Journal of Applied Research and Technology

www.jart.icat.unam.mx

Journal of Applied Research and Technology 21 (2023) 181-195

Original

Performance simulation of water turbines by using 6-DoF UDF

and sliding mesh methods

J. Betancoura L. Velásqueza A. Rubio-Clementea,b E. Chicaa*

a
Grupo de Investigación Energía Alternativa (GEA),
Facultad de Ingeniería, Universidad de Antioquia, Colombia
b
Escuela Ambiental. Facultad de Ingeniería,
Universidad de Antioquia, Colombia.

Received 04 02 2021; accepted 08 02 2022

Available 04 30 2023

Abstract: In this work, the differences between the six degrees of freedom (6-DoF) method and the
sliding mesh approach in computational fluid dynamics simulations of an Archimedes screw turbine
for hydrokinetic applications were studied using ANSYs Fluent software. The numerical curves
obtained by representing the power coefficient (𝐶𝑝) values versus the tip speed ratio (𝑇𝑆𝑅 or 𝜆) were
compared based on experimental data. Both methods allowed to simulate the rotor rotation; although
the 6-DoF method was more accurate in the prediction of the turbine performance, higher
computational resources were consumed in comparison with the sliding mesh method, which allows
to obtain a computational solution in affordable times.

Keywords: Archimedes screw turbine, hydrokinetic turbine, 6-DoF method,

sliding mesh method

∗
Corresponding author.
E-mail address: [email protected] (E. Chica).
Peer Review under the responsibility of Universidad Nacional Autónoma de México.

http://dx.doi.org/10.1016/j.jart.2017.02.005
1665-6423/© 2017 Universidad Nacional Autónoma de México, Instituto de Ciencias Aplicadas y Tecnología.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 J. Betancour et al. / Journal of Applied Research and Technology 181-195

1. Introduction approach and the sliding mesh model (SMM). Among the multiple
moving reference frame approaches, the multiple reference
Computational fluid dynamics (CFD) simulation is a cost- frame model (MRF) and the mixing plane model (MPM) can be
effective alternative used to validate the performance of named (ANSYS Inc., 2018a). The referred approaches are steady-
hydrokinetic turbines instead of conducting expensive state approximations for treating the conditions at the interfaces.
experimental research (Tiwari et al., 2020). CFD studies, both The MRF, which is widely known as the Frozen Rotor approach, is
two- and three-dimensional (3D) configurations, allow to generally used when the flow at the interface is nearly uniform. In
simulate the flow around the turbine, calculate the blade turn, the MPM allows to simulate the flow through domains
forces, determine the velocity and the pressure distributions, containing one or more cells in relative motion. On the other
predict the performance of the turbine of interest and study hand, when the unsteady interaction between the dynamic and
physical phenomena, including the flow separation and the the stationary subdomains is important, the utilization of the
geometry optimization of the turbine main components (Tian sliding mesh approach is plausible to capture the transient
et al., 2018; Yavuz et al., 2015; Zhao et al., 2020; Zitti et al., 2020). behavior of the flow (ANSYS Inc., 2018b) In the multiple moving
In order to numerically design a hydrokinetic turbine, the reference frames, the SMM is an accurate method for simulating
blade element momentum (BEM) theory or other approach is flows; however, this method is more computationally demanding
first employed to determine the hydrodynamic performance compared to other multiple moving reference frame approaches.
and the force generated on the turbine blades (Abutunis et al., In turn, when the mesh motion is a function of the solution or the
2019). Secondly, CAD model or computational domain of the mesh is deformed, the dynamic mesh method must be used
turbine must be constructed. The third stage consists of along with the degrees of freedom (DoF) solver (ANSYS Inc.,
dividing the domain into a finite number of cells by using a 2018b). It is highlighted that the dynamic mesh solution is
structured or an unstructured mesh. Afterwards, the Navier intended to be more robust and accurate; nevertheless, it is also
Stokes equations coupled with a turbulence model are used more time-consuming than other methods (ANSYS Inc. 2018b;
to build a mathematical model describing the changes on Prakoso et al., 2019; Qu et al., 2015;).
those physical properties for the fluid flow and the model In the literature, several studies reporting the use of the
geometry. Finally, several boundary conditions (e.g., velocity rotation modeling approach have been reported. As a matter
of the flow at the inlet, pressure, temperature and the density, of fact, Yavuz and coworkers utilized the MRF approach to
among others), and a rotor rotation modeling approach are define the blade rotation in order to study the performance of
defined by the researcher in order to solve the problem airfoil or hydrofoil arrangements for wind or hydrokinetic
throughout computer-based simulations. To verify the turbines, respectively (Yavuz et al., 2015). Similarly, Zitti et al.
numerical simulations, mesh and time step convergence employed the referred approach to calculate the efficiency of
studies should be also performed. When a hydrokinetic Archimedes turbines (AST) (Zitti et al., 2020).
turbine is numerically simulated, the flow domain consists of In contrast to MRF, the utilization of SMM has been
a dynamic and a stationary subdomain. Typically, the dynamic reported by Tian et al. (2018) and Wang et al. (2019) to simulate
subdomain involves moving parts like the blades, rotor or the rotor rotation of a horizontal-axis hydrokinetic turbine.
moving surfaces. During the numerical simulation, each The simulations performed by Tian et al. were focused on the
subdomain is independently discretized and then a study of the wake structure and the flow around the blades
boundary/interface is established in the surfaces that are in (Tian et al., 2018). Concerning the numerical simulation
contact between each subdomain (Tian et al., 2018; Yavuz et conducted by Wang and coworkers, it was aimed at
al., 2015; Zhao et al., 2020; Zitti et al., 2020). determining the force distribution along the blade (Wang et
For the rotation modeling of the turbine rotor, several al., 2019). The MRF and SMM approaches for the rotating
approaches such as the single moving reference frames (SRF) modeling are different from the experimental conditions,
and the multiple moving reference frames have been since the angular velocity (𝜔) and the rotation direction of the
proposed (Tabib et al., 2017). The SRF approach is used for turbine rotor are boundary conditions that are defined in the
modeling simple geometries when the computational domain setup of the problem, i.e., the hydrokinetic turbine rotor
is axi-symmetric. This method can be used when the rotates around a particular axis at a certain 𝜔, which is
computational domain is assumed to rotate. In the case of previously set by the researcher, whereas this is a measured
problems involving multiple moving parts, the stationary result under experimental conditions. To increase the
subdomain is discretized by means of the utilization of the precision in the results of the numerical simulation, 6 degrees
stationary frame equations, whereas the dynamic subdomain of freedom (6-DoF) user defined functions (UDF) method is
is discretized using the moving reference frame equations. In considered as an alternative approach since the turbine rotor
ANSYs Fluent, there are two methods to treat the equations at ω is an output variable within the numerical simulation
the interface, such as the multiple moving reference frame domain (Prakoso et al., 2020).

