Energy 133 (2017) 1027e1034

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Minimising the air demand of micro-hydro impulse turbines in

counter pressure operation
M. Kramer, S. Wieprecht*, K. Terheiden
Institute for Modelling Hydraulic and Environmental Systems, University of Stuttgart, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Following new and innovative concepts of energy recovery which use micro-hydro impulse turbines in
Received 18 December 2015 drinking water systems, more sophisticated turbine designs become essential to improve energy re-
Received in revised form covery efficiency. In the case of implementing a hydraulic impulse turbine with tailwater depression in
20 April 2017
such systems, one major challenge to optimising overall efficiency is the reduction of the turbine's air
Accepted 6 May 2017
demand. Firstly, a lower air demand would reduce the required dimensions of the ventilation system and
Available online 9 May 2017
therefore increase energy efficiency. Secondly, a minimised air demand would considerably reduce
negative effects downstream of the turbine casing, such as a reduced transport capacity and corrosion. To
Keywords:
Pelton turbine
achieve a minimised air demand, detailed experimental investigations are conducted, during which
Micro-hydropower different casing inserts are tested in a micro-hydro Pelton machine. A specially constructed turbine test-
Counter pressure operation bed, featuring two separately installed measuring units, allows for the partitioning and measurement of
Air demand dissolved and undissolved air. In the case of the used Pelton machine, the results clearly identify different
Air detrainment air detrainment processes. The undissolved air demand strongly depends on the particular casing insert
Casing setup and can be reduced by 90% through optimal flow condition using flow straighteners as insert. In contrast,
the amount of dissolved air demand cannot be controlled by additional casing inserts. This can be
explained due its different entrainment processes. The analysis of the dependency of the measurement
results on scale shows that a direct transfer to a prototype application will underestimate the air demand.
Nevertheless, the presented research findings show an effective way, also relevant to turbine manu-
facturers, to reduce air demand and improve overall efficiency of impulse turbines in counter pressure
operation.
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction pressure (B) or direct feedings into a supply zone (D).

In the last few decades, in order to improve the energy efficiency
of drinking water supply networks, energy recovery units have
increasingly been deployed. Fig. 1 shows possible locations of en-
ergy recovery units within water supply systems. A distinction is
made between open and closed systems: open systems are pipe
networks incorporating spring discharges (A), pressure chambers
(C) and tank overflows (E). The liquid within those systems is
exposed to atmospheric pressure as there is a free surface in the
tailwater of the hydraulic turbine. By contrast, the supply pressure
of closed systems is not exposed to atmospheric pressure within
the downstream network. Examples are tank inlets with counter
* Corresponding author.
E-mail address: silke.wieprecht@iws.uni-stuttgart.de (S. Wieprecht).
URL: http://www.iws.uni-stuttgart.de
Commonly implemented hydraulic machines are Francis tur-
bines, reverse running pumps and impulse (Pelton) turbines. Based
on the working principle that the runner can be partially loaded,
Pelton turbines have a particular efficiency advantage when
compared to other machines, especially under varying flow con-
ditions. Current knowledge of Pelton turbines, for example on
erosion mechanisms in the buckets [1] and their influence on tur-
bine performance [2], relies upon experimental and numerical
investigations.
However, conventional designs of Pelton turbines only allow
implementation into open systems. In order to facilitate imple-
mentations of Pelton turbines in closed systems independent of the
back water level conditions, the counter pressure operation of
micro-hydro Pelton turbines was developed in 2004 [3]. In this
operation mode, which was primarily used in large scale Pelton
arrangements [4], the runner of the hydraulic impulse turbine is
located below the downstream water level. In order to ensure

http://dx.doi.org/10.1016/j.energy.2017.05.043
0360-5442/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1028 M. Kramer et al. / Energy 133 (2017) 1027e1034

