Environmental Engineering and Management Journal April 2017, Vol.16, No.

4, 979-988
http://omicron.ch.tuiasi.ro/EEMJ/

“Gheorghe Asachi” Technical University of Iasi, Romania

EXPERIMENTAL INVESTIGATION AND COMPUTATIONAL FLUID

DYNAMICS (CFD) ANALYSIS OF AN ECO-FRIENDLY TURBINE

Daniela Popescu1∗, Constantin Popescu1, Andrei Dragomirescu2*, Mircea Bărglăzan3

1“Gheorghe Asachi” Technical University of Iasi, Department of Fluid Mechanics, Fluid Machines and Drives,
59A Dimitrie Mangeron Blvd, Iaşi, 700050, România
2University Politehnica of Bucharest, Department of Hydraulics, Hydraulic Machinery and Environmental Engineering,

313 Splaiul Independentei, Bucharest, 060042, România

3“Politehnica” University of Timişoara, Hydraulic Machinery Department, 1 Mihai Viteazu Blvd, Timişoara, 300222, România

Abstract

The global renewable energy share should reach and exceed 30% by 2030. Mini-scale hydropower installations are
environmentally-friendly technologies, unlike large hydroelectric power plants that have certain drawbacks from the
environmental point of view. A safe and clean solution is the cross-flow turbine, but the commercial types require a minimum
head, usually captured by additional hydro-technical works that can also damage the environment. In this context, a very good
solution could be a low head turbine that has no adverse environmental impact, since no water storage is necessary. The objective
of this paper is to investigate a cross flow pico-turbine prototype, in order to verify whether, when well designed, such a turbine
can operate with good efficiencies at very low heads. The tests revealed good efficiencies of the turbine that are in the range of
values calculated with the classical Hutton, Banki, and Sutton methods. However, unlike the classical methods, that predict an
increase in efficiency as the head increases, the results presented in this paper suggest that, for cross-flow turbines designed for
very low heads, the efficiency increases as the head decreases. This could be considered the most important result of this study.
Numerical simulations were performed to get a better insight into the air-water two-phase flow inside the turbine operating at low
heads. The results obtained suggest that air bubbles get entrapped into the blade channels during the first pass if the operating
parameters are far from the rated ones. The results could be used as a recommendation for saving the hydraulic energy that is
wasted in very low head hydraulic installations. By implementing the proposed cross-flow turbine in hybrid technologies or on
irrigation open channels, the energy is produced using a sustainable solution that is able to protect the environment.

Key words: cross-flow turbine, efficiency, environmental impact, low head hydro application, renewable energy

Received: December, 2016; Revised final: March, 2017; Accepted: April, 2017

1. Introduction technologies, by implementing energy policies at

The share of renewable energy in the gross
final energy consumption is a key indicator for
measuring the progress achieved by the
implementation of Directive 2009/28/EC of the
European Parliament and Council on the promotion
of the use of energy from renewable sources
(Directive 28, 2009). The global renewable energy
share can reach and exceed 30% by 2030 (IRENA,
2014), by applying the best existing and emerging
national level, and by using financial schemes for
further development and deployment.
Hydropower, an energy source that does not
pollute, generates approximately 3500 TWh per year,
equivalent to 15.8% of the global electricity
generation (IEA-EPSAT and IRENA, 2015). The
total installed capacity in hydroelectric power plants
was of 1060 GW in 2016 and it is expected to reach
2000 GW by the year 2050 (WEC, 2016). Most
hydropower is produced in large hydroelectric power

