Pelton, Francis and Kaplan -is not new. It was invented by an engineer named Michell who obtained a patent for it in 1903. Quite independently, a Hungarian professor with the name Donat Banki, re-invented the turbine again at the university of Budapest. By 1920 it was quite well known in Europe, through a series of publications. There is one single company who produces this turbine since decades, the firm Ossberger in Bavaria, Germany. More than 7'000 such turbines are installed worldwide, most of them made by Ossberger.

The main characteristic of the Cross-Flow turbine is the water jet of rectangular cross-section which passes twice through the rotor blades -arranged at the periphery of the cylindrical rotor - perpendicular to the rotor shaft. The water flows through the blading first from the periphery towards the centre (refer to fig. 17), and then, after crossing the open space inside the runner, from the inside outwards. Energy conversion takes place twice; first upon inpingerment of water on the blades upon entry, and then when water strikes the blades upon exit from the runner. The use of two working stages provides no particular advantage except that it is a very effective and simple means of discharging the water from the runner.

The machine is normally classified as an impulse turbine. This is not strictly correct and is probably based on the fact that the original design was a true constant-pressure turbine. A sufficiently large gap was left between the nozzle and the runner, so that the jet entered the runner without any static pressure. Modern designs are usually built with a nozzle that covers a bigger arc of the runner periphery. With this measure, unit flow is increased, permitting to keep turbine size smaller. These designs work as impulse turbines only with small gate opening, when the reduced flow does not completely fill the passages between blades and the pressure inside the runner therefore is atmospheric. With increased flow completely filling the passages between the blades, there is a slight positive pressure; the turbine now works as a reaction machine.

Cross-Flow turbines may be applied over a head range from less than 2 m to more than 100 m (Ossberger has supplied turbines for heads up to 250 m). A large variety of flow rates may be accommodated with a constant diameter runner, by varying the inlet and runner width (x in fig. 18). This makes it possible to reduce the need for tooling, jigs and fixtures in manufacture considerably. Ratios of rotor width/diameter, from 0.2 to 4.5 have been made. For wide rotors, supporting discs welded to the shaft at equal intervals prevent the blades from bending.

Fig. 18: Cross-Flow Runner

Fig. 19: Cross-Flow Schematic


A valuable feature of the Cross-Flow turbine is its relatively flat efficiency curve, which Ossberger are further improving by using a divided gate. This means that at reduced flow, efficiency is still quite high, a consideration that may be more important than a higher optimum-point efficiency of other turbines.

It is easy to understand why Cross-Flow turbines are much easier to make than other types, by looking at fig. 18 and 19.


Fig. 20 is a graphical presentation of a general turbine application range of conventional designs. The usual range for commercially available Cross-Flow turbines is shown in relation (dotted line). In the overall picture, it is clearly a small turbine.

Fig. 20: Turbine Application Range

Source: James Leffel Co. USA

Fig. 21 shows efficiencies of some of the more important turbine types in relation to gate opening, e.g. flow rate. Conventional and highly optimised turbines (including the Pelton turbine which is not shown) achieve efficiencies of more than 90 % in large units. The Ossberger Cross-Flow has around 80 % for a wide range of flow, and the Cross-Flow turbines built in Nepal achieve over 70 %. On a small unit of, say, 40 kW capacity, the maximum difference in efficiency of the Nepal Cross-Flow, and an imported conventional type would be around 10 % at the optimal point. Given the same head and flow condition, this gives a reduced output for the Cross-Flow turbine of around 5 kW. Depending on turbine type, this difference is likely to be smaller or even reversed at reduced flow (e.g. Cross-Flow compared to Francis or Propeller) and also in cases where a standardised conventional turbine is installed in non-optimal conditions.

For more specific reference, the application range of the two designs of Cross-Flow turbines T1 + T3 built by BYS in Nepal are shown (fig. 22) in relation to a range of standardised conventional machines of BELL in Switzerland. Locally built Cross-Flow turbines in other developing countries cover a similar range. Where need arises, it is possible to extend the application as regards head, flow and output. In Indonesia, for instance, the output range has been extended to 400 kW.

Fig. 21: Efficiency Curves of some Turbine Types

Source: Adapted from James Leffel Co.

Fig. 22: Application Range of Nepal Cross-Flow Turbines and Small, Conventional Types

Source: Adapted from BELL, Switzerland

There are a number of manufacturers, mostly in industrialised countries, who offer equipment specifically for outputs below 100 kW. Turbine types range from Propeller and Cross-Flow to Francis and Pelton. For greater clarity, these are not shown in the diagram. Addresses, though, are given in the annexe for reference.


Turbine T1, the first Cross-flow turbine that was built in Nepal has been specifically designed for manufacturing facilities available at BYS, and is of fully-welded construction. Due to the lack of motorable roads to most installation sites, special consideration had to be given to transportation on the back of porters. Individual parts of the turbine are therefore bolted together and kept in position by taper pins. Thus, a turbine carried to the site in individual parts, can readily be assembled. This is also an advantage, if it should become necessary after some years of operation to repair or replace one of the parts.

Fig. 23: Schematic of Turbine T1

Source: BYS, Nepal

The inlet (1) consists of two curved sheets that form a logarithmical spiral, welded to two plane side panels to form a rectangular inlet section and nozzle. The width of the inlet is denoted x in fig. 23 and in the table fig. 31. The rotor (3) consists of 28 blade segments (2) that are cut from standard diameter five inch pipe of 5 mm wall thickness, which fit into slots of two side discs of 400 mm diameter, where they are welded in. The central shaft (3) is also welded to the rotor discs and final machining of the rotor outside diameter, including the blade tips as well as the shaft diameter, is done after completed welding. The drum-like rotor is also provided with a central supporting disc for the blades, for sizes of x > 220 mm. The shaft extends from both sides of the rotor and is usually symmetric. Depending on the application of the turbine, either both shaft ends can be provided with pulleys to drive two machines via belt-drive, or, if a generator is connected on one side, the other end may be used for operating a speed governor. Bearings used are of the self-aligning spherical double-roller type, which makes accurate machining of the bearing supports unnecessary.