Nomenclature k empirical factor, ½

km velocity coefficient, ½
B bucket width, ½m L characteristic length, ½m
Bi distance of the bucket to the casing top, ½m M quantity in model scale
c0 jet velocity at nozzle exit, ½m=s N quantity in prototype scale
c2 absolute velocity at bucket outlet, ½m=s p pressure, ½bar
c2m meridian component, ½m=s pG counter pressure, ½bar
c2u perimetral component, ½m=s Qa air volume flow, ½min1 
dG casing diameter, ½m Qw water volume flow, ½m3 =s
Fr Froude-number, ½ R relative quantity ¼ prototype/model quantity
g gravity acceleration, ½m2 =s Re Reynolds-number, ½
h water level, ½m u perimetral velocity, ½m=s
hf setting height, ½m ub bubble rise velocity, ½m=s
hG casing height, ½m We Weber-number, ½
hA outlet diameter, ½m q flow rate parameter, ½
H fall head, ½m fB specific flow rate, ½
HP penetration depth, ½m u angular speed, ½1=s

Fig. 1. Possible locations of energy recovery units; adopted from Ref. [14]. A. spring Fig. 2. Differentiation between external and internal air demand; figure adopted from
discharge B. tank inlet with counter pressure C. pressure chamber D. direct feeding E. Ref. [14].
tank overflow.