∗
Author to whom all correspondence should be addressed: e-mail: daniela_popescu@tuiasi.ro; Phone: +40746036497; e-mail:
andrei.dragomirescu@upb.ro, Phone: +40214029710
 Popescu et al./Environmental Engineering and Management Journal 16 (2017), 4, 979-988
plants. However, such plants require the restriction of
the natural flow of water and the flooding of a large
area to create reservoirs, accompanied by specific
drawbacks: a strong negative impact on the aquatic
life-restriction of fish migration and destruction of
vegetation-, deposition of sludge that can produce
gases that are dangerous for health, relocation of
households etc. Unlike large hydroelectric power
plants, mini-scale hydropower installations do not
have significant adverse environmental impact, while
helping to reduce greenhouse gas emissions in the
power domain (Codruta et al., 2009). In this context,
micro-scale hydropower, which harnesses energy for
electricity production mainly from the available flow
of small rivers whereby little water storage is
provided to obtain the necessary head, represents an
interesting solution for the development of
sustainable energy resources able to protect the
environment.
Lately, as a result of research work on
developing new environmental-safely technologies,
new types of turbines and technologies were
proposed. Bostan et al. (2011) designed kinetic
turbines for micro hydro stations, Udroiu et al.
(2011) proposed an innovative e-method for
integrating design with manufacturing for a small
Pelton turbine. A very low head axial turbine (VLH
Turbine, 2016) was designed for heads from 1.5 m to
4.5 m and it is already in use at different locations in
France (Rondeau, Mayenne, Moulin de Hauterive, La
Moulasse), Switzerland (Martigny en Bourg),
Romania (Roman), Italy (Vito), and Croatia (Ilovac).
Another good choice of an environmental
protection product is the cross-flow turbine. Since
1922, when Donát Bánki obtained a patent for this
type of mini turbine (Banki, 1922), improved
versions were developed and realized, such as the
Ossberger turbine (Ossberger GmbH+Co, 2016) and
the Cink turbine (Cink Hydro-Energy, 2016). Recent
studies focused on the advantages of implementing
this hydraulic machine in other domains, e.g. along
water distribution pipelines for using the hydraulic
potential available between upper and lower
reservoirs that are part of the water supply network of
a city (Sinagra et al., 2014) or at the end of a sewer
pipe delivering the treated waste water to its final
receiving water body (Khan and Badshah, 2014).
The usual working domain of cross-flow
turbines covers a range that goes from 2 m to 200 m.
However, the research presented in Annex 2, Subtask
B2 “Innovative Technologies for Small-Scale
Hydro”, elaborated by the International Energy
Agency, proposes a new type of cross-flow turbine
dedicated to very low heads: from 0.6 m to 10 m
(IEA Annex-2 STB2, 2016).
The interest for harvesting hydro power at
very low heads is not uncommon nowadays (Elbatran
et al., 2015). The situation can be met at household
users interested in green energy, who want to
implement combined solar-wind-hydro renewable
systems or just hydro units. Most stand-alone energy
producers either are discouraged by complicated
980
administrative problems generated by the usage of
water from registered rivers or want to avoid legal
issues generated by ownership outside their property.
In such cases, a low head pico-turbine that can be fed
with water from a runlet could be the best solution.
The challenge in the present paper is to find
out whether an eco-friendly cross-flow pico-turbine
can be used efficiently, when only a very low head is
available. Experimental results obtained for a Banki
turbine fed from an open channel at very low heads
are presented and analyzed with an emphasis on the
maximum efficiencies that can be attained. The paper
also presents results of a Computational Fluid
Dynamics (CFD) analysis that complements the
experiments and offers a clearer picture of the air-
water two-phase flow inside the turbine.
2. Theoretical considerations and turbine design
Different methods for designing cross-flow
turbines are presented in literature. They are usually
based on the considerations made by Mockmore and
Merryfield (1949). Recently, Sammartano et al.
(2013) proposed a two-step design procedure for
cross-flow turbines in the Banki–Mitchell variant. In
the first step, some critical design parameters were
theoretically estimated, while in the second step the
influence of the remaining design parameters on
efficiency was analyzed by means of 2D numerical
simulations. It was found that the number of blades
and the diameter ratio have only little influence on
the peak efficiency and that the presence of the shaft
does not affect the characteristic curve, although it
considerably diminishes the efficiency. Based on this
method, Sinagra et al. (2014) presented a simple yet
rigorous design procedure for a cross-flow turbine
with discharge regulator, which allows to obtain a
constant hydraulic head as long as the head losses
between water source and turbine are negligible.
The cross-flow turbine investigated in this
study was designed for a rated power, P, of 1.5 kW
and a rated head, H, of 1 m. From the point of view
of the head, the turbine can be considered as a very
low head one, since cross-flow turbines are usually
designed for heads larger than 2m (Cink Hydro-
