Guide on How to Develop a Small Hydropower Plant ESHA 2004
gate
Inclined Kaplan in pit arrangement
Figure 6.20: Cross section of a pit Kaplan power plant
Photo 6.7: Siphon Kaplan
Siphons are reliable, economic, and prevent runaway turbine speed, however they are noisy if no
protection measures are taken to isolate the suction pump and valves during starting and stopping
operations. Even if not required for normal operation, a closing gate is strongly recommended as it
avoids the unintended starting of the turbine due to a strong variation of upstream and downstream
levels. In case of such a problem, the turbine will reach high speeds and the operator will not have
the means to stop it. A solution to this problem is the use of flap gate dams.
Underground powerhouses are best at mitigating the visual and sonic impact, but are only viable
with an S, a right angle drive or a pit configuration.
The speed increaser configuration permits the use of a standard generator usually turning at 750 or
1 000 rpm, and is also reliable, compact and cheap. The S configuration is becoming very popular,
however one disadvantage is that the turbine axis has to cross either the entrance or the outlet pipe
with consequent head losses. It is mainly used for medium heads and/or hydropower schemes with
penstock.
The pit configuration has the advantage of easy access to all the equipment components, in
particular the coupling of turbine and speed increaser, the speed increaser itself and the generator,
which facilitates inspection, maintenance and repair. This configuration is popular for very low
heads and high discharges allowing a runner diameter bigger than 2 m.
For the same reasons as for the Francis turbines, Kaplans must have a draft tube. Due to the low
heads, the kinetic energy is very important and the quality of this part of the turbine should not be
neglected.
167
Guide on How to Develop a Small Hydropower Plant ESHA 2004
6.2.2 Specific speed and similitude
The large majority of hydraulic structures, such as spillways, water intakes, etc. are designed and
built on the basis of the results obtained from preliminary model studies. The behaviour of these
models is based on the principles of hydraulic similitude, including dimensional analysis; the
analysis of the physical quantities engaged in the static and dynamic behaviour of water flow in a
hydraulic structure. The turbine design does not constitute an exception and actually turbine
manufacturers make use of scaled models. The problem of similarity in this case can be summarised
as follows: "Given test data on the performance characteristics of a certain type of turbine under
certain operating conditions, can the performance characteristic of a geometrically similar machine,
under different operating conditions be predicted?" If there is a positive answer to this question the
theory of similitude will provide a scientific criterion for cataloguing turbines that will prove very
useful in the process of selection of the turbine best adapted to the conditions of the scheme.
• Effectively the answer is positive provided that model and industrial turbine are geometrically
similar.
To be geometrically similar the model will be a reduction of the industrial turbine maintaining a
fixed ratio for all homogeneous lengths. The physical quantities involved in geometric similarity are
length, area A and volume. If the length ratio is k, the area ratio will be k2 and the volume ratio k3.
It is particularly important to notice that model tests and laboratory developments are the only way
to guarantee the industrial turbines efficiency and hydraulic behaviour. All the similitude rules are
strictly defined in international IEC standards 60193 and 60041.
No guarantees can be accepted if not complying with these standards and rules.
According to these standards, the specific speed of a turbine is defined as:
n⋅ Q
n QE
= 3 [-] (6.5)
4
E
Where: Q = Discharge [m3/s]
E = specific hydraulic energy of machine [J/kg]
n = rotational speed of the turbine [t/s]
nQE is known as specific speed. These parameters characterise any turbine.
As some old and non-standard definitions are still in use, the following conversion factors are given
hereafter:
ν = 2.11 ⋅ n QE
(6.6)
n Q
= 333 ⋅ n QE
(6.7)
n s
= 995 ⋅ n QE
(6.8)
Equation 6.8 corresponds to the ns definition calculated with SI units.
168
Guide on How to Develop a Small Hydropower Plant ESHA 2004
Figure 6.21 shows four different designs of runners and their corresponding specific speeds,
optimised from the efficiency viewpoint. The lower the specific speed, the higher the corresponding
head.
D0
ns= 80
Ds
D0
ns= 200
Ds
D0
ns= 300
Ds
D
ns= 514
Figure 6.21: Design of turbine runners in function of the specific speed ns
In general turbine manufacturers denote the specific speed of their turbines. A large number of
statistical studies on a large number of schemes have established a correlation of the specific speed
and the net head for each type of turbine. Some of the correlation formulae are graphically
represented in figure 6.22.