Flow is controlled by the flow regulator (4). Its shaft is parallel to the rotor shaft, with two U-channel parts connecting the regulator shaft with the rectangular tongue at the top. The latter acts as the gate and fits neatly inside the nozzle to keep leaks at the sides in the closed condition within limits. The device is operated by a pushrod (5) which is either connected to a handwheel (6) -requiring a thread on the pushrod and a nut in the handwheel - or, for automatic operation, to the hydraulic cylinder of a speed governor. The housing is completed with the base part (8) and the rear part (9), all bolted to the foundation frame (7).

In addition, two side panels of thin sheet, stuffing boxes and rubber gaskets are required to seal up the turbine housing. The photograph of fig. 24 shows an almost complete turbine assembly on a foundation frame that also accommodates the stand for a small alternator.

In all cases, an adaptor is provided at the turbine inlet that connects the penstock with the turbine. This part is of square shape at one end, to fit to the square inlet, and of circular cross section at the other end to fit to the penstock pipe used.

Depending on the setting above tailwater in an installation, a drafttube of square shape is also provided. For this, a flange made from sheet strips is welded to the foundation frame, so that the drafttube can be bolted on.

To cover the head and flow ranges as given in the table fig. 31 at the end of this section, the turbine is manufactured in 10 different nozzle widths x. The table in fig. 25 shows the standard sizes of x in millimeters and also corresponding other variable dimensions. The diagrams show measurements that remain constant.

In conclusion it may be noted that turbine T1 is suitable for manufacture in a non-specialised metal workshop. Machine tools required are standard, such as:

-Turning lathe with a centre height> 200 mm

-Drilling machine with a capacity up to 0 25 mm and boring attachment

-Milling machine or shaper

-Acetylene cutting torch, plate shear (optional)

-arc welding equipment

-a number of jigs and fixtures made for the purpose

-general hand tools

Manufacturing can be carried out by a team of three or four, consisting of a trained mechanic, a skilled worker trained on the job, and semi-skilled helpers.

Fig. 25: General Dimensions of BYS Cross-Flow Turbine T1

Source: BYS, Nepal


Photographs fig. 26, 27 show some stages of manufacture at BYS. Material used is all common mild steel.

In its application, generally speaking, the machine is clearly overdesigned for outputs of below 25 kW, giving it a long life. For higher output and depending on head and the width of the rotor (shaft bending load), engineering know-how is required to decide whether or not parts that give greater strength must be incorporated (such as bigger shaft diameter, supporting disc, strengthening ribs).

Turbine T3 has a more recent history: its design and development is based on experience with T1, and also on the design of turbine type 205 of BEW (Butwal Engineering Works, Butwal, Nepal). Interestingly, the two major Cross-Flow turbine manufacturers in Nepal, BYS and BEW, closely cooperate in technical matters relating to water turbines, although they are competitors in a commercial sense. On the initiative of an engineering student in Switzerland, T3 was developed with information from Nepal fed in through SKAT, and subsequently a prototype was built with a number of rotors of different design, making a laboratory optimisation of the runner configuration possible. The test program was carried out within the frame of student thesis work at the Swiss Engineering College (HTL) of Brugg/Windisch. Photograph fig. 28 shows the prototype turbine with an output of 30 kW under a head of 70 meters installed in the laboratory at HTL Brugg.


Lastly, with the test results and some minor adaptations of the design, standardisation and a final design were done by BYS in Nepal. At the time of writing, several turbines T3 have already been built.

Turbine T3 has a runner diameter of 200 mm and is therefore much smaller than T1. It is not necessary to make the housing in several parts for transportation purposes. The smaller runner diameter gives it double the speed of T1 which may be an advantage for high-speed applications such as electricity generation. On the other hand, for low-head applications, where a higher speed runner is usually desirable, T3 would necessitate more than double the inlet width as compared to T1, due to its lower specific discharge which results from the smaller rotor diameter. The advantage of smaller general dimensions is therefore compensated for at a certain width of the turbine, so that the required speed and limits of material strength become determininq factors to choose between the two designs.

Fig. 29: Schematic of Turbine T3

Source: Adapted from design A. Arter


The schematical presentation of the T3 design (fig. 29) shows no basic difference as compared to T1, except for the flow regulator arrangement. For T3, a wicket-gate (or butterfly) type was chosen which is situated directly in the inlet. The flow is divided into two sections in this design, which is of no obvious consequence in performance, but permits less material-intensive construction. The wicket-gate of relatively small mass and hydraulically balanced, requires smaller operating forces than the flow regulator of T1. This is an advantage if a hydraulic governor is applied. Another new feature is an accesshole at the top of the inlet. This permits removal of the gate with the turbine in place in the installation, and partial inspection of the rotor.

The design is also optimised to a degree regarding manufacture. The turbine is somewhat simpler to make, is more compact and has fewer parts than T1. Fabrication techniques used are essentially the same, except for the runner blades. These are stamped from 2.5 mm thick steel sheet strips with a hand-hydraulic press. For welding of all parts, a different set of jigs and fixtures is of course required. Turbine T3 has been standardised into inlet widths (denoted b0) 50, 70, 90, 120, 160, 220, 290, 390, 520, 690, 920 mm, to cover the range approximately shown in fig. 22. For large widths, up to 4 supporting discs (depending on head) in the rotor are required to give the blades the necessary strength. Bearings used are of standard self-aligning ball-type with flanges that are bolted to the turbine housing.

Fig. 30 shows a layout of turbine T3 with: (from left to right, pipe adaptor, base frame -both with rubber gaskets and nuts/bolts -turbine housing with gasket and stuffing box and bearing block in front, the rotor, wicket-gate with operating lever and bushing, and finally the access hole-lid with gasket. Parts of the manual gate operating mechanism are in front of the base frame.

For a comparison of the two turbines from the design and application point of view, the main specifications have been tabulated in fig. 31. It may be noted that the attempt at optimisation has resulted in a gain of 5 % efficiency for T3. Besides this, the speed range is more suitable for electricity generation -in some cases direct coupling to an alternator is possible without the need for a step-up transmission -and the head range covered is greater. T1 still has its place, specially for heads below 19 meters and for very low heads, for low-speed mechanical power transmission.