Vol. 21, No. 2, April 2023 182

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

In turbo-machinery, the 6-DoF model has been commonly in areas near the national electric grid, the electricity
used to analyze systems as cross-flow turbines (Adanta, generated by an AST could be supplied and sold, creating new
Budiarso et al., 2018; Adanta, Hindami & Siswantara, 2018; opportunities and jobs (Madrid et al., 2015).
Prakoso et al., 2019; Siswantara et al., 2018), water wheel Under this scenario, in this study, a comparison between
turbines (Adanta Budiarso, & Warjito, 2020) and open flume 6-DoF and SMM simulation approaches is conducted focusing
turbines (Adanta, Nasution et al., 2020). In hydrokinetic turbine on the performance of an AST prototype. Previous
rotor modeling, the 6-DoF approach has not been thoroughly investigations had used only one approach to study the
investigated. From the authors’ knowledge, only the work system rotation (Bouvant et al., 2021; Zitti et al., 2020). During
carried out by Wang and coworkers was found to report the 6- the CFD simulations, Unsteady Reynolds Averaged Navier
DoF model in the CFD simulation in order to evaluate the Stokes (URANS) equations coupled with the Shear Stress
performance of a vertical-axis Darrieus turbine (Wang et al., Transport (SST) turbulence model are considered. The
2011). Bouvant et al. (2021) reports an interesting work on the validation of the numerical results is carried out by means of a
geometrical optimization of an AST as a hydrokinetic turbine. case reported in the literature (Zitti et al., 2020) and
The results of the research were compared to the results experimental data obtained in the current work.
available in the literature, providing significant improvement
on the topic, especially on the methodology used for the 2. Materials and methods
optimization of the AST geometrical configuration. In the
referenced research, the CFD numerical simulation was 2.1. Archimedes screw turbine geometry
improved with respect to that one proposed by Zitti et al.
The main parameters taking part of an AST are the length of
(2020), when substituting the constant ω assigned to the the shaft (𝐿), the number of blades, the blade pitch (𝑝), the
turbine by a specific and a more realistic UDF function.
diameter ratio (inner diameter/outer diameter, i.e., 𝐷𝑖/𝐷𝑜)
Furthermore, the CFD results were extended with the use of and the inclination angle of the blade with respect to the axis
the response surface methodology. This led to the
(𝛼). These factors are shown in Figure 1. In this study, the AST
identification of an optimal configuration, which was parameters were set to be the same ones as those ones used
compared with other CFD, and experimental tests obtained
by Zitti et al. (2020), so that the obtained numerical results
from the literature. were validated by using the experimental data reported in the
It must be noted that Archimedes screw is an ancient
referenced study. The AST parameters utilized in the
technology and has been mainly used as a pump in several conducted numerical studies are listed in Table 1.
applications; nonetheless, in the last decades its use in
different fields has been increased (Lisicki et al., 2016; Waters
& Aggidis, 2015; Yu et al., 2016). While utilized in hydropower
generation, AST allows to take advantage of water flows in
sites with low heads and large flows to produce electricity
(Dellinger et al., 2018). The use of the AST for electricity
generation has a lot of technical, environmental and social
advantages. The principal technical advantage is the high
efficiencies achieved, from 60% or 80% (Hawle et al., 2012) to
92% in inclined-axis AST installations (ANDRITZ Hydro, 2012).
Concerning the environmental dimension, this kind of turbine
is considered as a renewable and an alternative energy source Figure 1. Main geometric parameters
to carbon-based power generation systems, contributing to involved in the design of an AST.
the reduction of greenhouse gas emissions. Moreover, these
turbines are able to take advantage of kinetic energy as well as Table 1. Geometric parameters of the studied AST.
the water potential energy, leading to a minimal
Parameter Value and units
environmental impact. On the other hand, the AST generally
Shaft length (𝐿) 320 mm
operates at a low rotational speed, causing reduced morbidity
Number of blades 1
and mortality to aquatic species (Kozyn & Lubitz, 2017). Finally,
Blade pitch (𝑃) 160 mm
on the social side, AST can provide electricity to isolated
Outer diameter (𝐷𝑜) 100 mm
communities with an acceptable investment, taking profit of
Diameter ratio (𝐷𝑖/𝐷𝑜) 0.4
its relatively low cost and simple construction. In addition,
Blade inclination (𝛼) 70°
micro-hydropower would increase live quality. Furthermore,