nitrogen, oxygen, carbon dioxide) in water. However, the amount of

successful operation under these conditions, an air cushion is
induced in the pressure-tight casing. Currently, Switzerland has
integrated approximately 20 counter pressure Pelton turbines
(GDPT) into its water supply systems, mainly at tank inlets with
counter pressure (B). First experiences indicate further potential for
an increase in energy efficiency and an improvement of the general
performance of a GDPT in such systems.
For energy and performance improvements of GDPT in closed
water supply systems, the minimisation of the air detrainment
during backpressure operation is still a major challenge. The
detrained air reduces energy efficiency and may cause corrosion
and reduced transport capacity in the tailwater system. In general,
there are two air detrainment mechanisms, the undissolved and
dissolved detrainment, see Fig. 2. In the case of undissolved
detrainment, water jets or droplets emitted from the Pelton bucket
enter the casing water body and lead to the entrainment of air
bubbles, which can be flushed out of the casing. Dissolved air
detrainment results from dissolution of the air components (e.g.
dissolved air detrainment increases with higher counter pressure
due to the pressure-dependent solubility of volatiles in liquids
(Henry's law). Compared to the air detrainment of impulse tur-
bines, the air demand is defined as air quantity required by the
turbine and supplied by internal recirculation and external aera-
tion. Therefore, a further distinction is made between external and
internal air demand, see Fig. 2. In counter pressure operation, the
external air demand is defined by the difference between the air
quantity in the impulse turbine's head- and tailwater (both outside
the casing), and has to be replenished by compressors under
additional energy use. Analogue to the air detrainment, the air
demand can be expressed as undissolved and dissolved air demand.
Until now, research on the air demand of Pelton turbines mainly
focused on large-scale hydropower plants. Osterwalder [5] presents
air demand measurements of vertical shaft turbines in model and
full size scale. However, no distinction is made between recircu-
lating air in the tailwater channel and air supplied by compressors.
 M. Kramer et al. / Energy 133 (2017) 1027e1034
Ceravola and Noseda [6] investigate the external air demand of a
Pelton turbine model with respect to different water levels and
pressure heads in the tailwater channel. Here, an increase of the air
demand at lower channel levels and higher pressure heads is
observed. The maximum external air demand is at approximately
Qa =Qw ¼ 6%, where Qa is the rate of supplied air flow and Qw the
turbine discharge. The published results of the external air demand
are not directly comparable due to the different turbine di-
mensions. In contrast to Ceravola and Noseda [6], Grein and Koeller
[7] observe external air demands up to Qa =Qw ¼ 60%, depending on
the tailwater channel length. Furthermore, a comprehensive liter-
ature review on the main aspects of research in the field of air
entrainment and detrainment processes of large-scale impulse
turbines is presented in Ref. [8].
In the field of micro-hydropower, Kramer et al. [8] conduct
detailed experimental investigations on the dissolved and undis-
solved air detrainment of a micro-hydro Pelton turbine with tail-
water depression. With regard to dissolved air detrainment, the
counter pressure and turbine discharge are identified as governing
parameters, whereas velocity coefficient, casing water-level, tur-
bine discharge and geometry of the casing have a major influence
for undissolved air detrainment. One resulting option for mini-
mising the undissolved air detrainment is to apply the following
design equations for cylindrical turbine casing dimensioning:
sffiffiffiffiffiffiffiffiffiffiffiffiffi
4Qw
dG  k (1)
pub
hG  hA þ HP þ hf þ B=2 þ Bi (2)
where dG is the casing diameter, k an empirical factor related to the
velocity distribution in the casing, Qw the turbine discharge, ub the
bubble rise velocity, hG the casing height, hA the outlet diameter, HP
the penetration depth, hf the setting height of the runner, B the
bucket width and Bi the distance of the bucket to the casing top.
Based on the analysis of velocity measurements in the casing of a
prototype Pelton turbine [9], a first approximation of the empirical
factor leads to k ¼ 1.83. Future research concerning flow velocity
measurements in the casing of impulse turbines is still needed for
verification.
Recent investigations on reducing the deaeration length of a
large-scale Pelton plant, conducted on a laboratory model of the
large-scale turbine Veytaux 2 (Switzerland), showed that the
deaeration length can be reduced by flow straighteners in the
tailrace channel [10]. Therefore an air demand reduction should
also be expected if implementing additional installations, e.g. flow
straighteners, into the Pelton turbine casing of energy recovery
units. In the micro-hydro sector, no optimisation of the casing with
additional installations has been undertaken thus far. Conse-
quently, the influence of casing inserts on the undissolved and
dissolved air demand of a counter pressure Pelton turbine is
investigated within this study.
2. Experimental work
Fig. 3 shows the experimental setup containing the Pelton tur-
bine. The main pipe leads from a large basement tank over two
high-level tanks with fixed overfalls to a booster pump, which is
used to create the required energy head at the turbine inlet. The
investigated turbine is a two-jet counter pressure Pelton turbine
(GDPT) with a vertical axis, a design fall head of H ¼ 60 m and a
design discharge of Qw ¼ 4.3 l=s. The test turbine is provided by the
Revita GmbH. It is a two-jet counter pressure Pelton turbine with a
double-pole asynchronous induction machine (type: HU 90 LP2,
1029
Rüetschi). The electrical machine is connected to the runner with a
belt drive (transmission ratio: 0.73). The runner consists of 22
buckets and has a pitch circle of 140 mm diameter. The air cushion
in the turbine casing is realised with a compressor (type: Mahle
KMK 250 W, max. 10 bar). The water level in the turbine casing is
regulated by a Programmable Logic Controller (PLC), which is
steering two air valves implemented in the connection tube be-
tween compressor and turbine casing. Further details of the test
turbine are published in Ref. [8]. The flow rate of the turbine is
measured with the magnetic flowmeters Q1 to Q3, while the
pressure is recorded at different locations of the setup using pres-
sure gauges p1 to p5. Furthermore, the counter pressure Pelton
turbine is equipped with a rotational speed sensor (n) and a power
meter (P).
The undissolved air detrainment is quantified in a measurement
section located in the tailwater of the Pelton turbine. This mea-
surement section contains two deaeration chambers (AK1 and AK2)
which work according to the displacement principle. As soon as air
bubbles reach a deaeration chamber, they ascend due to their lower
density and displace the equivalent volume of water. To measure
the corresponding water level decrease, the deaeration chambers
are equipped with ultrasonic sensors (h3 and h4). During mea-
surements, undissolved air accumulates primarily in the first
deaeration chamber while the second chamber has mainly a con-
trolling function, indicating whether the whole amount of air is
separated in the first chamber. The first chamber contains a
piezoelectric pressure transducer (p4) and a thermo couple (T),
which determine the essential parameters used for calculating the
normalised undissolved air detrainment with the ideal gas law.
Since the feed water of the turbine generally does not contain
undissolved air bubbles, the undissolved external air demand cor-
responds to the undissolved air detrainment and is detected based
on the gradient (dh=dt) of the water level depression in the first
deaeration chamber. Conversely, dissolved air is present in the
head- and tailwater of the turbine. The dissolved external air de-
mand is measured by two oxygen meters (O2 1 and O2 2) installed at
the inlet and the outlet section of the hydraulic Pelton turbine.
Table 1 gives a survey of the measuring devices implemented in the
test-bed, including error analysis. As high quality and reliability of
the measured data is the basis for consolidated findings, sensors
with low errors ranges are installed. Further details of validation of
the measurement section are presented in Ref. [8].
2.1. Casing configurations
The casing of the counter pressure Pelton turbine is modified
with three different inserts denoted as configurations C1 to C3:
 C1: No additional casing inserts
 C2: Implementation of a baffle plate mounted on three de-
flectors. Due to the baffle plate, the direct impingement of water
jets or droplets from the Pelton bucket on the tailwater body
inside the casing is prevented.
 C3: Implementation of flow straighteners with honeycomb-like
structure. Due to the flow straighteners, the water flow is less
turbulent.
Fig. 4 shows sectional and top views as well as the particular
inserts of the different casing configurations. The height of the
casing inserts is denoted as z.
2.2. Test series
The influence of casing inserts (configurations C2 and C3) on the
1030 M. Kramer et al. / Energy 133 (2017) 1027e1034