Fig. 31: Technical Features of Cross-Flow Turbines T1 and T3

Source: Compiled from BYS-data

b) Other equipment



For turbine installations in Nepal, the penstock pipes -which deliver the water to the turbine -are usually made from rolled steel sheet welded longitudinally. Flanges at each end are required so that penstock sections can be bolted together in place. For sealing between flanges, flat rubber gaskets are used. Alternatively, it is also possible to use a sealing method developed by BEW, where an o-ring (a standard rubber ring with circular cross-section) is used. One of the flanges is provided with a groove in which the ring rests, so that nearly half of its section protrudes. This is then bolted to the flat face of the second flange, thereby compressing the seal It is reported that this kind of seal is very effective and is cheaper to make than flat rubber gaskets. Another technique that should be promising where portable welding sets are available, is to weld penstock sections together at the site. While this eliminates flanges, bolts and gaskets, much depends naturally on the skill of welders.

For a safe design of penstock strength, general engineering practice applies. If water hammer -due to sudden closure of the turbine gate -is under control by appropriate governor design, it is usually sufficient to calculate the required sheet thickness with the hoop-stress formula.* For considerations of corrosion during the life-time and for a welding factor of less than 1, it is the common BYS practice to add 1.5 to 2 mm to the calculated value and then take the nearest standard sheet thickness. Results with this procedure have been satisfactory. Often, however, permissible stress is not known. In such a case it would be sound to make a pressure-test using a water pressure testing pump on a section of penstock to which end lids are bolted. This is also very suitable for checking welding seams for tiny leaks and for determining the performance of flanges and gaskets. If in such tests a pipe withstands the threefold operating pressure, it may be considered safe.

It should be noted, that the procedure described does not consider other stresses that a penstock may be subjected to; notably stresses due to elongation, bending stress due to the weight of water and the pipe between supports, and collapsing strength in case of valve closure at the top of the penstock. As a rule, it is more economical to take measures avoiding such additional stresses to a great degree than to dimension the sheet thickness based on a combined-stress calculation. For one, it is relatively easy to provide a sufficient number of supports. Elongation should be taken up by expansion joints (usually one at the higher end between two anchor blocks) such as shown in fig. 32. The packing-type of expansion joint is quite simply made by making one section of the pipe larger in diameter so that the other section fits inside and overlaps for a length of approximately twice the diameter. A stuffing-box type of packing is provided, that acts as a seal and in which the inner pipe

To avoid other stresses, two simple parts may be incorporated at the penstock top. A vertical air inlet-pipe of slightly greater height than the head water level and a small bypass valve across the main valve. The former admits air to the penstock when the main valve is closed and the water rushes out and the latter is used to fill the empty penstock slowly, while the main valve is still closed. All three components are shown in fig. 33, the main valve being of the common gate type, also locally made and of welded construction, as can be seen on the left-hand side photograph.

Where alternative material to sheet-metal pipe is available, it is of course worthwhile to compare costs and suitability. It may often be found that PVC and/or PE (PVC = Rigid Polyvinyl-Chloride, PE = High Density Polyethylene), pipes in diameters up to 200 mm cost much less than steel pipe. In Nepal, so far only a few installations are provided with PVC or PE penstocks, mainly because of very limited availability. The use of locally-made ferrocement pipe is also reported from China.

An important consideration in choosing the penstock diameter is the loss of head -due to friction of water flow -versus cost of the pipe. It is difficult to state a generally applicable rule. Power output required, head and flow available and costs of the pipe are factors that must be investigated in each case. Depending on these, an economic solution is a trade-off between cost and head loss, the latter representing a loss of energy. A general rule often applied is that the flow velocity in the pipe should be well below 4 m/s. The diagram fig. 34 shows the typical relationship between flow rate, the diameter of pipe used and the head-loss (in meters) in a pipe section of 10 m length, made from rolled steel sheet. Values for a Ø 200 PE (plastic) pipe are also given, showing that a plastic pipe has smaller losses due to a smoother surface.

Fig. 34: Typical Head Loss in Pipe

Source: BYS Cross-Flow turbine, Adams 1974



Because of the use of a constant-diameter runner in the Cross-Flow turbines and its use for a big range of heads, it is usually necessary to provide a step-up speed transmission. In the case of mechanical power use, it may be that several machines should be operated. A first transmission step from the turbine to a line shaft is then done using either flat belts or vee-belts with pulleys that give the desired line shaft speed. Two, three or more machines may be operated again with belts with a set of pulleys, one for each machine on the line shaft and one on the machine shaft itself. This is a very simple and conventional technique that is used everywhere. It is fairly efficient if shaft alignment and belt tension is correct and without problems if safety measures -such as providing guard-rails - are taken. Fig. 35 shows a typical turbine-mill, where three machines are operated by flat belts.

With the heads under which T1 is used, step-up transmission is required in all cases for electricity generation using standard equipment. A step-up ratio of 5 in a single step is relatively simple with standard vee-belt drives. This would be needed for a head of about 9,5 m with T1 and a generator of 1500 RPM, where the largest size would generate over 25 kW. To transmit this power, a multiple belt arranqement is required, using belt-manufacturer's instructions for the design. For lower heads with turbine T1, a two-step transmission would be required with vee-belts, making it somewhat expensive. This may be a reason for selecting the faster speed T3-turbine or, alternatively, another step-up arrangement, such as a chain drive or gearbox.

The latter are quite sophisticated pieces of equipment and therefore costlier than most alternatives. In Nepal, no gear boxes have been used so far mainly for this reason. A chain drive,on the other hand, has been applied in a number of installations. Using Duplex or Triplex roller-chains of standard pitch, sprocket and pinion were made locally from mild steel. It may be considered a cost-effective solution even though a housing with oil-bath lubrication is required (see fig. 36). Initially, there was wear on the mild steel sprocket. This was later made from case-hardened steel.

Except in a few cases, vee-belt drives may prove to be the best solution. Belts and multiple-groove pulleys are standardised and mass-produced, making them relatively cheap. In addition, a properly designed vee-belt drive requires little care and is quite efficient and durable. This view is also corroborated for Latin America by OLADE (personal communication with OLADE).

For turbine T3 it is possible in some cases to couple the turbine directly to the generator. At a head of 58 m,for instance, turbine speed is 1500 RPM, making it ideal for direct coupling to standard 1500 RPM (4 pole) generators. Due to a quite flat peak of the efficiency/speed curve of the Cross-Flow turbine, the head range for direct coupling may be extended above and below 58 meters without losing efficiency, since a step-up transmission would also incur losses on the order of 3 to 4 %. Speed of the turbine may be increased or decreased to match 1500 RPM in the range from 41 to 78 m head. As fig. 37 shows, this will result in a lower turbine efficiency by about three percent. Since in that case no transmission is required, costs are saved and overall efficiency remains the same.