Vol. 21, No. 2, April 2023 183

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

2.2. Rotating models approaches The 6-DoF approach can be used to predict sequentially
2.2.1. Sliding mesh model (SMM) the angular positions according to the turbines mass
The SMM is a special and simplified case of the dynamic mesh properties (mass and moment of inertia) in real time. The 6-
model (DMM), where the nodes in the mesh are rigidly moved DoF method computes the AST 𝜔 based on forces and
in a specific dynamic mesh zone. DMM allows to configure the moments acting on the rotor. Some examples regarding the
relative movement of the boundaries in a cell zone (ANSYS application of the 6-DoF model can be found in Table 3.
Inc., 2018a). The solution process starts from a computational
domain, which is subdivided into a moving and a static part, 2.3. Numerical simulations
with respect to the basic frame of reference. These two The performance curve for a hydrokinetic turbine is obtained
domains are separated by the sliding interface to ensure a by representing the power coefficient (𝐶𝑝) versus (vs.) the tip
conservative distribution of the fluxes going from one speed ratio (𝑇𝑆𝑅 or 𝜆). It is important to note that 𝐶𝑝 is the
subdomain to an adjacent one, as the mesh faces on each side ratio between the power generated by the AST and the total
do not usually match. In SMM, the mesh zone must be energy contained in the flow passing through it. In turn,
configured in a way that enables the non-conformal interfaces 𝜆 characterizes the AST rotational speed regarding the inflow
to be in permanent contact with each other. This is the reason velocity.
it is called SMM, i.e., interfaces slide among them. The mesh 𝐶𝑝 and 𝜆 can be computed as described by Equations
moves like a rigid body (ANSYS Inc., 2018b), and even knowing (3) and (4), respectively, where 𝑃 is the turbine power output,
that the mesh is moving in space, the nodes that define the 𝜌 stands for the density of water, 𝐴 is the cross-flow surface of
cells do not deform it. A fixed rotational speed is imposed on the turbine (𝐴 = 𝜋𝑅2 ), 𝑉 refers to the water inlet velocity and
the rotor by the SMM approach before solving the governing 𝑅 is the turbine radius.
equations. Some examples of the use of SMM in numerical
𝑃
simulation can be found in Table 2. 𝐶𝑝 = (3)
0.5𝜌𝐴𝑉 3

2.2.2. Six degrees of freedom (6-DoF) method TSR = λ =

ωR
(4)
V
The 6-DoF model has many capabilities and is used in
simulations of different nature and complexity, since the 6-
To compare the SMM and the 6-DoF models, the simulations
DoF model utilizes the forces and moments acting on the body
were configured in the same way, the rotating speed the
to know the actual movement state. In Equation (1), the
differentiating element. In each simulation, a rotating domain
translational movement for the center of mass of the body is
or internal domain (cylindrical body, which includes the turbine
calculated with respect to the inertial reference frame (ANSYS
footprint) was defined, as well as a fixed domain or external
Inc., 2018a).
domain (parallelepiped of length, width, and height equal to 2
m, 0.6 m and 1 m, respectively), which encloses the rotating
⃗V̇ = 1
⃗⃗⃗⃗G
∑F (1) geometry and contains the boundaries. The fixed domain was
m
designated to represent a water channel, whose general
where v⃗ ̇ , 𝑚 and ⃗⃗⃗⃗
FG refer to the translational motion for the dimensions and the dimensions of the rotating body are
body, the mass of the body and the vector of forces, represented in Figure 2. The turbine shaft was positioned at 0.5
respectively. m from the top surface of the domain and at 0.3 m from each of
In turn, Equation (2) is used to compute the rotation the walls. The dimensions for the water channel representation
and the AST geometry were the same as those ones reported by
movement of the body, applying body coordinates; where ⃗⃗⃗⃗⃗ ωḂ
Zitti et al. (2020), aiming at achieving comparable results. The
is the vector of 𝜔, 𝐿 stands for the matrix with the inertia
3D CFD simulations were developed in a parallel mode with 16
moments and ⃗⃗⃗⃗⃗ MB is the moment vector of the object (ANSYS
CPU cores, configured in a LENOVO Thinkstation P520 and
Inc., 2018a).
equipped with an Intel Xeon W2145 @ 3.7 GHz processor and a
64 GB RAM memory. Each time step was configured to be run for
⃗ ̇ B = L−1 (∑M
⃗ω ⃗⃗⃗⃗⃗B − ⃗⃗⃗⃗⃗
ωB × Lω
⃗⃗ B ) (2) a maximum of 30 iterations with a criterion of convergence of
10−4 for all the residuals.