Energy, 2016; Ossberger GmbH+Co, 2016). The
design was carried out according to simple yet
efficient recommendations formulated by Bărglăzan
(2001) and based on the classical theory of
turbomachinery. The design procedure is presented
below.
Cross-flow turbines usually have specific
speeds in the range 50–150. The specific speed is an
essential similarity parameter, given by Eq. (1):
n PHP0.5
ns = (1)
H 1.25
where n is the rotational speed expressed in rpm and
PHP is the power expressed in horsepower. Since ns is
in inverse ratio to a power of H, it increases when the
head decreases. Hence, for the turbine designed for a
 Experimental investigation and CFD analysis of an eco-friendly turbine
very low head, the value ns=150 was assumed. The
rotational speed that results from Eq. (1) is
n=105.04rpm. The value n=105 rpm was adopted.
Another important similarity parameter of
hydraulic turbines is the unit speed:
n D1
n11 =
H 0.5
where D1 is the outer diameter of the turbine runner.
The optimal unit speed is of about 40rpm. With this,
the value D1=0.3809m results from Eq. (2). To ease
the manufacturing, a slightly higher diameter of
D1=0.382m was adopted. For this value, the unit
speed becomes n11=40.11rpm and does not deviate
significantly from the optimal value.
The inner diameter of the cross-flow runner,
D2, must be calculated as well. In many studies, the
ratio D2/D1 is chosen equal to 0.66 (Chattha et al.,
2014; Mockmore and Merryfield, 1949). For this
ratio, the diameter D2=0,252m was obtained.
The efficiency was estimated with the
equation proposed by Banki (1922) (Eq. 3):
D1
η = 0.771 − 0.384
H was
The value η≈0,625 was obtained. It was then
straightforward to calculate the discharge (Eq. 4):
PkW
Q=
9.81 H η
where PkW is the power expressed in kW. The value
Q≈0.245m3/s was obtained.
Khosrowpanah et al. (1988) studied the
optimum number of blades and proposed an equation
to calculate it, based on the optimum solidity ratio.
This parameter was not yet established by further
work. According to the study “Parametric evaluation
of cross-flow turbine performance”, conducted later by
Desay and Aziz (1994) to quantify the effect of key
parameters influencing the maximum efficiency of
cross flow turbines, “there is no concurrence of
opinion regarding the effect of the number of blades
on cross flow turbine performance”. Most researchers
prefer to choose the number of blades in the range 15–
30, according to recommendations found in the
scientific literature. Few blades can cause pulsations
and more blades can increase friction losses. An
experimental study conducted on 27 runners, also by
Desai and Aziz (1994a), recommends 25 blades for a
good efficiency. For this design, the number of blades
was chosen equal to Z=24blades, so that the angle
between two successive blades is15, a rounded value
easy to be measured during manufacturing.
The absolute velocity angle at runner entry,
α1, was chosen equal to 16°, similar to the choice
made by Mockmore and Merryfield (1949) and in
most published papers. The absolute velocity at
runner entry is given by Eq. (5):
v1 = k v1 2 g H (5)
where the coefficient of the absolute velocity at
runner entry, kv1, should be chosen in the range 0.96–
0.98. The value kv1=0.98 was adopted. It resulted the
velocity v1=4.341 m/s. From Eq. (6) it results the
(2) meridional velocity at runner entry vm1 = 1.197 m/s.
The blade angle at runner entry is given by Eq. (7):
v m1 = v1 sin( α 1 ) (6)
 v m1 
β1 = arctan  (7)
 1
v cos(α 1 ) − u1 
where u1 is the peripheral velocity and is given by
Eq. (8):
π D1 n
u1 = (8)
60
The peripheral velocity corresponding to the
chosen rotational speed and diameter has the value
u1=2.101 m/s. It resulted the angle β1 ≈ 30°.
At the interior of the runner, the blade angle
(3)
chosen β2=90°, according to the
recommendations made by Mockmore and
Merryfield (1949).
3. Experimental study
(4)
The turbine designed previously was
manufactured and tested at the Fluid Mechanics and
Fluid Machinery Laboratory from “Gheorghe
Asachi” Technical University of Iaşi, Romania. Fig.
1a presents the schematic view of the experimental
set-up that consists of an open channel and the main
parts of the turbine – the nozzle and the runner –,
while Fig. 1b shows a photo of the runner.
Two centrifugal pumps located in the
basement of the laboratory deliver the water through
a pipe to the open channel 1, that has a width of
0.76 m. At the end of the open channel, the water
falls freely into the vertical nozzle 2, as it is
presented in Fig. 1a. From the nozzle, the water is fed
to the turbine runner 3. The nozzle is made from steel
sheet and the blades of the runner are made of plastic.
It should be mentioned that the runner is divided into
three parts of equal width, so that tests at one third
and two thirds of the discharge are also possible.
The turbine shaft is connected to a mechanical
brake via a chain drive. The brake allows to control
the turbine load and to change the speed from zero to
the maximum possible value.
The discharge was measured with a Siemens
SiTrans FUE1010 portable non-intrusive clamp-on
ultrasonic flowmeter, accuracy ±0.5 to 1%, placed on
the pipe that feeds the channel with water. The
turbine head was measured inside the nozzle with
NIVOCAP CBR-215-4 level transducers, with the
measurement range 0.2–3 m and the linearity error
981
 Popescu et al./Environmental Engineering and Management Journal 16 (2017), 4, 979-988