Fig. 37: Turbine T3: Efficiency/Speed Curve

Source: adapted from Diploma work HTL Brugg

Fig. 38: Discharge/Speed Curves


The discharge of water through the turbine is also affected. The diagram fig. 38 shows that if speed is increased to 1500 RPM, relative discharge falls by about 20 %. This means that under this condition, turbine inlet width has to be increased by 20 % to achieve a relative discharge of 1. If speed is decreased on the other hand, discharge increases, but by a much smaller amount. This, for partical purposes, may be neglected. The two discharge curves shown in diagram fig. 38 correspond to the conditions in fig. 37.


Generators are machines used for the production of electrical energy. Their operation is based on the principle of electrical induction whereby a periodic flow of electricity is produced in a loop-type conductor as a result of the periodic variation of the flux of the magnetic lines of force passing through this loop. To do this, one can either cause the loop to rotate in a constant magnetic field or, alternatively, the loop can be kept stationary and the magnetic field rotated. In the latter arrangement the armature is stationary, and the magnetic poles revolve instead; the stator consists of an iron ring with induction coils mounted on the inside; the magnetic poles on the rotor move past the ends of these coils at a very short distance from them (fig. 39 and 40). In this case the current produced by the generator is taken direct from the stator. The relatively low ouptput of direct current needed for producing the rotating magnetic field is fed to the rotor by means of slip-rings and carbon or copper-mesh brushes (fig. 40). Fig. 40 shows a smaller generator which likewise operates on the principle described above (rotating magnetic field, stationary armature winding). In this case the magnet wheel is in the form of a two-part T-rotor.

Fig. 39: Alternating-Current Generator (schematic)

Fig. 40: Details of a Generator (internal pole machine)

Source: (fig. 39 + 40), How Things Work, p 65


There are generators that produce direct current (DC) but these are used for special applications and low outputs only. In the context of rural electrification they are of limited interest and are therefore not discussed here. Generators producing alternating current (AC) are very often called alternators. They may be single phase, supplying a voltage of 200 to 240 Volts (depending on standard) or more often, three phase. Single phase alternators are available in sizes commonly not exceeding 12 kW and may well be used in small installations.

Three-phase alternators are more versatile in relation to electricity end-use. Units in the micro-range produce a voltage of 380 to 440 Volts between phases and of 200 to 240 Volts between any phase and neutral, depending on the standard applied. The "four-wire-system" is required on the distribution side and the load on the three phases must be balanced within prescribed limits.

For isolated small hydropower stations the most convenient machine to use is the self-excited, synchronous 3-phase alternator. The other type available is the asynchronous generator which requires excitation from an existing grid or through batteries. Due to this, it is of limited interest, although it has no need for speed governing in certain cases, and is cheaper than synchronous alternators.

Two types of synchronous alternators are mass produced and therefore relatively cheap. The more conventional type has slip-rings and brushes trough which exitation is provided from a booster-transformer/rectifier, static excitation system. Many bigger developing countries produce this type and it may therefore be available on the regional market. A newer type of alternator is provided with rotating rectifiers directly mounted on the main shaft. Excitation current can be directly supplied to the rotor windings without the need for slip-rings and brushes. Such machines are therefore called brushless alternators. It is difficult to select one of the two types on the basis of technical criteria. A brushless alternator requires less maintenance than a slip-ring alternator where brushes need to be replaced periodically. Yet the former is a more sophisticated piece of equipment than the latter. Should any repair be required, it is in most cases easier to work on a slip-ring alternator. Notwithstanding this, the reason to select one or the other type is probably its price and availability.


Fig. 41: Speed of Standard Alternators

No of poles

Speed (RPM)


at 50 Hz

at 60 Hz




















Fig. 42: Brushless Alternators (2-pole & 4-pole)

Source: Leroy-Somer, France


There is still another technical aspect that needs looking at: Depending on the number of poles on the rotor, speed at which alternators are to be operated varies. For a given frequency, the higher the number of poles the lower the speed needs to be. More poles imply a greater weight and a larger size; the relationship of these parameters is shown in an exemplary table in fig. 41 and fig. 42.

Lower speed alternators in the range up to 160 kW are produced less often, and are obtainable from many manufacturers upon request only, while 2 pole alternators are not available for outputs above 60 kW.

Several manufacturers in the People's Republic of China offer a large range of low-speed alternators in execution with or without brushes, such as shown in fig. 43. These are specially built for use with water turbines and are worth investigating. An address is given in the annexe.


Fig. 43: Low Speed Synchronous Alternator

Source: CMEC, Beijng, China

Worldwide, by far the most easily available alternators are 4 pole, running at 1500 RPM for a frequency of 50 HZ. These machines -as most other types - are usually built for operation with thermal prime movers, e.g. diesel sets for small sizes. There is a fundamental difference in run-away speed as compared to water turbines in that a combustion engine for instance may have a runaway speed 20 to 40 % higher than its rated speed under load.

For water turbines, this run-away speed is typically 80 % to 100 % higher than the rated speed. If now in a plant in operation the load is suddenly switched off and the speed governor does not react in case of failure, turbine speed and thus generator speed go up to run-away speed within seconds. An alternator built for a lower permissible run-away speed may then not be able to withstand the higher centrifugal forces -which incidentally increase with the square of speed - and a coil loop on the rotor could be deflected outwards. If it touches the stator, the result would be disastrous and would wreck the machine. Even where this is not happening, increased load on the bearings will shorten the lifetime of the alternator. The only way out is to specifically ask manufacturers to guarantee for a permissible runaway speed above that of the turbine. In the experience of Nepal, most manufacturers seem to be in a position to meet such requirements, often at the same price as for the standard machine.

When selecting an alternator, two more points deserve attention. For one, standard insulation is not always suitable for tropical (humid) climate. To be on the safe side, users should specify in what relative humidity an alternator is to be installed. The other point concerns the elevation above sea level of an installation and ambient temperature. Both affect alternator maximum-output. Usually, manufacturer's output data relate to max. 1000 m elevation and to max. 40° C ambient temperature. Above these values, factors for aerating must be applied, by multiplying them with the rated output to arrive at the situation specific output. The table fig. 44 shows typical factors which may of course vary with different products.