Fig. 3. Sketch of the experimental setup.

Table 1 the counter pressure of the particular test series is indicated in the
Implemented measurement devices in the test-bed. last column. The measuring principle of the deaeration chambers
Name Device Measurand Error
requires a minimum overpressure, and therefore, a pressure of
pG ¼ 0.1 bar is applied throughout test series 1, 2 and 3. The upper
Q1 ABB Copa-XL Qw : Discharge ± 0.5%
Q2 E þ H Promag 50 W Qw : Discharge ± 0.5%
part of the table shows all test series of the undissolved air demand
Q3 E þ H Promag 50 W Qw : Discharge ± 0.5% measurement (test series index a), while the lower part refers to
p1 Jumo MIDAS 401001 p: Pressure ± 0.5% dissolved air demand measurements (test series index b). Since
p2 Jumo MIDAS 401001 p: Pressure ± 0.5% counter pressure and turbine discharge are found to be the gov-
p3 Keller PD-39X p: Pressure  0.1%
erning parameters with regards to dissolved air detrainment [8], a
p4 Keller PA-21Y p: Pressure ± 0.25%
p5 Keller PA-21Y p: Pressure ± 0.25% comparison of different test series was only made between con-
P EMU 32.61.2Q P: Power ± 2.0% figurations with and without installations (configurations C1 and
T Thermal element Type T T: Temperature ± 0.5 C C3).
O2 1 E þ H COS51D O2: Dissolved Oxygen ± 1.0%
O2 2 Hach-Lange LXV416 O2: Dissolved Oxygen ± 0.2 mg=l
h3 Pepperl þ Fuchs UB500 h: Water-level (chamber) ± 1.0% 3. Results and discussion
h4 Pepperl þ Fuchs UB500 h: Water-level (chamber) ± 1.0%
h1 Capacitive sensor h: Water-level (casing) e 3.1. Undissolved external air demand
h2 Capacitive sensor h: Water-level (casing) e
N ifm IY5052 n: Rotational speed e
Series 1a: variation of the velocity coefficient km . The results of the
test series 1a for the configuration C2 are shown in Fig. 5. The
undissolved external air demand is plotted as dimensionless air-
external air demand of a counter pressure Pelton turbine is inves- water volume flow ratio on the primary ordinate against the ve-
tigated in different test series, where the operating parameters of locity coefficient of the turbine. Each measurement is repeated four
the impulse turbine are varied according to Table 2. These param- times (n ¼ 4) and fluctuation ranges are indicated. Furthermore, the
eters are the velocity coefficient km ¼ cu0 ¼ pffiffiffiffiffiffiffi
u , the flow rate
absolute velocity c2 at the bucket outlet and the meridian compo-
2gH