±0.3% FS. The torque was measured with an AEP 2π n [rad/s]

ω= (10)
Type RT2 torque transducer with linearity and 60
hysteresis accuracy of ±0.2%. The rotational speed
was measured with a Compact A2102 optical where n is the rotational speed.
tachometer with the accuracy ±0.05%. The turbine performance was studied in terms
of unit quantities corresponding to the specific
turbine (a turbine that has D1=1m and operates at
H=1m). The unit speed, denoted n11, was calculated
with Eq. (11):

nD1 [rpm]
n11 = (11)
H 0.5

where the rotational speed, n, is expressed in rpm, the

runner diameter, D1, is expressed in m, and the head,
H, is expressed in m. It is customary to express n11 in
rpm, despite the fact that this is not the real
measurement unit.
Results of the study on efficiency versus unit
(a)
speed are presented in Figs. 2-4 for different turbine
heads that correspond to the three distances from the
bottom of the channel to the axis of the turbine.

(b)

Fig. 1. Experimental set-up: a) sketch, b) detail photo of

the turbine nozzle and runner. The main components are:
1 – open channel, 2 – turbine nozzle, 3 – turbine runner Fig. 2. Experimental results measured at Y1=0.6 m

The tests were performed for turbine heads in

the range H=0.6–1.6 m. To change the head, the
distance from the bottom of the channel to the axis of
the turbine, denoted Y (Fig. 1a), was changed by
means of a lifting/descending system in three
positions: Y1=0.6 m, Y2=0.9 m, and Y3=1.2 m. The
measured discharge varied from 0.1 m3/s to 0.22m3/s.
For the heads and discharges investigated, the
mechanical power obtained was in the range 130–
1300 W, while the rotational speed varied
from 40 rpm to 220 rpm. The maximum efficiency
rate was in the range 51–66%, in accordance to the
best efficiencies calculated with classic theoretical
methods, valid for turbines that work at higher heads. Fig. 3. Experimental results measured at Y2=0.9 m
The overall turbine system efficiency η, was
calculated with Eq. (9): The data presented in Figs. 2-4 was fitted with
third order polynomials which offered very good
Mω [-]
η= (9) aproximations. The obtained R squared values can be
ρ gQH read in Table 1. As Fig. 2 shows, during experiments
at Y1 = 0.6 m, efficiencies above 40% were obtained
where M is the torque, ω is the angular speed, Q is irrespective of the discharge, for unit speeds n11 =
the measured discharge, H is the turbine head, ρ is 30…65 rpm, calculated at rotational speeds in the
the density, and g is the gravitational constant. The range n=80...125 rpm.
angular speed ω is given by Eq. (10):

982
 Experimental investigation and CFD analysis of an eco-friendly turbine
Table 1. R squared values of the polynomials that fit the experimental data
Y1 = 0.6 m
Q (l/s) 97 120 150 170 197
R2 0.998 0.999 0.996 0.996 0.976
Fig. 4. Experimental results measured at Y3 = 1.2 m
Fig. 2 suggests also that, at very low heads,
the optimal unit speed, for which the maximum
efficiency is obtained, shifts towards 43rpm. For
Y2=0.9 m (Fig. 3), efficiencies above 40% were
obtained for unit speeds n11=25…65 rpm, calculated
at rotational speeds n=100...150 rpm. At Y3=1.2m
(Fig. 4), the highest efficiencies were obtained for
unit speeds n11=25…55rpm, that correspond to
rotational speeds n=100...150rpm (Popescu et al.,
2016).
The most important diagram that summarizes
results of all cases is the turbine hill chart that plots
curves of constant efficiency in coordinates Q11–n11.
Fig. 5 presents the efficiency hill chart of the studied
turbine, based on a fit of the experimental data with a
two-dimensional third order polynomial.
Fig. 5. Turbine hill chart
Fig. 5 clearly shows that efficiencies of at
least 55% or higher were obtained over the entire
range of unit discharges. However, the range of unit
speeds for which η ≥ 55% is narrow at low values of
Q11 and broadens as the unit discharge increases.
Y2 = 0.9 m Y3 = 1.2 m
135 162 180 203 142 160 191 232
0.988 0.989 0.960 0.958 0.991 0.989 0.968 0.994
4. CFD analysis
CFD has become a standard tool for analyzing
the flow through turbomachinery (Ansys, 2011). Many
CFD-based studies were carried out for different
designs of cross-flow turbines aiming at getting a
better insight into the flow through the nozzle and
runner and at optimizing the main geometrical
parameters. Pereira and Borges (1996) performed
numerical simulations only for the flow inside the
turbine nozzle. Kobuku et al. (2011) investigated a
cross-flow turbine fully filled with water, although
during normal operation an important part of the
turbine contains air as well. There are important
differences between only-water and water-air results
that were emphasized by Arzola et al. (2008) after a
3D numerical study on the flow inside a nozzle. Choi
et al. (2008) investigated the effect of the structural
configuration on the performance and internal flow
characteristics of a cross-flow turbine. De Andrade et
al. (2011) carried out 3D simulations for the steady-
state flow through a cross-flow turbine at different
rotational speeds. They used a homogenous free
surface model to track the water-air interface and the
k-ε turbulence model. The results showed that the
absolute velocity angle at runner entry (entry of the
first stage), α1, deviates significantly from the design
value of 16°.
But the worst deviation from the design value
(of 90°) was found for the relative velocity angle at the
entry of the second stage, β3, most probably due to the
interference of the streamlines that leave the first stage.
Additionally, other fluid dynamic phenomena, such as
shocks due to the interaction of the water with the
turbine shaft, recirculation zones in the blade passages,
and non-cross-flow were found. Choi et al. (2010)
investigated both experimentally and numerically the
internal flow and the performance of a cross-flow
turbine intended to be used as a wave energy
converter. They used the SST turbulence model that
offers the advantage of resolving to a good extent both
high and low energy areas on a complicated grid with
relatively low grid dependency.
In this study, due to the simplicity of the
turbine geometry, that remains identical on any plane
perpendicular to the turbine axis between the two end
discs of the runner, the numerical analysis was
performed for a two-dimensional geometry that
represents a cross-section of the turbine. The
computational domain is presented in Fig. 6a. It
consists of three sub-domains: the end part of the
nozzle, the runner, and the draft tube. The domain is
bounded by the inlet, through which the water enters
the nozzle, the inner wall of the nozzle, the surfaces
983
 Popescu et al./Environmental Engineering and Management Journal 16 (2017), 4, 979-988
of the runner blades, the outer surface of the turbine
shaft and the outlet, through which the water-air
mixture leaves the computational domain. The three
sub-domains are separated by a circular interface
having a radius of 192.2 mm (with 1.2 mm larger
than the outer radius of the runner). The
computational domain was meshed with quadrilateral
cells forming both structured and unstructured grids.
Fig. 6b presents two mesh details.
(a) (b)
Fig. 6. a) Sketch of the computational domain, b) mesh
details. All dimensions are in mm
Since a cross-flow turbine has partial
admission and processes only the kinetic energy of
the water jet formed at nozzle exit, a water-air two-
phase mixture flows through the runner. In the
numerical simulations, this flow was considered
unsteady (due to the runner movement) and
turbulent. The flow is described by the continuity and
momentum equations, the equation that tracks the
water-air interface, and the closure equations for the
turbulent quantities. The closure was provided by the
k-ω SST turbulence model.
As boundary condition at inlet, a constant water
mass flow rate of 502.2 kg/(s m) was imposed. The
value corresponds to a volumetric flow rate of 232 l/s
distributed over the entire width of the runner. A total
pressure of 0 Pa was imposed at the outlet, since here
Table 2. Maximum efficiency obtained during experimental tests
Y [m] 0.6
H [m] 0.62 to 1.07 m
L [%] 33% 66% 100%
ηmax, [%] 65.5 65.5 65.9
nopt [rpm] 90.4 95.9 115.0
984
the pressure equals the atmospheric pressure. At all
solid walls, the usual no-slip condition was used.
As reference pressure and reference
temperature the values p0=105Pa and T0=293.15K,
respectively, were considered. The corresponding
values of the densities are ρw=1000kg/m3 for water and
ρa=1.188kg/m3 for air. The dynamic viscosities are
μw=1.002x10-3Pa s for water and μa=1.824x10-5Pa s for
air.
The flow equations together with the
boundary conditions were integrated in time and
space with the Finite Volume Method and the water-
air interface was tracked with the Volume of Fluid
Method. The runner movement was simulated with
the Sliding Mesh Technique, allowing the mesh
nodes of the runner sub-domain to rotate with the
turbine speed. Simulations were carried out for the
following speeds: 50, 75, 100, 125, 150, 175, and
225 rpm. The time steps were decreased from 0.5 ms
to 0.1 ms as the speed increased. At each time step,
the convergence criterion was the drop in scaled
residuals below 10-3.
5. Analysis of results
Tests were performed for different sections
of the rotor, to cover a large field of input
parameters. For each distance Y, measured from the
bottom of the channel to the axis of the turbine,
tests were run firstly at the lowest discharge rates,
when water passes through one section of the rotor,
the next tests were run when water passes through
two sections and the final tests when the full width
of the rotor is used.
The best efficiency point ηmax and the
optimum rotational speed nopt to reach it are
presented in Table 2, for each series of
investigation.
The experimental tests lead to best efficiency
points obtained in the range ηmax=51.5 to ηmax=65.5%
when the water passes through the first section
(L=33%), ηmax=58.5 to ηmax=65.3% when the water
crosses two sections (L=66%) and ηmax=58.7 to
ηmax=65.9%, when the water passes through the full
width of the rotor (L=100%).
In order to understand if the studied turbine
has good performance, the experimental results
were compared to the theoretical expected
maximum efficiency values, calculated by
equations, in Table 3.
0.9 1.2
0.85 to 1.38 m 1.16 to 1.6 m
33% 66% 100% 33% 66% 100%
53.3 62.8 64.5 51.5 58.8 58.7
95.0 108.0 135.3 109.7 118.0 136.3
 Experimental investigation and CFD analysis of an eco-friendly turbine