Fig. 44: Output Derating Factors

Source: Leroy-Somer, France


Electricity generators are complex machines that require numerous considerations prior to procurement, to make sure that a suitable type is selected. Manufacturers are usually quite ready to supply information either with standard brochures or when specifically asked about the various aspects. It is important to specify all relevant requirements and ask all the relevant questions to arrive at a satisfactory solution. A typical and detailed enquiry then, could look as follows:


3-phase Alternator for use with hydraulic turbine


self-excited, synchronous, with built-in excitation and voltage regulation


1500 RPM at 50 cycles (Hz) with a permissible run-away speed of 1.8 rated speed


380/220 Volts, 3 phase


40 kW at power factor 0.8 e.g. 50 KVA continous operation

Operating situation:

1200 m elevation, 30° C ambient temperature, high humidity

Mounting mode:

horizontal shaft, foot mounted, with splashwater protection


connectors for 4-wire system

•Manufacturers' information required on:

- Moment of inertia and total weight


- General dimensions of the machine


- Mounting plan


- Derating factors applicable


- Execution with or without brushes


- Cooling mode


- Efficiency/load curve


- Permissible overload


- Permissible imbalance in phase loading


- Motor starting capacity


- Limits of voltage regulation from no-load to full-load


- Possibility of parallel operation with other generators


- Maintenance requirements


- Guarantee period


- Delivery time


- Price (CIF & FOB)

Since there are many potential suppliers, it is well worthwhile to send out a number of enquiries, at least initially. Prices and product quality may differ considerably but delivery time and after-sales services may also be important.

In summary, procuring alternators for small hydro-electric projects is not an easy but a time-consuming task and an effort is necessary for a satisfactory solution regarding costs and performance. Local manufacture -as for turbines - may seem an attractive possibility in some countries as a new venture in the long run. In the case of Nepal, an evaluation of this possibility was done at one point. The conclusion was -at that time -negative. The lack of necessary component material such as transformer sheets, copper wire and insulating materials, and mainly competition with outside suppliers on a relatively small market and the lack of trained personnel, were the reasons. Based on this, it may be acceptable to conclude that local manufacture of alternators need not be a priority for a new venture in those countries where no production facilities exist yet.

The same applies to switchgear and control instruments. Such components exist in large varieties and are easily available. To find the right types again requires investigation of the local market. Assembly and wiring in a locally made panel or box is according to electrical engineer's practice. Provided a skilled man is available who knows his job, there is no problem. Usually only basic components are required such as a main switch with relay, a set of appropriate fuses, a volt-meter with multi-stage switch to check voltage between different phases, three ampere-meters (one for each phase) and finally a frequency-meter. Meters for kW, kWh and power factor are convenient but not strictly necessary permanently, and whether or not an under-voltage/over-voltage drop relay is required, depends on the kind of governing applied.

Electricity generation, switching, transmission and distribution is a very expansive subject and can hardly be discussed in sufficient detail here. There is a lot of specific and very practical literature available on the subject. Some handbooks have been written specially for the context of rural electrification and these are recommended.




Electric motors and appliances require a stable voltage and frequency. An electric generator, on the other hand, produces such stable voltage and frequency only if it is run at constant speed. A water turbine delivers such constant speed at a given gate opening if the load on its shaft is kept constant. Changing the load, e.g. switching power on and off, results in speed variations which in turn cause variations in voltage and frequency of the electricity produced. To keep such variations within acceptable limits, it is necessary to incorporate a turbine controller, also called a governor. To achieve control, there are chiefly two possibilities that may be applied:

· load-control, where the flow of water through the turbine is kept constant and where therefore the load has to be kept constant within tolerable limits. This is achieved today mainly with electronic controllers that switch any part of the load not consumed by the regular circuit into a ballast circuit, thus keeping the total load on the turbine-alternator set constant.

· flow-control, where the volume of water flowing through the turbine is adjusted depending on the load on the turbine-generator set.

Both systems keep the speed of the turbine within tolerable limits which, in turn, results in a voltage and frequency that remains within a specified range.

In the development of Cross-Flow turbines in Nepal, the question of an appropriate governing device was raised at an early stage. While there exist virtually dozens of designs for flow-control governors, it is not easy to copy one of them. Most are complex and difficult to manufacture for a non-specialised workshop. If imported from one of the major suppliers, they tend to be very costly and more accurate in performance than would really be required in a rural situation. Keeping financial viability in mind, it was decided at an early stage to look for another solution. An electronic load-controller was procured at one stage for a 20 kW three-phase unit. While this type performed well in single-phase application, there was a problem with the three-phase configuration. When phases were not properly balanced, it aggravated the problem by overloading the phase that already had the highest load, on the ballast side.

This experience led to the development of a new type of load-controller based on the principle of current-sensing versus voltage-sensing that was used on the first device. Research work was carried out at the Swiss Federal Institute of Technology at Lausanne (EPFL) and a prototype was later successfully tested in Nepal. Based on this, efforts are now underway for simplifying the device and to assemble it locally using components from India. This work is being done by BEW with the help of an expert of the Centre for Electronics Design Technology (CEDT), Bangalore, India. Fig. 45 shows the full prototype assembly for demonstrating purposes on a plane board -including full wiring with fuses and overload switches.

At the same time, development work on a simple, water-hydraulic, mechanical flow-control governor was undertaken. For this purpose a full-scale Cross-Flow turbine of type T1 was installed in the laboratory of the Swiss Federal Institute of Technology, Zurich (ETHZ), where the Institute for Fluid Technology carried out the work. To keep the device simple, a proportional type was developed and water pressure, under the head of the installation, was to be used as a working fluid. Since the water is drawn from the penstock and is discharged continuously through the control valve, there is no need for a pump, resulting in a really simple device which can be manufactured locally. The schematical diagram in fig. 46 shows the components of the governor. Its working principle is chiefly based on a two-throttle system and on the principle of a centrifugal pendulum: Two throttles mounted in line, the first with a constant-area orifice and the second with a variable discharge area, create a variable pressure in the line between the two throttles, depending on the discharge area of the variable throttle. The discharge area may be varied by moving a piston in axial direction. The centrifugal pendulum (or flyweight) on the other hand, moves a pushrod to a certain position depending on speed, if counterbalanced by an appropriate spring. If now the flyweight is mounted on the turbine shaft, and is connected via a lever to the piston of the variable throttle (or pilot valve), the variable pressure can be made proportional to the speed of the turbine. A piston, connected to the turbine gate and counterbalanced by a spring, is subjected to the variable pressure and thus, a turbine gate position is determined by speed. Since a certain gate opening corresponds to a certain load, speed of the turbine is now determining its loading condition.