parameter q ¼ pQRw3 u, the casing water level h and the counter pres-
sure pG . In the equations, u is the perimetral speed on the pitch
circle, c0 the water velocity at the nozzle exit, g the gravitational
acceleration, H the fall head, Qw the turbine discharge, R the radius
of the runner and u the angular speed. The results are also
compared to the measured data of casing setup C1, which has been
presented in Ref. [8].
During a test series, stationary hydraulic conditions are fulfilled
and all operating parameters, besides the one altered, are kept
constant. Table 2 gives an overview of the conducted test series. The
altered operating parameters are listed in the second column and
nent c2m as well as the perimetral component c2m are calculated
based on the vector outlet velocities and plotted on the secondary
ordinate.
The measured results of the configuration C2 show a bell-
shaped dependency of the volume flow ratio on the velocity coef-
ficient with a maximum detrainment at a velocity coefficient of km
z 0.47. This is due to the perimetral component (c2u ) of the ab-
solute velocity having a minimum at the best point of operation at
km z 0.47. At this point of operation, water jets impinge in general
vertically on the surface of the water body inside the casing. As a
result the highest amount of undissolved air is entrained and
consequently detrained out of the casing. A similar bell-shaped
 M. Kramer et al. / Energy 133 (2017) 1027e1034 1031

Fig. 4. Different casing setups of the counter pressure Pelton turbine (GDPT).

Table 2
Conducted test series.

altered parameter C2 C3 pG [bar]

undissolved velocity coefficient km 1a 1a 0.1

flow rate parameter q 2a 2a 0.1
water level h 3a 3a 0.1
counter pressure pG 4a 4a 0.1; 1.0
dissolved velocity coefficient km e 1b 0.1
flow rate parameter q e 2b 0.1
water level h e 3b 0.1
counter pressure pG e 4b 1.0; 1.5; 2.0