Pelton (1 nozzle) n QE
= 0.0859
0.243 (Siervo and Lugaresi) [-] (6.9)
Hn
Francis n QE
= 1.924
0.512 (Lugaresi and Massa) [-] (6.10)
Hn
Kaplan n QE
= 2.294
0.486 (Schweiger and Gregory) [-] (6.11)
Hn
Propeller n QE
= 2.716
0.5 (USBR) [-] (6.12)
Hn
Bulb n QE
= 1.528
0.2837
(Kpordze and Warnick) [-] (6.13)
Hn
Once the specific speed is known the fundamental dimensions of the turbine can be easily
estimated. However, one should use these statistical formulae only for preliminary studies as only
manufacturers can give the real dimensions of the turbines.
169
Guide on How to Develop a Small Hydropower Plant ESHA 2004
In Pelton turbines, the specific speed increases with the square root of the number of jets. Therefore
the specific speed of a four jet Pelton (only exceptionally they do have more than four jets, and then
only in vertical axis turbines) is twice the specific speed of one jet Pelton.
Table 6.2 shows the typical specific speed of the main turbines types.
Table 6.2: Range of specific speed for each turbine type
Pelton one nozzle 0.005 ≤ nQE ≤ 0.025
Pelton n nozzles 0.5
0.005 ⋅ n ≤ nQE ≤ 0.025 ⋅ n
0.5
Francis 0.05 ≤ nQE ≤ 0.33
Kaplan, propellers, bulbs 0.19 ≤ nQE ≤ 1.55
Figure 6.23 shows the specific speed evolution function of the net head and of the turbine type.
10
Kaplan
Propeller
Bulb
1 Francis
Pelton
0.1
0.01
1 10 100 1000
Hn = E/g
Figure 6.22: Specific speed in function of the net head Hn = E/g.
In addition, some basic similarity laws are given hereafter.
2
Q t
= H ⋅D t t
[-] (6.14)
2
Q m H D m m
170
Guide on How to Develop a Small Hydropower Plant ESHA 2004
n t
= H ⋅D t m
[-] (6.15)
n m H D m t
Where t correspond to the industrial turbine and m to the laboratory model.
The following example illustrates the use of the similarity laws.
If we intend to build a model with a 1:5 scale of a turbine working with an 80 m net head at 10 m3/s
and running at 750 rpm, then to test it under a net head of 10 m, the model discharge will be
0.141 m3/s and its rotational speed 1'326 rpm.
Another example is the case where a turbine would be designed for 120 net Head at 1 m3/s, and 750
rpm, but is now used under 100 m net head. In this case Dt = Dm. In order to work properly, the
turbine should have a rotational speed of 685 rpm and the maximum admissible flow would be
0.913 m3/s.
6.2.3 Preliminary design
This chapter will give some statistical formulae allowing for the determination of the main
dimensions of the turbine runner for Pelton, Francis and Kaplan turbines.
It has to be remembered that the turbine design is an iterative process depending on miscellaneous
criterion as cavitation limits, rotational speed, specific speed, etc. (see chapter 6.1.4). Clearly, it
means that after using the following equation, one has to control that the preliminary designed
turbine complies with the above-mentioned criterion.
For all turbine types, the first step is to choose a rotational speed.
Pelton turbines
If we know the runner speed its diameter can be estimated by the following equations:
H
D = 0.68 ⋅
1 n
n
[m] (6.16)
Q 1
B = 1.68 ⋅ ⋅ [m] (6.17)
H
2
n jet n
Q 1
D e
= 1.178 ⋅
n jet
⋅
gH
[m] (6.18)
Where n is the rotational speed in t/s and njet, the number of nozzles.
D1 is defined as the diameter of the circle describing the buckets centre line. B2 is the bucket width,
mainly depending on the discharge and number of nozzles. De is the nozzle diameter.
As a general rule, the ratio D1/ B2 must always be greater than 2.7. If this is not the case, then a new
calculation with a lower rotational speed or more nozzles has to be carried out.
171
Guide on How to Develop a Small Hydropower Plant ESHA 2004
The discharge function of the nozzle opening Cp - in one jet turbine the total discharge – can be
estimated according to the following formulae:
2
Q = K ⋅ π ⋅ D4
jet v
e
⋅ 2 ⋅ gH [m3/s] (6.19)
Where Kv is given in the figure 6.23 function of the relative opening a = Cp/De.
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1 1.2
Cp/De
Figure 6.23: Nozzle characteristic
For the other dimension calculations, please refer to the De Siervo and Lugaresi article 10.