Fig. 46: Schematic of Generating-Set with Mechanical Water-Pressure Governor

Source: Meier, Manual for the Design of a Simple Mechanical Governor

If load is switched off, speed increases, which results in a movement of the flyweight pushrod. This in turn moves the pilot valve piston, which results in a lower pressure on the servo-cylinder piston. The turbine gate in turn closes until piston force and closing spring are in balance again, thus adjusting the turbine gate to the new situation. In case of switching load on, the reverse happens, and in case of pressure loss in the supply pipe, the gate of the turbine closes completely and shuts the plant down (A more complete description of this governor and design manual is available from SKAT upon request). The system was initially designed for a maximum speed-deviation of ± 10 %. After testing of the prototype in Nepal it was found that the governor was stable enough to permit speed regulation within + 5 %, a value that seems better than acceptable in most situations.

The photograph of fig. 47 on the left shows the rotating fly-weight mounted directly on the turbine shaft. The connecting lever with flyweight-spring on the left and connection to the pilot-piston on the right. The picture on the right-hand side shows the turbine with the governor unit in front and the servo-cylinder to the left. Connection between cylinder and pilot valve is by diameter 50 mm PE pipe.

There are now two relatively cheap governing systems available for application and the question arises as to which should be used for what situation. The answer -as is often the case when there is a choice - is not simple. Both systems, though past the prototype stage, have no substantial performance records to date. The mechanical governor seems suitable for installations with a single turbine. It has the advantage of keeping water that is not strictly required for power generation from running through the turbine. This may be important where excess water from the head-race canal is used for irrigation. A mechanical governor may require much more frequent adjustments and maintenance than an electronic device although the skills required are fewer by far.

An electronic load-controller, on the other hand, is a sophisticated piece of equipment. Still, it is solid state, without moving and wearing parts, and if a breakdown occurs, repair could be attempted by semi-skilled personnel, with plug-in modules and according to a prescribed routine. In addition, it might prove cheaper to use one electronic load controller on a station where two turbine-alternator sets work in parallel, instead of two mechanical governors.

Transfer of technology for both devices described is in principle possible. In the case of the electronic load controller much depends, though, on the state of the local electronics industry. For the mechanical governor, the situation is clearer. Technology involved is simpler and can therefore more easily be documented and transferred. Although the actual design has been done for use on the Tl-turbine design, it will be relatively easy to adapt for another design of turbine. For prototype testing purposes, though, some sort of testing-facility must be available. Refinements are possible only empirically, if comprehensive mathematical modelling of the dynamic behaviour is to be avoided.

There are still other solutions possible to the problem of governing: In cases where load fluctuations can be limited, e.g. where changes in load on a plant are small in relation to total output, and relatively infrequent, hand-regulation may be acceptable. An operator on duty would in this case make necessary adjustments on a hand-wheel, as soon as load changes become evident from instrument readings. This simple method is still quite widely used on small plants.

The second possibility is to use an imported governor. Usually, these are prohibitively costly but there is information available of governor-manufacturing in the People's Republic of China. The National Export agency is offering various types of oil-hydraulic, mechanical governors such as shown in fig. 48. At the time of writing, firm prices were not available but costs are reportedly only a fraction of what a governor of this kind from the West would cost. This, may therefore be of interest.

Fig. 48: Oil-Hydraulic, Mechanical Speed Governor from PR China

Source: CMEC Brochure



A Cross-Flow turbine with adapter and penstock, a step-up transmission, a mechanical governor with flywheel, a coupling, connecting the flywheel to the alternator, and finally the alternator itself, with switchboard and connecting terminals, comprise what is called the complete electro-mechanical generating equipment, necessary for a rural electrification project (refer fig. 49).

All items are bolted to a common base-frame for ease of alignment and the whole set in turn is bolted onto a solid foundation. The flywheel, which has not been discussed so far, may be considered a part of a flow-control governor. Because governor action in case of a load change does not follow instantly, but occurs at a definite speed, a transitory state occurs, during which the turbine speeds up or looses speed depending on the load change. A flywheel, together with all other rotating masses, takes up energy in the process of being accelerated and so slows down the rate of speed increase. In the case of falling turbine speed, the rotating masses deliver energy due to deceleration, thereby trying to maintain speed. Thus, a flywheel serves to "smooth-out" speed deviations in the transitory period immediately following load changes, and gives the governor more time to do its job, while keeping speed deviations within limits.

It is easy to understand that the necessary flywheel size is determined by the governor characteristic, by the maximum permissible speed-deviation and by the size of other rotating masses already incorporated in the set. A flywheel has the biggest effect if mounted on the fastest turning shaft available. This will usually mean mounting it in line with the alternator. From the design point of view, it is important that stresses due to centrifugal forces remain within safe limits. As all other components, the flywheel must withstand run-away speed. For a maximum speed of 2800 RPM, for instance, a max. safe diameter of 750 mm was determined in Nepal for flywheels welded up from mild steel plates. When using the electronic load-controller, it was found that a flywheel is not strictly necessary. Because an electronic device -without moving parts -acts immediately, existing rotating masses of turbine, speed transmission and alternator are sufficiently large to give the system stability.

Other steel-fabricated parts not shown are: A trashrack situated in the fore-bay in front of the penstock inlet. This serves to keep floating particles such as leaves and wooden sticks from entering. For easy cleaning, it is removable. It must be of strong construction so it can withstand water pressure even if fully covered by leaves. At the same time, spacing of rods must be small enough to retain floating parts (e.g. usually slightly smaller spacing than the distance between runner blades), while the total flow area must be large enough to admit sufficient flow.