dependency of the volume flow ratio on the velocity coefficient was

observed in investigations without additional installations [8].
The dependency of the undissolved air demand of configuration
C3 on the velocity coefficient is investigated at three different flow
rate parameters of q ¼ 0.0064, 0.0076 and 0.0099 and at a casing
water level of h ¼ 26.6 cm. The conditions of configuration C3 and
C2 are slightly different, mainly in terms of the casing water level.
This is done to clearly represent the dependency of undissolved air
demand on the velocity coefficient, and the results are shown in
Fig. 6. The undissolved air demand and the components of the
absolute velocity are plotted against the velocity coefficient. In
contrast to configuration C2, a different dependency is observed.
The undissolved air demand decreases with higher velocity co-
efficients and is proportional to the meridian velocity component:
Qa =Qw fc2m (3)
This behaviour can be explained by the effect of the casings
insert's honeycomb-like structure. Due to this structure, the cas-
ing's internal flow is straightened vertically and the perimetral
velocity component c2u is therefore damped. As a result, the me-
ridian velocity component c2m governs the air demand of the im-
pulse turbine with casing setup C3.
Series 2a: Variation of the flow rate parameter q. In series 2a, the
discharge of the turbine, expressed as dimensionless flow rate
parameter q, is varied at a constant velocity coefficient of km ¼ 0.47
(point of best operation). In general, hydraulic impulse turbines
have an efficiency maximum at km ¼ 0.45 to 0.48. The air demand
measurements of the configurations C2 and C3 are undertaken at a
1032 M. Kramer et al. / Energy 133 (2017) 1027e1034
Fig. 5. C2 Series 1a - relation of the undissolved air detrainment (air demand) and the
velocity coefficient km ; c2 : absolute velocity at bucket outlet; c2u : perimetral compo-
nent; c2m : meridian component; h ¼ 34.6 cm; pG ¼ 0.1 bar; n ¼ 4.
Fig. 6. C3 Series 1a - relation of the undissolved air detrainment (air demand) and the
velocity coefficient km ; c2 : absolute velocity at bucket outlet; c2u : perimetral compo-
nent; c2m : meridian component; h ¼ 26.6 cm; pG ¼ 0.1 bar; u ¼ const.; n ¼ 4.
constant casing water level of h ¼ 32.6 cm for more accurate
comparison. The results are shown in Fig. 7, together with results of
configuration C1 (no casing inserts).
The results clearly show that the undissolved air demand in-
creases with increasing flow rate. This is due to higher flow ve-
locities and therefore higher flow forces acting on entrained air
bubbles. In comparison to configuration C1, the undissolved air
demand of configuration C2 is reduced by approximately 50% at
q ¼ 0.0087 and q ¼ 0.0099. Furthermore, the implementation of
flow straighteners (C3) leads to a reduction of the undissolved air
demand of about 93% to 99% at the investigated flow rates. These
results could additionally be confirmed through visual flow
observation. During measurements with configuration C3, a
Fig. 7. Series 2a - relation of the undissolved air detrainment (air demand) and the
flow rate parameter q; km ¼ 0.47; pG ¼ 0.1 bar; u ¼ const.; n ¼ 4.
considerable less turbulent flow was observed inside the turbine
casing.
Series 3a: Variation of the casing water level h. The water level h
above the casing bottom is adjusted by positioning two capacitive
water level sensors at a hydraulically connected stand pipe. For the
measurements of the series 3a, different water levels of h ¼ 30.6 cm,
32.6 cm and 34.6 cm are adjusted while the flow rate is kept con-
stant at q ¼ 0.0099. Fig. 8 shows the relation of the measured un-
dissolved air demand and the casing water level for the
configurations C1 to C3. The results demonstrate a strong influence
of the water level on the undissolved air demand, which increases
at lower water levels. Comparing the different configurations, the
undissolved air demand of the configurations C2 and C3 is signifi-
cantly reduced. In case of configuration C3, a reduction of more
than 90% is achieved at all investigated water levels.
Series 4a: Variation of the counter pressure pG . Published mea-
surements of the undissolved air detrainment (configuration C1) at
different counter pressure levels showed no dependency of the
undissolved air detrainment on the counter pressure in a pressure
range up to 1.0 bar overpressure [8]. Those measurements are
repeated with configuration C2 and C3 of the counter pressure
Pelton turbine. Therefore, pressure levels of 0.1 bar and 1.0 bar
overpressure are applied to the casing. The undissolved air
detrainment is measured at different water levels while the flow
rate parameter and the velocity coefficient are held constant. The
measured results of configuration C2 (Fig. 9) and configuration C3
(not presented) also show no dependency of the investigated
pressure level on the undissolved air demand, confirming the
measurements of configuration C1.
3.2. Dissolved external air demand
The external dissolved air demand is determined through dis-
solved oxygen measurements carried out on the configuration C3 of
the counter pressure Pelton turbine. This approach for determining
dissolved air is in accordance with common practice, see Ref. [11].
As turbulence and phase interface are diminished through the
implementation of flow straighteners, it can be expected that dis-
solved air demand of configuration C3 is reduced.
Series 2b: Variation of the flow rate parameter q. Fig. 10 shows the
dependency of the dissolved air demand on the flow rate param-
eter. The dissolved air contents in both the head- and the tailwater
of the turbine are plotted as dissolved oxygen mass flows (O21 and
O22 ) against the flow rate parameter. On the secondary ordinate, the
measured pressures in the turbine casing (p3) and at the oxygen
meter (p5) are presented. It shows pressure losses in the tailwater
pipeline system. At higher flow rate parameters, the dissolved
Fig. 8. Series 3a - relation of the undissolved air detrainment (air demand) and the
casing water level h; km ¼ 0.47; pG ¼ 0.1 bar; n ¼ 4.
 M. Kramer et al. / Energy 133 (2017) 1027e1034 1033

higher pressures. The oxygen saturation in the tailwater reaches

values of SO2 ¼ 84% and remains constant at the investigated
pressure levels. Control samples taken at the test bed show a
maximum oxygen saturation of SO2 z 91%. Full saturation is almost
reached, with some minor deviations due to the impurities of the
water.
However, a noticeable reduction of dissolved air demand could
not be observed during the measurements. A possible reason is that
the water in the turbine casing is still turbulent enough to reach
saturation. As the results of configuration C3 align with the results
of configuration C1 [8], only test series 2b and 4b have been pre-
sented exemplarily.