Francis turbines
Francis turbines cover a wide range of specific speeds, going from 0.05 to 0.33 corresponding to
high head and low head Francis respectively.
Figure 6.24 shows schematically a cross section of a Francis runner, with the reference diameters
D1, D2 and D3.
Figure 6.24: Cross section of a Francis
Runner
172
Guide on How to Develop a Small Hydropower Plant ESHA 2004
The de Siervo and de Leva11 and Lugaresi et Massa13 articles, based on a statistical analysis of more
than two hundred existing turbines, enables a preliminary design of the Francis Turbine. As with all
statistical analysis, the results will not be sufficient on their own for complete turbine design. They
only correspond to standard average solutions, particularly if we consider the cavitation criterion
(see chapter [Link]).
The outlet diameter D3 is given by equation 6.20.
H
D 3
= 84.5 ⋅ (0.31 + 2.488 ⋅ n QE) ⋅ n
60 ⋅ n
[m] (6.20)
The inlet diameter D1 is given by equation 6.21
D = (0.4 + 0.095
1
n
)⋅D
QE
3
[m] (6.21)
The inlet diameter D2 is given by equation 6.22 for nQE > 0.164
D = D 3
[m] (6.22)
0.96 + 0.3781 ⋅ n QE
2
For nQE < 0.164, we can consider than D1 = D2
For the other dimension calculations, please refer to the above-mentioned articles.
Kaplan turbines
The Kaplan turbines exhibit much higher specific speeds than Francis and Pelton.
Figure 6.25: Cross section of a Kaplan turbine
173
Guide on How to Develop a Small Hydropower Plant ESHA 2004
In the preliminary project phase the runner outer diameter De can be calculated by the equation
6.23.
H
D e
= 84.5 ⋅ (0.79 + 1.602 ⋅ n QE) ⋅ n
60 ⋅ n
[m] (6.23)
The runner hub diameter Di can be calculated by the equation 6.24.
D = (0.25 + 0.0951
i
n
)⋅D
QE
e
[m] (6.24)
For the other dimensions calculation, please refer to the De Siervo and De Leva12 or Lugaresi and
Massa14 articles.
6.2.4 Turbine selection criteria
The type, geometry and dimensions of the turbine will be fundamentally conditioned by the
following criteria:
• Net head
• Range of discharges through the turbine
• Rotational speed
• Cavitation problems
• Cost
As previously expressed, the preliminary design and choice of a turbine are both iterative processes.
Net head
The gross head is well defined, as the vertical distance between the upstream water surface level at
the intake and the downstream water level for reaction turbines or the nozzle axis level for impulse
turbines.
As explained in chapter 6.1.1, equation 6.4, the net head is the ratio of the specific hydraulic energy
of machine by the acceleration due to gravity. This definition is particularly important, as the
remaining kinetic energy in low head schemes cannot be neglected.
The first criterion to take into account in the turbine's selection is the net head. Table 6.3 specifies
the range of operating heads for each type of turbine. The table shows some overlapping, as for a
certain head several types of turbines can be used.
174
Guide on How to Develop a Small Hydropower Plant ESHA 2004
Table 6.3: Range of heads
Turbine type Head range in metres
Kaplan and Propeller 2 < Hn < 40
Francis 25 < Hn < 350
Pelton 50 < Hn < 1'300
Crossflow 5 < Hn < 200
Turgo 50 < Hn < 250
Discharge
A single value of the flow has no significance. It is necessary to know the flow regime, commonly
represented by the Flow Duration Curve (FDC) 12 as explained in chapter 3, sections 3.3 and 3.6.
Figure 6.26: Turbines' type
field of application.
175
Guide on How to Develop a Small Hydropower Plant ESHA 2004
The rated flow and net head determine the set of turbine types applicable to the site and the flow
environment. Suitable turbines are those for which the given rated flow and net head plot within the
operational envelopes (figure 6.26). A point defined as above by the flow and the head will usually
plot within several of these envelopes. All of those turbines are appropriate for the job, and it will
be necessary to compute installed power and electricity output against costs before making a
decision. It should be remembered that the envelopes vary from manufacturer to manufacturer and
they should be considered only as a guide.