For shutting down the plant, a gate-valve is incorporated in the penstock above the turbine inlet, and another valve at the bottom of the forebay is needed to empty it for maintenance work and to flush out sediment. Depending on the situation, another (coarse) trashrack may be required near the canal intake, to keep stones from entering the canal. A simple lifting gate may also permit to empty the canal for maintenance work. At less expense, though, the same can be achieved with simple stop-logs, e.g. a number of wooden planks that fit into slots on both sides of the canal to form a temporary barrier across it.

The arrangement of the electro-mechanical equipment is usually perpendicular to the direction of the penstock and will require fairly little space. As indicated in the schematic fig. 50, a 20 kW set would usually require less than 5 m² of area. A power house with a floor area of 20 m² could in fact accommodate two sets conveniently, with sufficient access space on all sides.

Not only the turbine but also all other components of the system incur losses in operation, e.g. their efficiency is smaller than 1 due to hydraulic losses, mechanical friction, and electro-magnetic losses. Based on the technology described and specific to each situation, efficiencies and losses may realistically be assumed as follows:

Fig. 50: Schematic Layout of a 20 kW Hydro-Electric Generating Set

Source: BYS, Nepal

· Efficiencies:


0.65 to 0.75

-Step-up transmission

0.94 to 0.98


0.78 to 0.92

overall efficiency:

0.48 to 0.68

· Additional losses: these occur in all hydraulic conduits, in the operation of a governor and in transmission and distribution of power. Depending on conduit-length and on the length of transmission and distribution networks as the main parameters, they may constitute between 10 to 25%.

Power available to the user, expressed as a percentage of theoretical generating-potential, is therefore in the range from 36 to 61 %. A highly optimised (imported) turbine might achieve an efficiency of 88 % under ideal conditions, which would supply useful power amounting to a maximum of 71 % of the theoretical value.

c) Survey and Civil Engineering

In the context of field-surveys and site-identification, it must be stressed that the first problem to be solved is not of a technical but of a social nature. Instead of doing a detailed and precise survey directly, as soon as a site that looks feasible is identified, it should be a priority to get in contact with the local population, village elders and community leaders. It should be made clear in detail, how a small hydropower project might affect their lives and what would be expected from them in terms of participation during all stages of a project, and also of what it would cost them in terms of monthly electricity bills. Only after people have understood the implications, and consider further activities their project, need more detailed investigation work be done.


To locate an optimal site is a question of the available potential related to the expected power requirements and, further, of how easily the potential may be developed, and also how far the site is away from the consumers to be supplied. In cases where mechanical power use is envisaged, it is also a question of easy accessability since most likely some goods will have to be transported to and from the site, as for instance in the case of grain milling. A method to arrive at the best choice of site may be to evaluate first the parameters approximately of the site that looks most promising from a purely visual investigation. This serves as a basis for reference. Other sites are then surveyed superficially and the one or two that look best in a comparison are then surveyed in detail.

The type of development in the range of interest and with the equipment referred to will normally be run-of-river, in the low to medium head-range, e.g. an installation where water is supplied through a canal and penstock to a power house that is built near the river bank. On rivers with steep gradients, only a short canal is required but still better is an abrupt fall in the river. Such conditions usually permit a medium-head development at relatively low cost. The practice adopted in Nepal is not to build permanent dams but only a sort of intake-structure that diverts water to a canal-intake low enough in relation to the water level of the river to guarantee sufficient inflow. One method is to go upstream from the previously identified installation site, up to a point that would give a sufficient head when a horizontal line is sighted along the hill slope towards the site. A team must then walk along this horizontal contour line and also at a higher elevation, along a line parallel to the lower one, to determine whether construction of a canal is possible, and whether head could also be increased from the point of view of canal alignment. The necessary slope of the canal must also be considered at this point and a suitable intake point is then looked for still further up-stream, to permit inclusion of the canal slope. During all stages of surveying, it is important to keep in mind the kind of structures and equipment that are going to be used.

Experience in Nepal shows that it is in fact of little use if a survey is done by a team of competent surveyors who otherwise know nothing of hydropower generation. Utmost accuracy is far less important for small projects than the knowledge of what structural components should best be built at which point. As a rule, a simple engineering-level will be good enough, although where a team does many projects, a theodolite may be justified for greater speed. In areas where local people have a tradition in constructing irrigation canals, it may be found that canal alignment can entirely be left to them. Leveling of head is then also possible without a surveying instrument, by using staffs, a tape, and a spirit level. This simple method is briefly explained in fig. 51. It will permit the establishment of head up to 30 to 40 meters within an error of a few centimenters. Also, it can quite easily be managed by semi-skilled personnel or local people who were trained on the job for a day or so.

The flow available is the other factor that determines theoretical power. A plant that produces energy all year round is desirable and therefore flow-measurements in the dry season are crucial. A few measurements alone of course will not give a meaningful result regarding minimal flow, because it is very often the case that variations in the long term are considerable. Many literature sources suggest that data from the nearest gauging stations, percipitation size of drainage area and other climatological data should be correlated with the few data available from the site of interest. For practical purposes this seems not to bear promise in most situations. Data will definitely be lacking for small streams or be insufficiently reliable.

Fig. 51: Method of leveling head without instrument

Source: BYS, Nepal


Better than doing specialised theoretical work in the office, is perhaps to visit the site several times, do spot-measurements and question local people extensively how the present flow relates to their long term experience. Specially people that are used to gravity irrigation -as is often the case - have considerable knowledge of "their" river. The point is to refrain from wishful thinking and to question also the lowest estimate whether it couldn't have been still less. Experience over the last few years in Nepal shows, that failures are unlikely to occur if such a pragmatic procedure is followed seriously. In investigating a site, it is within limits a good practice to develop maximum head. This will make the use of less voluminous equipment possible and output will be less affected by flow variations. If the conservatively assumed flow in the next few years appears to be more favourable, project-expansion will be quite simply possible by adding another turbine and by using larger flow.

A second turbine may be of considerable interest already initially, where flow variations are certain to occur and minimum-discharge is likely to be much smaller during a short period, than average-discharge. If, from the consumption point of view, it is desirable to have more power available during the greater part of the year than the minimum-flow would provide, a single turbine unit may be inappropriate. Since part-load efficiency is lower than the optimum, such a plant would work least efficient at a time when generating potential is already badly affected by low discharge. Two turbines of unequal size -ideally of an output-ratio of 1 : 2 - would in fact provide three optimum efficiency points: one at low flow with only the small turbine in operation, the next at medium flow with only the bigger turbine working, and the last at high flow with both turbines at optimum gate opening and near maximum plant output.