3.3. Comparison of undissolved and dissolved air demand

Fig. 9. C2 Series 4a - relation of the undissolved air detrainment (air demand) and the
counter pressure pG ; km ¼ 0.47; q ¼ 0.0087; n ¼ 4.
In order to compare the undissolved and dissolved air demand
of the counter pressure Pelton turbine, the different amounts of
dissolved oxygen mass flows are converted to volume flows and
further components of air (N2 , Ar and CO2 ) are calculated based on
the measured oxygen content. Fig. 12 shows the dimensionless
undissolved and dissolved air demands of the configuration C3
denoted as Qa;u =Qw and Qa;d =Qw . The indices u and d indicate un-
dissolved and dissolved, respectively. The dissolved air demand is
constant over the flow rate parameter with Qa;d =Qw ¼ 1.30% (blue).
However, as dimensionless parameters are considered, the absolute
dissolved air demand increases with higher flow rates. The undis-
solved air demand reaches a maximum of Qa;u =Qw ¼ 0.23% (black)
at q ¼ 0.0010 and thus is five times lower than dissolved air
demand.
Fig. 10. C3 Series 2b - relation of the dissolved air demand and the flow rate parameter
q; pG ¼ 1.0 bar; km ¼ 0.47; h ¼ 32.6 cm; u ¼ const.; n ¼ 4.

oxygen mass flow in the tailwater shows a stronger increase than in

the headwater. As dissolved air demand results from the difference
between both oxygen mass flows, the air demand increases with
the flow rate of the hydraulic Pelton turbine.
Series 4b: Variation of the counter pressure pG . Fig. 11 shows
dissolved oxygen mass flows in the head- and the tailwater of the
Pelton turbine being recorded at different pressure levels in the
casing. The investigations are carried out at counter pressures of
pG ¼ 1.0, 1.5 and 2.0 bar overpressure. The oxygen saturation in the
tailwater (SO2 ) is plotted on the secondary ordinate. The saturation
refers to the equilibrium concentration resulting from the partial
oxygen pressure and temperature in the casing. The measurements
Fig. 12. C3 - comparison of undissolved and dissolved air demand; km ¼ 0.47; h ¼ 32.6
show a linear increase of the downstream oxygen mass flow at cm; pG ¼ 1.0 bar; u ¼ const.; n ¼ 4.