One of the most popular efficiency scale
formulae (Eq. 12), recommended by IEC to be used
in order to find the best expected efficiency is that of
Hutton (Turton, 1984):
1 was obtained at the lowest head H=0.6 m, not at the
1 − ηr  Re  5
= 0.3 + 0.7 ⋅  m 
1 − ηm  Re r 
where the subscript m refers to the model and the
subscript r refers to the prototype or to another
model. The Reynolds criteria for turbomachines can
be calculated with Eq. (13):
D1 2 gH
Re =
ν
The diameter of the tested cross flow turbine
is D=0.382 m, which means (Eq. 14):
Re = 1692048 ⋅ H
In the case Hm = 1.2 m and ηm = 58.7 %, the
Reynolds criteria for the model is Rem = 1853546.
By applying Hutton’s theory, for turbine head
Hr = 0.9 m, the expected maximum efficiency of the
turbine should be ηr = 57.85 %. At turbine head Hr =
0.6 m the maximum efficiency of the turbine should
be ηr = 56.62 %.
According to the equation proposed by Banki,
the maximum hydraulic efficiency of the turbine can
be calculated with Eq. (15):
D1
η h' max = 0.771 − 0.384
H 18 blades, at a head of 0.91 m to 1.07 m, similar to
and, according to Sonnek, with Eq. (16):
D1
η h'' max = 0.863 − 0.264
H experimental measurements. Fig. 7 presents a
The efficiency of the turbine is given by
Eq. (17):
η turb = ηh·ηv·ηm
The maximum expected efficiency of the
turbine working at full length (L=100%) is presented
in Table 2. According to the theory of Hutton and to
formulas proposed by Banki and Sonnek, the
maximum efficiency should decrease when the
prototype is tested at lower heads. The experimental
results obtained in this work are in the expected
range of efficiency calculated by the equations
presented above, but the best energy efficient point
(12) highest. This is the main finding of this paper: a well
designed low head turbine can have good efficiency,
when it works at heads up to 2 m.
Others studies also mentioned that for certain
types of cross flow turbines, the efficiency might not
decrease when the turbine works at very low heads.
For instance, the work of Kaunda et al. (2014) on
experimental tests on a cross flow turbine working at
H=3m to H=10 m, identifies the best efficiency point
(13) ηmax=79 % at the head H=5 m and a valve opening of
80%. In the following, the results presented in this
paper were compared to previous research that had
major contribution to the domain. The most cited
research work on Banki turbines was done by
Mockmore and Merrryfield (1949) who tested a
(14)
runner, external diameter D1=332.7 mm, width
B=305 mm, at turbine head values in the range
H=2.74 m to H=5.49 m. The best efficiency,
ηmax=68% was obtained at H=4.88 m.
Tests performed by Desai and Aziz (1994b)
showed that the optimal angle of attack of the flow to
the runner is in the range 220-240, and a decrease in
diameter ratio D2/D1 from 0.75 to 0.68 leads to an
increase of the maximum efficiency, up to 76.5%.
Johnson et al. (1982) designed and tested a
cross flow turbine with guide vanes. The tests
indicated that the maximum efficiency was in the
range ηmax=60-80%. The interesting result is that the
best efficiency point was reached for a runner having
(15)
the case study presented in the present paper.
The behavior of the turbine can be explained
based on results of the numerical simulations. These
results were validated by comparing them with the
(16)
comparison between measured and simulated
torques, obtained at full runner width. The values of
the measured torques are those obtained in the set of
measurements for which the discharge was kept
(17) constant at 232 l/s. The results of the simulations are
slightly higher than the measured ones. However, it
should be considered that the simulations account
neither for the mechanical losses inside the two shaft
bearings nor for the volumetric losses due to water
that bypasses the runner at its end discs.