Vol. 21, No. 2, April 2023 184

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

Table 2. Reports concerning the use of SMM.

Analyzed
Reference Goal Setup Concluding remarks
system
Improving
Bienz et al. F1 braking 3D numerical – SMM provided better results than MRF in
aerodynamics in
(2003) systems MRF and SMM braking systems analysis.
F1 cars
Comparing the Each analyzed turbine had a different
3D numerical –
performance of 3 optimal inclination.
Dellinger et Inclined Comparison with
AST with different
al. (2019) AST experimental
number of blades 5-blade AST showed the best
results – SMM
and inclination performance.
3D numerical – SMM described the system in a better
Correctly
Gullberg and Cooling Comparison with way.
simulating the
Sengupta fan for a experimental
efficiency of a
(2011) car results – SMM MRF could not properly simulate the fan
cooling fan
and MRF pumping capacity.
Presenting the
correlation 3D numerical –
Truck CFD simulations with SMM were able to
Gullberg et al. between Comparison with
cooling predict, in an accurate way, the fan
(2013) simulations and experimental
fan characteristic curve of the pressure rises.
experiments of results – SMM
axial cooling fans
3D numerical –
MRF grooves took advantage of the
Proposing an Comparison with
strengths of rotative wall model.
Hobeika and approach for experimental
Vehicle
Sebben improving the results –
wheel MRF allowed to get satisfactory results in
(2018) simulations of Proposed MRF
wheel system simulations without the
rotating wheels grooves for
computational cost ascribed to SMM.
rotating model
The rotating methods were compared to
Wind previous numerical results where the
Comparing 3D numerical –
turbine at FAST methodology was used.
Tabib et al. rotating methods MRF, SMM and
an
(2017) for wind turbine RANS based
industrial The SMM showed the best results,
simulation actuator line
scale followed by MRF and the RANS based
actuator line.
Using SMM, the 𝐾 − ɛ, 𝐾 − ɛ RNG and
SST-Kl turbulence models were
Comparing
3D numerical – contrasted.
turbulence models
Sammartano Cross flow Comparison with
for the simulation
et al. (2016) turbine experimental When comparing with experimental
and the design of
results – SMM data, the best results were found using
cross-flow turbines
the SST-Kl model, followed by the 𝐾 − ɛ
RNG and the 𝐾 − ɛ models.
Note: Initials in table stands for: F1 (Formula 1), RANS (Reynolds-averaged Navier–Stokes equations), RNG (Re-
Normalization Group) and SST-Kl (Shear Stress Transport – Kato Launder production limiter).

Vol. 21, No. 2, April 2023 185

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

Table 3. Reports concerning the use of the 6-DoF model.

Analyzed
Reference Goal Setup Concluding remarks
system
Studying the effect
of the channel By performing simulations with 0, 10, 15,
Adanta,
Water wheel inclination angle 3D numerical – 6- 20 and 25° of channel inclination, higher
Budiarso and
turbine in the DoF efficiencies were obtained with the
Warjito, (2020)
performance of channel at 10° of inclination.
the system
6-DoF model was a great tool in the
study of systems whose operating
Analyzing and
conditions depend entirely on their
Adanta, Open flume proposing the 6-
3D numerical – 6- interaction with the fluid.
Nasution et al. turbine DoF method as a
DoF
(2020b) (OFT) tool in OFT
The 6-DoF model in OFT simulations
simulations
delivered better patterns than those
ones in SMM.
Comparing the Better results were found for the turbine
Adanta, system 3D numerical – designed with traditional blades.
Budiarso, et Cross-flow performance Comparison with
al. turbine when using NACA experimental The 6-DoF model was used to obtain an
(2018) and traditional results – 6-DoF adequate description of the turbine
profiles impeller rotation.
Developing CFD simulations with the 6-
Investigating the
DoF model, linear and non-linear
Go and Ahn behavior of the 3D numerical – 6-
Tow-fish hydrodynamic damping coefficients
(2019) system using CFD DoF
could be determined more precisely
methods
than by using other methods.
The 6-DoF method used in CFD could
Throttleable calculate aerodynamic forces and
Simulating the
Hopfe et al. Ducted 3D numerical – 6- moments significantly well and quite
trajectory of a
(2013) Rocket DoF close to what would be obtained in
ramjet missile
(TDR) experimental studies, leading to cost
reduction.
Comparing the Comparing with experimental results
system behavior obtained by Sammartano et al. (2016),
Prakoso et al. Cross-flow 3D numerical – 6-
while using the 6- deviations of 6.8% and 12.4% were found
(2019) turbine DoF and SMM
DoF and SMM for the 6-DoF and SMM simulations,
methods respectively.
Analyzing the
The 6-DoF model allowed to study the
Qu et al. emergency 3D numerical – 6-
Jet behavior of the body while it is
(2015) landing on water DoF
submitted to the impact forces.
of a jet
Note: initials in the table stands for NACA (National Advisory Committee for Aeronautics)