For measuring river flow, there are various possibilities that are described elsewhere in detail. Using current-meters or even travel-speed of a float on a straight river section, multiplied with the cross-sectional area of the river, yields an approximate value. More accurate and relatively easily done for a discharge of less than 1 m 3/s is the weir-method briefly described in fig. 52.

Besides all the worries concerning the minimum-discharge, it is finally also important to investigate the danger of high or even flood-flow, which may have a damaging effect on structures and equipment. It is in this case not necessary to know the discharge-rate but rather to know the highest possible level that the water will reach, in order to be able to place the equipment floor safely above this level. In the absence of statistical data, it is again local inhabitants who can be of help. It is necessary to question them as regards flood occurrences as far back in history as possible.

Fig. 52: Flow Measurement with Rectangular Weir

Source: BYS, Nepal

In case of small streams or channels a temporary measuring weir can be easily erected as indicated. It may be made of strong timber or metal sheet, with the bottom and sides of the rectangular notch bevelled, to a width of about 2 mm. The distance from the bottom of the notch to the downstream water level should be at least 75 mm (3 inches) and sufficient to allow for complete aeration. This weir is let into the stream and made watertight with clay or plastic sheet so that all the water will pass through the rectangular notch. For accurate measurments a stake should be driven into the stream bed about 2 metres upstream, the top of the stake being level with the crest of the weir. The depth of water flowing over the weir can be obtained by measuring the height (h) from the water surface to the top of the stake. The flow may then be calculated from the table below.


Structural works required are probably the greatest variable component of a hydropower project, since the situation of a particular site tends to be quite unique. While structural components may be of the same principle in different projects, their actual design needs to be site specific to make them cost effective, but still safe. For obvious reasons, structures should be kept simple and what is actually needed may be summarised as follows: A diversion dam (1 in fig. 53) - usually partial and of semi-permanent construction -is required to divert water from the river into the head-race canal (2). A simple lifting gate (3) or grooves to accommodate stop-logs is provided at this point to block off water inflow for canal maintenance. The canal may be an unlined earth canal where the amount of seepage and the danger of slides in unstable terrain are minimal. Lining will avoid such problems and it will give the canal a greater capacity. Flow velocity may be higher in a lined canal as compared to an earth canal, where erosion may be a problem. Lining can be done with stone slabs or with a thin cement mortar. Of much interest would be hydraulically stabilised mud bricks, but there is no experience available that would prove this possible. Lining with plastic sheets, while cheap, has not met with success in Nepal. Cattle use to walk in the canals and this would puncture plastic sheets badly. A rather interesting technique is reported by and is also used by Las Gaviotas in Colombia. It is to cover canals up to prevent them from damage. This is described in fig. 54.

Fig. 53: Structural Components of a Small Run-of-River Project

Source: Adapted from BYS, Nepal


Except where water is free from suspended particles, a sedimentation chamber should be included. This may be a section of canal near the intake where flow velocity is reduced to much less than 50 cm/s. (depending on the nature of suspended matter), by widening the cross-section and reducing the slope. Sediment can be flushed out periodically if a bottom discharge-sluice is provided or it can otherwise be removed manually.

Where the danger of land-slides exists, much care should be taken to follow the contour of the hill with the canal and to avoid excessive "shortcuts". This will minimise excavation necessary. Where a slope has to be cut-in considerably, it may be important to make measures at slope protection such as driving in stakes, or still better, big-engineering methods.

A forebay (4) in stone masonry or concrete, is the next structural component. It is situated at the end of the head-race canal and comprises a small basin with a trash-rack in front of the penstock inlet, and if necessary a sandtrap that has again the job of retaining sand and sediment. The penstock-inlet needs to be submerged 60 to 80 cm below the surface to avoid drawing in air in operation. A lateral spillway (5) along the last canal section serves to discharge excess water and keeps the head constant. A gate (6), similar in construction to the one in the intake, permits emptying the forebay for cleaning and maintenance. The penstock (7), which connects the forebay with the turbine, requires appropriate anchor-blocks and supports, for which stone-masonry is a suitable technique. The last item, the powerhouse, which accomodates the electro-mechanical equipment, completes the installation. Good foundations and a lined pit to take care of the tail-race water are essential here. Otherwise, the local style will be acceptable in most cases.

In all hydraulic structures where a solid foundation is required or where water pressure has to be retained, the use of gabions -wire mesh baskets which are filled with large stones - is a method that deserves attention. Except for the imported galvanized wire, which is commonly used, all material and labour can be entirely local. Complete river training and small dam projects are known to be based on gabion-technology, and their life may well exceed 60 years under normal conditions

A long thin walled polyethylene bag is filled with water to form a flexible sausage. This is done with the plastic already in place in a large enough excavation on a bed of very lean soil and stone-cement mixture (about 6:1). The pipe is then covered with the same mixture while adding water pressure by raising the ends of the "sausage". The mixture is "vibrated" by treading on the pipe near the place of filling. After completion of one section, and a short while of setting, the water is drained from the pipe and the plastic sheet pulled out while twisting it. An adjacent section may then be started and an inspection shaft be made between two sections. Thus, a "concrete-pipe" is cast in place.

Fig. 54: "In-site" Casting of Head-Race Conduit Pipe

Source: Guerrero,

Appropriate hydraulic structures can by definition not be standardised if they are to be effective and low-cost. In most situations, the share of civil engineering in total project cost is somewhat above 50 % and therefore, effective cost reduction in civil engineering will have a greater impact on overall cost than any other single measure. On the development front today - in the field of small hydropower -there is still a tendency to over-design, and to use too much cement and concrete due to a lack of experience and, often, innovation. Small projects in the micro-range are a very suitable training ground, where experiments and the application of new technologies are possible without unbearable risks. In this sense, small hydropower projects can serve as a basis for local technology development, with a scope for up-scaling to larger projects. In design, all structures in contact with water, such as conduits, intakes, spill-ways and basins, require the knowledge of basic hydraulic principles which cannot be explained here. The same is true for questions of soil stability and foundations, where the theory of soil mechanics and foundation technology apply. Some titles worth referring to are from: Jagdish Lal, Grummann, U.S. Bureau of reclamation, Peck and Mata.