As a turbine can only accept discharges between the maximal and the practical minimum, it may be
advantageous to install several smaller turbines instead of one large turbine. The turbines would be
sequentially started, so that all of the turbines in operation, except one, will operate at their nominal
discharges and therefore will have a high efficiency. Using two or three smaller turbines will mean
a lower unit weight and volume and will facilitate transport and assembly on the site. Sharing the
flow between two or more units will also allow for higher rotational speed, which will reduce the
need for a speed increaser.
In case of strong flow variation in the range of medium head, a multi-jet Pelton with a low
rotational speed will be preferred to a Francis turbine. A similar remark can also be made for
Kaplan and Francis in low heads.
The final choice between one or more units or between one type of turbine or another will be the
result of an iterative calculation taking into account the investment costs and the yearly production.
Table 6.4: Flow and head variation acceptance
Turbine type Acceptance of Acceptance of
flow variation head variation
Pelton High Low
Francis Medium Low
Kaplan double regulated High High
Kaplan single regulated High Medium
Propeller Low Low
Specific speed
The specific speed constitutes a reliable criterion for the selection of the turbine, without any doubt
more precise than the conventional enveloping curves, just mentioned.
If we wish to produce electricity in a scheme with 100-m net head and 0.9 m3/s, using a turbine
directly coupled to a standard 1500-rpm generator we should begin by computing the specific speed
according equation (6.5).
n QE
= 0.135
176
Guide on How to Develop a Small Hydropower Plant ESHA 2004
With this specific speed the only possible selection is a Francis turbine. Otherwise if we accept the
possibility of using a lower speed, it could be possible to select, in addition to the Francis, a 4-
nozzles Pelton with 600-rpm generator.
If we intend to install a turbine in a 400 m head, 0.42 m3/s scheme, directly coupled to a 1000-rpm
generator, we will begin computing the specific speed:
n QE
= 0.022
Which indicates the 1 jet Pelton option, with a diameter D1 = 0.815 m according to equation (6.15).
A two or more jet Pelton would also be possible if required by a highly variable flow requiring a
good efficiency at part load.
As previously explained, the Pelton turbines are generally defined by the D1/B2 ratio rather than by
the specific speed. As a general rule, this ratio has to be higher than 2.7. Such a ratio cannot be
obtained without model laboratory developments.
Cavitation
When the hydrodynamic pressure in a liquid flow falls below the vapour pressure of the liquid,
there is a formation of the vapour phase. This phenomenon induces the formation of small
individual bubbles that are carried out of the low-pressure region by the flow and collapse in
regions of higher pressure. The formation of these bubbles and their subsequent collapse gives rise
to what is called cavitation. Experience shows that these collapsing bubbles create very high
impulse pressures accompanied by substantial noise (in fact a turbine undergoing cavitation sounds
as though gravel is passing through it). The repetitive action of such collapse in a reaction turbine
close to the runner blades or hub for instance results in pitting of the material. With time this pitting
degenerates into cracks formed between the pits, and the metal is snatched from the surface. In a
relatively short time the turbine is severely damaged and will need to be shut-off and repaired - if
possible.
However cavitation is not a fatality. Laboratory developments allow for a proper hydraulic design
to be defined and the operating field of the turbines to be fixed, which can both help in avoiding this
problem.
Cavitation is characterised by the cavitation coefficient σ (Thoma's coefficient) defined according
to IEC 60193 standard as:
σ = NPSE [-] (6.25)
gH n
Where NPSE is the net positive suction energy defined as:
2
NPSE = Patm −Pv + v − gHs [-] (6.26)
ρ 2
177
Guide on How to Develop a Small Hydropower Plant ESHA 2004
Where: Patm = atmospheric pressure [Pa]
Pv = water vapour pressure [Pa]
ρ = water specific density [kg/m3]
g = acceleration due to gravity [m/s2]
V = outlet average velocity
Hn = net head [m]
Hs = suction head [m]
To avoid cavitation, the turbine should be installed at least at the Hs as defined by equation (6.27).
2
Hs = Patm −Pv + v − σ⋅H [m] (6.27)
ρ⋅g 2⋅g n
A positive value of Hs means that the turbine runner is over the downstream level, a negative value
that it is under the downstream level.
As a first approach, one can consider that V = 2 m/s.