Vol. 21, No. 2, April 2023 186

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

Figure 2. General dimensions and boundary conditions.

The entire set of simulations were carried out using ANSYs For the 6-DoF simulation, the AST footprint and the entire
19R3 and ANSYs workbench, allowing all the necessary steps domain were allocated in a way that with the water coming
to perform the simulation. The CAD geometry was generated from +X to -X, the generated momentum would make the
in ANSYs design modeler and was then connected to the turbine rotates around the +X axis (one DoF rotation); the
ANSYs meshing module. In turn, the generated mesh was inertia tensor for the 6-DoF UDF configuration was computed
imported in ANSYs Fluent. with the help of SolidWorks software. The AST is limited to
The 𝐾 − 𝜔 SST turbulence model was used to solve the 3D rotating around the X-axis, while the other DoF is not available.
transient Reynolds-averaged Navier-Stokes equations and the The specifications of the AST, which depend on the geometry
SIMPLE scheme was set for the pressure-velocity coupling. and material of the turbine, are shown in Table 4. Small values
This turbulence model is commonly utilized for hydrokinetic of the moment of inertia promote the startup of the AST.
turbine modeling since it has been demonstrated to have However, excessively small values will impair the stability of
better performance for complex flows, including adverse the AST rotor rotation. For determining the moment of inertia,
pressure gradients and flow separations, as occurs in iron cast was selected, and the material density was 7150 Kg/m3.
horizontal-axis hydrokinetic turbines. This turbulence model
offers an improved prediction of adverse pressure gradients in Table 2. Specification of the 6-DoF body.
the near wall regions when compared to standard 𝐾 − 𝜔 and
𝐾 − 𝜀 models (Menter, 1994). The 𝐾 − 𝜔 SST turbulence Parameter and units Value
model has been widely adopted by several researchers for the Mass [kg] 1.107
design of hydrokinetic turbines (Chitrakar et al., 2020). Moment of Inertia [kg m2] 0.00047
For the channel inlet, considering that rivers in developing Initial 𝜔 [rad/s] 0
countries do not usually have flow velocities higher than 2 m/s Initial center of mass (x, y, z) (-0.173, 0, 0)
(Andreadis et al., 2013), a velocity inlet of 1 m/s was used. For [m]
this boundary condition, a default turbulence intensity of
10% was implemented. At the domain outlet, a relative pressure Once the configuration was completed and the simulation
of 0 Pa and a backflow turbulent intensity of 5%, was imposed was run, the 6-DoF model began to accelerate the body
as the pressure outlet boundary condition. For the walls because of the interaction between the fluid and the turbine
(including the AST surface), no slip shear condition was enabled. walls, and the 𝜔 rose to a stable and maximum value. From
For the SMM simulation, 𝜔 of the rotating body (including that point, a pre-load was set to the 6-DoF UDF, and the
the turbine footprint) was varied from 1.2 rad/s to 24 rad/s, turbine slowed down until the condition of 0 rad/s was
with steps of 1.2 rad/s. Under this 𝜔 range and considering the reached. The collected data from the imposition of the pre-
inlet velocity of 1 m/s, a 𝑇𝑆𝑅 range between 0.06 and 1.2 was load allowed to draw the 𝐶𝑝 vs. the 𝑇𝑆𝑅 curve. As the
analyzed. Such 𝑇𝑆𝑅 range leads to a good representation of simulation has only one rotational DoF, it was not necessary
the AST performance curve. For the 6-DoF simulation, the to employ any of the mesh methods found in the setup of
angular position of the rotating body was registered during dynamic mesh options of Fluent software; that allows to
each time step, making possible to compute the define the rotation body as a rigid body whose rotational
instantaneous 𝑇𝑆𝑅 and to achieve a performance curve in the movement depends on the total momentum acting over the
same range as in the SMM simulation. footprint surface. To ensure the correct transfer of information

Vol. 21, No. 2, April 2023 187

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

between the stationary and the moving domains, an interface (f −f )