Table 3. Maximum expected efficiency calculated according to classical methods

Method Hutton Banki Sonnek

H [m] 0.6 0.9 1.2 0.6 0.9 1.2 0.6 0.9 1.2
ηmax [%] 56.62 57.85 58.7 43.69 50.46 53.83 57.67 62.32 64.64

985
 Popescu et al./Environmental Engineering and Management Journal 16 (2017), 4, 979-988
When the mechanical and volumetric
efficiency are estimated at 98% and 95%,
respectively, the simulated torques corrected to
account for the combined mechanical and volumetric
losses (the dashed curve in Fig. 7) agree very well
with the measurements.
Fig. 7. Comparison between measured torques and torques
obtained by numerical simulations for a discharge of
232 l/s at full runner width
To get an insight into the flow through the
turbine, Fig. 8 shows volume fractions of water (in
black) and air (in white) inside the turbine at four
rotational speeds – 50, 75, 100, and 125 rpm – that
are close to the rated one of 105 rpm. For n = 50 rpm,
less than half the rated speed, two negative effects
can be observed. One of them is the already known
water-shaft interaction (i.e. the interaction of the
water jet, that forms inside the runner, with the
turbine shaft). This interaction is expected to cause
important hydraulic losses inside the runner. The
other negative effect is the entrainment of air inside
the blade passages in form of air bubbles that remain
close to the suction side of the blades. These bubbles
are visible only during the first pass (in the first
stage). The appearance of these bubbles is likely to
be caused by the inappropriate kinematic conditions
at runner entry. Since the rotational speed is lower
Fig. 8. Water-air volume fractions inside the turbine (water in black, air in white) obtained
for the discharge Q = 232 l/s at four rotational speeds. The runner rotates counter-clockwise
986
than the rated one, the relative velocity angle (the
angle between the relative velocity and the peripheral
one) is lower than the blade angle. Consequently, the
water flow does not attach properly on the suction
side of the blades and cannot evacuate rapidly
enough the air pushed by the blades from the region
where they are idle (i.e. the upper region of the
runner, filled with air, where there is no interaction
between blades and water). It could be assumed that
the appearance of the air bubbles diminishes the
volumetric efficiency of the turbine, since the blades
do not fully come into contact with water. Both
negative effects discussed above are expected to
significantly reduce the turbine efficiency.
The picture of the flow inside the runner
changes visibly at n = 75 rpm. There is still a water-
shaft interaction, although this interaction is reduced
when compared to the case n = 50 rpm. However, the
blade passages fill completely with water during the
first pass and no air bubbles form. The flow inside
the blade passages seems to be optimal, but there is
still the water-shaft interaction that is expected to
diminish the turbine efficiency.
The case n = 100 rpm, which is very close to
the rated speed of 105 rpm, is likely to be the closest
to an optimal operation. There is only a very slight
interaction water-shaft and some small air bubbles
tend to form, but they are relatively rapidly washed
away.
Finally, the case n = 125 rpm is free of any
water-shaft interaction, but large air bubbles reappear
inside the blade channels. These bubbles are not
evacuated during the first pass. However, it is
interesting to note, that, unlike the case n = 50 rpm,
the bubbles form now closes to the pressure side of
the blades. The rotational speed higher than the rated
one leads to a relative velocity angle that is larger
than the blade angle. Consequently, the water flow
cannot attach properly on the pressure side of the
blades and air remains entrapped there.
 Experimental investigation and CFD analysis of an eco-friendly turbine