Fig. 11. C3 Series 4b - relation of the dissolved air demand and the counter pressure pG ; km ¼ 0.47; q ¼ 0.0063; h ¼ 32.6 cm; n ¼ 4.
1034 M. Kramer et al. / Energy 133 (2017) 1027e1034
3.4. Scaling of measurements results
The experiments on the hydraulic impulse turbine are under-
taken in a model scale (LR ¼ 2). Therefore, scale dependent effects
have to be taken into account by ensuring similarity between
model and nature, implying geometrical, kinematic and dynamic
similarity [12]. In case of the Pelton turbine, various flow phe-
nomena have to be considered in different stages including pipe
flow, free-jet flow, unsteady flow with free surface inside the
buckets and two-phase flow in the casing. These flow regions are
controlled by different dimensionless relationships. Dynamic sim-
ilarity of the overall flow requires equality of the Froude-, Reynolds-
and Weber-numbers as well as same values of the velocity coeffi-
cient km and the specific flow rate fB [13].
In the free surface casing-flow, pressure forces and viscous
forces are inferior compared to gravity and inertia forces. It has to
be ensured that the Froude number of model and prototype are
equal:
FrN vR
FrR ¼ ¼ pffiffiffiffiffiffiffiffiffi
ffi¼1
FrM gR LR
where the index R implies relative quantity, N a quantity in pro-
totype scale and M a quantity in model scale. Fr is the Froude
number, v the velocity, g the gravitational acceleration and L a
characteristic length. As gR ¼ 1, rearrangement of Eq. (4) results in:
pffiffiffiffiffi
vR ¼ LR
An application of Eq. (5) to other dimensionless relationships
leads to smaller Reynolds- and Weber-numbers in model scale
compared to prototype scale:
ReN   ReN
1=2 3=2
¼ ReR ¼ vR ,LR ¼ LR ,LR ¼ LR /ReM ¼ 3=2
ReM LR
WeN   We
1=2 2
¼ WeR ¼ v2R ,LR ¼ LR ,LR ¼ L2R /WeM ¼ 2N
WeM LR
where Re is the Reynolds-number and We the Weber-number. Eqs.
(6) and (7) show that the influence of viscosity and surface tensions
is higher in model scale than in nature. Therefore, scaling effects
and consequently a higher air demand are to be expected in the
prototype scale by applying Froude-similitude. In case of the dis-
solved air demand, the specific interfacial area is expected to be in
the same order of magnitude in model and prototype and therefore,
the assumption of full saturation seems to be reasonable.
Although a direct scaling of the obtained measurements results
is not possible, the obtained results show new findings for an
efficient air demand reduction of existing hydropower plants
through implementation of flow straighteners. In this context, it
should be noted that the design equations for dimensioning the
casing of impulse turbines (Eqs. (1) and (2)) are also valid in pro-
totype scale and should be applied for future installations.
4. Conclusion
The minimisation of air demand is a crucial issue when seeking
to increase the energy efficiency of drinking water supply networks
using micro-hydro counter pressure Pelton turbines. For the design
process of turbines, recently devised design equations can be used
to achieve higher efficiency. Nevertheless, there is still an urgent
need to optimise already implemented systems. Therefore different
casing setups have been investigated in detail and under varying
conditions, with specific attention to the undissolved and dissolved
air demand. Results show that the implementation of flow
straighteners is an extremely powerful tool in minimising the un-
dissolved air demand by more than 90% compared to casings
without inserts. These newly obtained research results provide
industry applicable findings for efficient air demand reduction in
existing hydropower plants. In contrast, the dissolved air demand is
barely reduced through the implemented casing inserts and thus
remains a partial air demand in an otherwise optimised casing.
Based on the experiments undertaken, a pressure dependent
saturation of tailwater is found. Therefore, the presented findings
provide both a quantification of the air demand and can be utilised
already at planning stage of hydraulic impulse turbines in counter
pressure operation. This supports the appropriate design of tail-
(4)
water systems, including deaeration devices, and makes the results
extremely valuable for turbine manufacturers in minimising air
demand and improving overall efficiency of impulse turbines in
counter pressure operation.
Acknowledgments
(5)
The authors kindly thank Mr. Patrick Kamber (Revita GmbH) for
providing the test turbine. They acknowledge the fruitful discus-
sions with Dr. Zh. Zhang (Ruetschi) and Professor Stefan Rie-
delbauch (University of Stuttgart).
(6)
References
[1] Padhy MK, Sainin RP. Study of silt erosion on performance of a Pelton turbine.
Energy J 2011;36:141e7.
(7) [2] Padhy MK, Saini RP. Study of silt erosion mechanism in Pelton turbine buckets.
Energy J 2012;39:286e93.
[3] Bretscher A, Schindelholz B. Energy use in in closed water supply systems (in
German). Tech. Rep. Langenbruck, Switzerland: Revita GmbH; 2004.
[4] Dawson PD. Multi-jet impulse turbines. 1964. ASME 64-WA/FE-24.
[5] Osterwalder J. Tailwater depression of multijet impulse turbines. Water Power
1966;18(9):358e62.
[6] Ceravola O, Noseda G. Back-pressure operation of pelton-turbines for ternary
units in pumped storage plants. In: IAHR-Symposium Rome, paper C3; 1972.
[7] Grein H, Koeller HK. Pelton turbines in counter pressure operation (in
German). Escher Wyss Mittl 1981;1:32e6.
[8] Kramer M, Terheiden K, Wieprecht S. Optimized design of impulse turbines in
the micro-hydro sector concerning air detrainment processes. Energy J
2015;93:2604e13.
[9] Arch A. Air entrainment and detrainment processes in power plants with
Pelton turbines in counter pressure operation (in German). Ph.D. thesis.
Austria: Graz University of Technology; 2008.
[10] Lucas J, Lais A, Hager WH, Boes RM. Air-transport processes in tailrace
channels. J Hydraulic Eng 2015;141(7).
[11] Huber H, Strobl T. In: A model test for the investigation of gassing processes in
the pressure surge tank of Kopswerk II (in German). Symposium Graz, Austria;
2006. 2(91-106).
[12] Kobus H. Hydraulic modelling. Bull Ger Assoc Water Resour Land Improv
(DVWK) 1984:7.
[13] Grein H, Klicov D, Wieser W. Efficiency scale effects in pelton turbines. Water
Power Dam Constr 1988;40(5):32e6.
[14] Revita. Universal turbine for water supply systems (in German). Langenbruck,
Switzerland: Flyer, Revita GmbH; 2005.