((f 3−F2 ))
boundary condition was employed, which also facilitated P= 2 1
(5)
ln(r)
convergence.
E21 = f2 − f1 (6)
2.3.1. Mesh independence
Mesh independence study is conducted in order to achieve a E32 = f3 − f2 (7)
mesh with several elements that gives results close to reality
and keeps a reasonable computational cost. As the 6-DoF Once 𝑃 has been calculated, the estimate of f0 is possible
simulations require better discretization conditions than the to be determined, as shown in Equation (8). The term r refers
SMM simulations, the convergence studies were carried out to to the grid refinement ratio and could be calculated as 𝑟 =
the numerical simulation results obtained by the application ℎ3 /ℎ2 = ℎ2 ⁄ℎ1 , where ℎ𝑖 is the number of elements in each
of the 6-DoF method. The mesh and time step configurations mesh. Although in this study the value of 𝑟 is not constant, 𝑟 is
under which the results convergence is achieved were also assumed to be equal to 2, since this value is near to the one in
utilized for the simulations using the SMM approach. Due to the meshes used and is higher than 1.3, representing the
the complex geometry of the AST, an unstructured-tetrahedral minimum value found by Roache (1998), under which good
mesh was created for the entire domain. Additionally, in order results were obtained.
to determine the influence of the spatial discretization, the
(E21 )
Richardson extrapolation method was used. For this purpose, f0 = f1 − (8)
rP −1
three different meshes were created and named coarse,
medium and fine mesh. The target parameter for the An additional and widely used concept is the Grid
comparison of the meshes was the area under the curve Convergence Index (GCI), which brings an assessment of the
formed by representing 𝐶𝑝 vs. 𝑇𝑆𝑅. The Richardson discretization error among meshes. Equations (9) and (10)
extrapolation is based on Taylor’s series and allowed us to represent the way in which the GCI is calculated for each pair
obtain an improved estimate of the numerical result in of mesh. In the referred equations, 𝐹𝑠 refers to a safety factor,
derivatives, integrals or differential equations (Richardson, which was set at 1.25 (Xing & Stern, 2010).
1911). The generalized Richardson extrapolation was well
presented by Roache (1994), standing a way to estimate the 𝑓 −𝑓
|( 2 3)|
𝑓2
error associated with the spatial discretization in CFD 𝐺𝐶𝐼23 = 𝐹𝑠 ∗ (9)
𝑟 𝑃 −1
simulations, which is considered as one of the numerical error
main sources. This method has been successfully used in 𝑓 −𝑓
|( 1 2)|
𝑓1
several CFD studies. For example, Prakoso et al. (2019) and 𝐺𝐶𝐼12 = 𝐹𝑠 ∗ (10)
𝑟 𝑃 −1
Siswantara et al. (2018) applied the Richardson extrapolation
in order to analyze the Grid Convergence Index (GCI) for the Finally, the results obtained within the used meshes must
domain of a cross-flow turbine. In turn, Franke and Frank be checked to be in an asymptotic region of convergence. For
(2008) used the mentioned procedure to analyze its suitability this purpose, Equation (11) is used. A value close to 1 is
for problems related with computational wind engineering, expected to be achieved.
achieving good results.
One of the benefits ascribed to the Richardson 𝐼=
𝐺𝐶𝐼23
≈1 (11)
𝑟 𝑃 ∗𝐺𝐶𝐼12
extrapolation is the possibility of computing an estimate of the
exact value (f0 ) for the analyzed variable, i.e., without the error
Figure 3 shows the results of the simulations (the evolution
due to the discretization. In order to calculate f0 , the
of 𝐶𝑝 vs. 𝑇𝑆𝑅) conducted during the mesh independence test.
convergence coefficient (𝑃) must be calculated using the
For the number of elements used, the variation in the results
expression presented in Equation (5), where f3 , f2 , and f1 are
due to the number of elements in the mesh can be observed
the area under the curve of the coarse, medium and fine mesh,
to be not significant. In Figure 4, the asymptotic tendency of
respectively. For the rest of the section, the sub-indexes 3, 2
the results obtained with the used meshes is illustrated. This
and 1 stand for the coarse, medium and fine mesh, in that
agrees with the value of 𝐼 = 1.0011, which was obtained for
order, being established a normalized grid spacing. A helpful
the asymptotic convergence analysis. The 𝐺𝐶𝐼23 and 𝐺𝐶𝐼12
step in the calculus of 𝑃 is the estimation of the difference for
values were 0.22% and 0.076%, respectively, showing that the
the solution variation between meshes. The calculus of the
grid independence was achieved. Thus, the medium mesh
above-mentioned parameter is obtained as described in
was selected for the study simulations. Medium mesh was set
Equations (6) and (7).