6. Conclusions Systems”, type PN-II-PT-PCCA-2013-4, contract

45/1.07.2014, financed by the Romanian Executive Agency
A pico-unit for clean energy production, more for Higher Education, Research, Development and
specifically a cross-flow turbine designed for very Innovation Funding.
low heads, was successfully tested experimentally Nomenclature
and simulated numerically. The experiments showed
that the proposed design can work with reasonable D [m] –diameter
efficiencies, at measured heads of up to 1.6 m and g [m/s2] –acceleration due to gravity, g = 9,81 m/s2
measured discharges of up to 0.25 m3/s, limits below H [m] –turbine head
which any other types of turbines are difficult to be kv [—] –coefficient of the absolute velocity
used. It is remarkable that at the lowest head, H = M [Nm] –torque (shaft moment)
0.6 m, the obtained efficiency is higher than the n [rpm] –rotational speed
n11 [rpm] –unit speed
expected values calculated with the classical Hutton, ns [—] –specific speed
Banki and Sutton methods. The tested turbine, R2 [—] –Pearson product moment correlation
designed to work at very low heads, reaches the best coefficient
efficiency point at this very low head, independent of Re [—] –Reynolds number
whether the water passes only through the first P [W] – power
section (33% of the runner width), or crosses two PHP [hp] –power expressed in metric horsepower
sections (66% of width), or passes through the full (hp)
runner (100% of width). PkW [kW] –power expressed in kW
The numerical simulations were validated p [Pa] –pressure
Q [m3/s] –discharge (volume flow rate)
based on the experimental results. A good T [K] –absolute temperature
agreement between measured torques and torques u [m/s] –peripheral velocity
obtained from the numerical simulations was found. v [m/s] –absolute velocity
The simulations revealed an interesting picture of Y [m] –distance from the bottom of the channel to
the flow through the very low head cross-flow the axis of the turbine
turbine. Their results suggest that, apart from the Z [—] –number of runner blades
water-shaft interaction (the interaction of the water α [°] –absolute velocity angle
jet that forms inside the runner with the turbine β [°] –relative velocity angle, blade angle
shaft), another negative phenomenon appears: air η [—] –efficiency
μ [Pa s] –dynamic viscosity
bubbles get entrapped into the blade channels ω [rad/s] –angular speed
during the first pass (or in the first stage) at ρ [kg/m3] –density
rotational speeds that are far from the rated one. At
speeds lower than the rated one, the air bubbles tend Indexes
to form close to the suction side of the blades, while 1 –at the exterior of the runner
at speeds higher than the rated one, the bubbles 2 –at the interior of the runner
move towards the pressure side of the blades. The a –air
appearance of the air bubbles is likely to be a result m –meridional component
w –water
of the operation at a very low head (i.e. at a low
max –maximum
pressure inside the turbine nozzle) combined with opt –optimum
the kinematic conditions at runner entry, that
become inappropriate at speeds that differ References
significantly from the rated one. Both the water-
shaft interaction and the entrapment of air bubbles Andrade J.D., Curiel C., Kenyery F., Aguillón O., Vásquez
are expected to diminish the turbine efficiency. A., Asuaje M., (2011), Numerical investigation of the
However, for an operation close to the rated speed, internal flow in a Banki turbine, International Journal
of Rotating Machinery, 2011, 1-12.
the two negative phenomena are only barely
Ansys, (2011), Ansys Fluent Theory Guide, Release 14.0,
present. Ansys Inc., Canonsburg.
The results presented show that a very low- Arzola F., Rodriguez C., Martin J., De Andrade J., Vásquez
head cross-flow turbine could work with good A., Asuaje M., (2008), Technical Assessment for
efficiencies when properly designed and operated. Overhaul Project in Small Hydro Power Plant, Proc.
This recommends the type of turbine investigated in of the 24th IAHR Symposium on Hydraulic
this paper for harnessing very low hydraulic potential Machinery and Systems, Foz do Iguassu, Brazil.
in households that want to benefit from the Banki D., (1922), Water turbine. U.S. Patent, No. 1436933.
advantages offered by renewable resources either in Bărglăzan M., (2001), Hydraulic Turbines and
Hydrodynamic Drives (in Romanian), Politehnica
form of combined solar-wind-hydro systems or as
Press, Timișoara, Romania.
standalone electricity production units. Bostan I., Dulgheru V., Bostan V., Sochireanu A., Trifan
N., Dicusară I., Ciobanu O., Ciobanu R., (2011), A
Acknowledgements micro hydro-power station for the conversion of
The paper represents a result of the research project “Cross flowing water kinetic energy, Environmental
Flow Hydro Turbines for Pico Renewable Energy Engineering and Management Journal, 10, 1033-