The Thoma's sigma is usually obtained by a model test, and it is a value furnished by the turbine
manufacturer. The above-mentioned statistical studies also relate Thoma's sigma with the specific
speed. These specify the equation giving σ as a function of nQE for the Francis and Kaplan turbines:
2
Francis
1.41
σ = 1.2715 ⋅ n QE + v [-] (6.28)
2⋅g⋅Hn
Kaplan
1.46
σ = 1.5241 ⋅ n QE + v [-] (6.29)
2⋅g⋅Hn
It must be remarked that Patm decreases with the altitude, from roughly 1.01 bar m at the sea level to
0.65 bar at 3000 m above sea level. So then a Francis turbine with a specific speed of 0.150,
working under a 100 m head (with a corresponding σ = 0.090), that is in a plant at sea level, will
require a setting of
2
Hs = 101'000 − 880 + 2 − 0.09 ⋅ 100 = 1.41 [m]
1000 ⋅ 9.81 2⋅9.81
installed in a plant at 2000 m above the sea level will require
2
9'440 − 880 + 2 − 0.09 ⋅ 100 = - 0.79
Hs = 71000 [m]
⋅ 9.81 2⋅9.81
a setting requiring an excavation.
178
Guide on How to Develop a Small Hydropower Plant ESHA 2004
Figure 6.27 gives an overview of cavitation limits.
10
Kaplan
Francis
1
σ
0.1
0.01
0.01 0.1 1
nQE
Figure 6.27: Cavitation limits
Equation 6.30 gives a mean to control the concordance between specific speed nQE and cavitation
limits.
n QE
≤ 0.686 ⋅ σ0.5882 [-] (6.30)
It has to be noted that local cavitation can occur on Pelton buckets if the inlet edge is not properly
designed or if the laboratory tested shape has not been fully respected during manufacture.
Rotational speed
According to equation 6.5 the rotational speed of a turbine is directly linked to its specific speed,
flow and net head. In the small hydro schemes standard generators should be installed when
possible, so in during turbine selection it must be considered that the generator, either coupled
directly or through a speed increaser to the turbine, should reach the synchronous speed, as given in
table 6.5.
179
Guide on How to Develop a Small Hydropower Plant ESHA 2004
Table 6.5: Generator synchronisation speed
Number Frequency Number Frequency
of poles of poles
50 Hz 60Hz 50 Hz 60 Hz
2 3000 3600 16 375 450
4 1500 1800 18 333 400
6 1000 1200 20 300 360
8 750 900 22 272 327
10 600 720 24 250 300
12 500 600 26 231 377
14 428 540 28 214 257
Runaway speed
Each runner profile is characterised by a maximum runaway speed. This is the speed, which the unit
can theoretically attain in case of load rejection when the hydraulic power is at its maximum.
Depending on the type of turbine, it can attain 2 or 3 times the nominal speed. Table 6.3 shows this
ratio for miscellaneous turbines.
It must be remembered that the cost of both generator and eventual speed increaser may be
increased when the runaway speed is higher, since they must be designed to withstand it.
Table 6.6: Runaway speeds of turbines
Turbine type Runaway speed nmax/n
Kaplan single regulated 2.0 - 2.6
Kaplan double regulated 2.8 - 3.2
Francis 1.6 – 2.2
Pelton 1.8 – 1.9
Turgo 1.8 – 1.9
180
Guide on How to Develop a Small Hydropower Plant ESHA 2004
6.2.5 Turbine efficiency
It is really important to remember that the efficiency characterises not only the ability of a turbine to
exploit a site in an optimal manner but also its hydrodynamic behaviour.
A very average efficiency means that the hydraulic design is not optimum and that some important
problems may occur (as for instance cavitation, vibration, etc.) that can strongly reduce the yearly
production and damage the turbine.
Each power plant operator should ask the manufacturer for an efficiency guarantee (not output
guarantees) based on laboratory developments. It is the only way to get insurance that the turbine
will work properly. The origin of the guarantees should be known, even for very small hydro
turbines.
Figure 6.28 shows an example of a real site developed without efficiency guarantees and laboratory
works.
0.90
0.89
0.88
0.87
0.86
0.85
0.84
0.83 With laboratory
η 0.82 development
0.81 Real turbine without
0.80 laboratory development
0.79
0.78
0.77
0.76
0.75
0.74
0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000
Q/Qn
Figure 6.28: Efficiency measurement on a real turbine built without laboratory development.
For the owner who wishes to control the output of a turbine, two methods are available:
The first is to carry out on site tests after putting the turbine into service. In order to obtain
adequate measurement precision, elaborate techniques, which are difficult to implement and which
most often are not suitable for small installations must be used. It is therefore generally necessary to
resort to simpler methods, the results of which are always questionable. If the tests demonstrate that
guaranteed output is not achieved, it is usually too late to improve the machine. Payment, by the
181