Vol. 21, No. 2, April 2023 188

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

with a first layer thickness of 1.58𝑥10−3 m, value that allowed good indicator of convergence. A representation of the medium
to obtain a maximum dimensionless wall distance (y+) of mesh used in the numerical study is illustrated in Figure 5.
90.37, which is found in the range of 30 and 300, making
possible to use the wall functions approach (Nieto et al., 2015). Table 3. Mesh independence test.

Parameter Mesh
Coarse Medium Fine
Number of nodes 42873 82130 153303
Number of 218100 419907 794911
elements
Max skewness 0.659 0.651 0.642
Max aspect ratio 4.698 3.894 4.838
Min orthogonal 0.341 0.349 0.358
Quality
Error with respect 0.325 0.113 -
to the fine mesh
(%)

Figure 3. Power coefficient (Cp) vs. tip speed ratio (λ) for
the different meshes.
Figure 5. Mesh is used for numerical simulation.

2.3.2. Time step independence

As in the mesh independence study, the time step independence
was developed by using 6-DoF simulations. The Richardson
extrapolation process was adapted to analyze the time step
influence on the simulation results; the test was carried out for
three different time steps (0.00025 s, 0.0005 s and 0.001 s) and in
order to find the convergence time step Index (CTI), the analyzed
time steps were normalized to be 1, 2 and 4, respectively. CTI is
the reciprocal of the GCI (Prakoso et al., 2019) and in this study, a
value of 2% or less for the CTI was chosen as a criterion of
selection for the definitive time step. The 𝐶𝑝 vs. 𝜆 curves for each
of the above-mentioned time steps are represented in Figure 6.
Little differences could be observed at the end of the curves;
however, a general good concordance in the data behavior was
also found. CTI12 and CTI23 were 1.69% and 2.28%, respectively;
Figure 4. Richardson extrapolation for the mesh satisfactory results for CTI were noted in the scale of Figure 7,
independence. which covers a variation of less than 0.014. A time step of 0.0005 s
was used in all the simulations. Additionally, when analyzing
Details for the different meshes are presented in Table 5. The asymptotic convergence for the time step, a value of 0.9952 was
meshes exhibited excellent quality parameters and the errors obtained. This value is close enough to 1 as to consider that the
between the mesh results were less than 1%, which is also a selected time step is in the asymptotic convergence region.

Vol. 21, No. 2, April 2023 189

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

was verified with a flow watch air and water speed flow meter
(FW450 with a resolution of ±0.01 m/s). The turbine was 3D
printed with fused deposition modeling, using a layer thickness
of 0.2 mm and polylactic acid (PLA) as the printing material.
To measure the power generated by the turbine, a torque
sensor Futek TRS 605-FSH02057 with encoder was used. This
device has measurement ranges of 0-50 Nm and from 0 to 7000
RPM. The sensor is connected to the rotor and to the DC
motor. The motor is energized so that its direction of rotation
is opposite to that of the turbine, simulating the load of an
electric generator and allowing operating data to be obtained
at different ω.
The other components of the measurement and data
collection system correspond to mechanical couplings, a
watertight casing and supports of the system. In Figure 8, the
setup used for the characterization of the turbine is illustrated.
Figure 6. Power coefficient (Cp) vs. tip speed ratio (λ)
for the different time steps.

Figure 8. Setup used for the experimental

characterization of the turbine.

3. Results and discussion

The final simulations were developed with a mesh of 419907

elements, a time step of 0.0005 s and a water velocity of 1 m/s.
Figure 7. Adapted Richardson extrapolation for the time Figure 9 shows the results when each one of the rotating
step independence. models was used. As observed in the figure, a significant
difference in the AST performance curve is evidenced. A
2.4. Experimental setup maximum 𝐶𝑝 of 0.574 at a 𝜆 of 1.319 was found utilizing the 6-
Experimental tests were conducted in a hydraulic channel. DoF model, in comparison with the maximum 𝐶𝑝 value
The channel has a cross-section of 0.30 m x 0.45 m, with a achieved using the SMM model (0.2755 at a 𝜆 of 0.78). The
length of 7 m. The water circulates through the channel with results obtained with the SMM approach are close to the
the help of centrifugal pumps, which are controlled by variable numerical results presented by Zitti et al. (2020), since the
speed drives that allow to control the flow of water circulating referenced authors achieved a maximum 𝐶𝑝 of 0.238 at a 𝜆 of
through the system. For the tests, a target flow rate of 0.53 m3/s 0.75 with a water velocity of 0.2 m/s, which is a significant
was established. At this flow rate, a speed of 0.5 m/s was lower value for this parameter. Furthermore, concerning the
obtained for the water in the channel. The velocity of 0.5 m/s difference obtained in 𝐶𝑝, it is important to note that Zitti et