987
 Popescu et al./Environmental Engineering and Management Journal 16 (2017), 4, 979-988
1039.
Chattha J.A., Khan M.S., Iftekhar H., Shahid S., (2014),
Standardization of Cross Flow Turbine Design for
Typical Micro-Hydro Site Conditions in Pakistan,
ASME 2014 Power Conference, paper V002T09A006.
Choi Y.D., Lim J.I., Kim Y.T., Lee Y.H., (2008),
Performance and internal flow characteristics of a
cross-flow hydro turbine by the shapes of nozzle and
runner blade, Journal of Fluid Science and
Technology, 3, 398-409.
Choi Y.D., Kim C.G., Kim Y.T., Song J.I., Lee Y.H.,
(2010), A performance study on a direct drive hydro
turbine for wave energy converter, Journal of
Mechanical Science and Technology, 24, 2197-2206.
Cink Hydro-Energy, (2016), On line at: www.cink-hydro-
energy.com/en.
Codruta J., Diaconescu D., Neagoe M., Saulescu R.,
(2009), The eco impact of small hydropower
implementation, Environmental Engineering and
Management Journal, 8, 837-841.
Desai V.R., Aziz N.M., (1994a), An experimental
investigation of cross-flow turbine efficiency, Journal
of Fluids Engineering, 116, 545-550.
Desai V.R., Aziz N.M., (1994b), Parametric evaluation of
cross-flow turbine performance, Journal of Energy
Engineering, 120, 17-34.
Directive 28, (2009), Directive 2009/28/EC of the
European Parliament and of the Council of 23 April
2009 on the promotion of the use of energy from
renewable sources and amending and subsequently
repealing Directives 2001/77/EC and 2003/30/EC,
Official Journal of the European Union 2009, L 140,
16-62.
Elbatran A.H., Yaakob O.B., Ahmed Y.M., Shabara H.M.,
(2015), Operation, performance and economic analysis
of low head micro-hydropower turbines for rural and
remote areas: a review, Renewable and Sustainable
Energy Reviews, 43, 40-50.
IEA Annex-2 STB2, (2016), Data Sheet about Innovative
Technology, On line at: http://www.small-
hydro.com/Programs/innovative-technologies.aspx.
IRENA, (2014), REmap 2030: A Renewable Energy
Roadmap, June 2014. IRENA, Abu Dhabi. On line at:
www.irena.org/remap.
IEA-ETSAP, IRENA, (2015), Hydropower-Technology
brief, On line at:
http://www.irena.org/DocumentDownloads/Publicatio
ns/IRENA-ETSAP_Tech_Brief_E06_
988
Hydropower.pdf.
Johnson W., Ely R., White F., (1982), Design and Testing
of an Inexpensive Cross-Flow Turbine, Proc. of
ASME Winter Annual Symposium “Small Hydro-
power Fluid Machinery”, Phoenix, AZ, 129-133.
Kaunda C., Kimambo C., Nielsen T., (2014), Experimental
study on a simplified crossflow turbine, International
Journal of Energy and Environment, 5, 155-182.
Khan M., Badshah S., (2014), Design and analysis of cross
flow turbine for micro hydro power application using
sewerage water, Research Journal of Applied
Sciences, Engineering and Technology, 8, 821-828.
Khosrowpanah S., Fiuzat A.A., Albertson M.L., (1988),
Experimental study of cross-flow turbine, Journal of
Hydraulic Engineering, 114, 299-314
Kobuku K., Son S.W., Kanemoto T., Choi Y.D., (2011),
Internal flow analysis on a micro cross-flow type
hydro turbine at very low specific speed range, Proc.
of the 11th Asian International Conference on Fluid
Machinery, IIT Madras, Chennai, India.
Mockmore C.A., Merryfield F., (1949), The Banki water
turbine, Engineering Experiment Station, 25, 5-30.
Ossberger GmbH+Co, (2016), On line at:
www.ossberger.de/cms.
Pereira N.H.C., Borges H.E., (1996), Study of the nozzle
flow in a cross-flow turbine, International Journal of
Mechanical Sciences, 38, 283-302.
Popescu D., Popescu C., Dragomirescu A., Drosescu R.,
Ciobabu B., Schiaua M., Panaitescu V., Nastase V. E.,
Panaite E.C., Alexandrescu A., (2016), Report 3/2016,
project PICOTURB, On line at:
http://www.picoturb.tuiasi.ro/Raport_2016.pdf.
Sinagra M., Sammartano V., Aricò C., Collura A.,
Tucciarelli T., (2014), Cross-flow turbine design for
variable operating conditions, Procedia Engineering,
70, 1539 – 1548.
Turton R.K., (1984), Principles of Turbomachinery,
Springer, Netherlands.
Udroiu R., Nedelcu A., Deaky B., (2011), Rapid
manufacturing by polyjet technology of customized
turbines for renewable energy generation,
Environmental Engineering and Management Journal,
10, 1387-1394.
VLH Turbine, (2016), MJ2 Technologies, France,
www.vlh-turbine.com/FR/html/Pdt_VLH.htm.
WEC, (2016), Hydropower 2016, On line at:
https://www.worldenergy.org/wpcontent/uploads/2017
/03/WEResources_Hydropower_2016.pdf.