Vol. 21, No. 2, April 2023 190

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

al. (2020) used the MRF method for the rotation of the turbine, tions. However, when the rest of the points are analyzed, the
which is less accurate than the SMM method (Gullberg & data provided by the 6-DoF method are the closest to a greater
Sengupta, 2011; Hobeika and Sebben, 2018; Tabib et al., 2017). number of the experimental data obtained.
In Figure 9, the experimental data are found to be better The solid lines in Figure 9 correspond to cubic fits
described when the 6-DoF model is employed. The fit of the conducted to each one of the CFD simulation results. The
SMM results is like those ones presented by Zitti et al. (2020), equations for the regressions are listed in Table 6.
being the experimental data larger than those ones obtained The discrepancies between experimental and numerical
from the simulations. The referred authors found an results can be caused by several factors. As a matter of fact, the
explanation in the flow confinement. Nevertheless, bearing frictional losses of the turbine and the effect of the
considering the better fit of the simulation data obtained by support on the flow field were not included in the simulation.
the 6-DoF method, the error associated with the utilization of Additionally, the support used in the experimental test can
the rotating model seems to be the most probable cause. produce an effect on the AST hydrodynamic performance. The
For 𝜆 values less than 0.4, the experimental data obtained numerical results have smaller values compared to the
are in the middle of the results for the 6-DoF and SMM simula- experimental ones.

Figure 9. Comparison of rotating models and experimental data.

Power coefficient (Cp) vs. tip speed ratio (λ).

Table 4. Cubic fits for CFD results.

Data Cubic fit R2

6-DOF simulation 𝐶𝑝𝑑𝑜𝑓 = 0.0914 ∗ 𝜆3 − 0.5966 ∗ 𝜆2 + 1.07 ∗ 𝜆

(10) 0.9989
results. − 0.01321

SMM simulation 𝐶𝑝𝑠𝑚𝑚 = −0.106 ∗ 𝜆3 − 0.269 ∗ 𝜆2 + 0.624 ∗ 𝜆

(11) 0.997
results. − 0.00232

CFD results (MRF) Non-

𝐶𝑝𝑚𝑟𝑓 = 0.0656 ∗ 𝜆3 − 0.5166 ∗ 𝜆2 + 0.6653 ∗ 𝜆 (12)
(Zitti et al., 2020). registered.

Vol. 21, No. 2, April 2023 191

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

4. Conclusions Funding

The results obtained in this study showed a good concordance The authors received no specific funding for this work.
with the information reported in the literature. In this regard,
the behavior of the AST system was described in a better way
when the 6-DOF model was used. References
Even though a higher precision was found by using the 6-
DOF method, the SMM method continues to be a valuable Abutunis, A., Hussein, R., & Chandrashekhara, K. (2019). A
option for the development of CFD analysis of hydrokinetic neural network approach to enhance blade element
turbines. The SMM method can be used in order to estimate momentum theory performance for horizontal axis
specific operation points with less computational cost, hydrokinetic turbine application. Renewable Energy, 136,
leading to faster design processes. In turn, the 6-DOF method 1281-1293.
could be a good option to be employed in the last stages of https://doi.org/10.1016/j.renene.2018.09.105
the turbine design, as more reliable data are needed. On the
other hand, when a complete estimate of the turbine Adanta, D., Budiarso, B., Warjito, W., Siswantara, A. I., &
performance curve is required, the 6-DOF method could work Prakoso, A. P. (2018). Performance comparison of NACA 6509
it out in a shorter time, since by using the SMM method each and 6712 on pico hydro type cross-flow turbine by numerical
operation point must be set and computed until good method. Journal of Advanced Research in Fluid Mechanics and
convergence conditions are reached. Thermal Sciences, 45(1), 116-127.
When comparing the MRF simulation results obtained by Retrieved from
Zitti et al. (2020) with those ones achieved in this study, the https://akademiabaru.com/submit/index.php/arfmts/article/
MRF approach seems to be a better approach than the SMM view/2188
method; however, there are few experimental points for the
comparison. Additionally, these few points are in a region Adanta, D., Hindami, R., & Siswantara, A. I. (2018). Blade depth
where the SMM method has slightly better behavior. investigation on cross-flow turbine by numerical method.
Therefore, the SMM approach is still considered as a better In 2018 4th International Conference on Science and
approximation method than the MRF approach. When Technology (ICST) (pp. 1-6). IEEE.
compared with the other methodologies, the 6-DoF https://doi.org/10.1109/ICSTC.2018.8528291
simulation was observed to be closer to the experimental
data. Adanta, D., Budiarso, & Warjito. (2020). The effect of channel
Further experimental research is needed to discern the real slope angle on breastshot waterwheel turbine performance by
behavior of the AST turbine under working conditions. The numerical method. Energy Reports, 6, 606-610.
results of these investigations would allow us to validate the https://doi.org/10.1016/j.egyr.2019.11.126
simulation methods used here and select the most suitable
one for the study of this type of turbine. Adanta, D., Nasution, S. B., Budiarso, Warjito, Siswantara, A. I.,
& Trahasdani, H. (2020). Open flume turbine simulation
Conflict of interest method using six-degrees of freedom feature. In AIP
Conference Proceedings (Vol. 2227, No. 1, p. 020017). AIP
The authors do not have any type of conflict of interest to Publishing LLC.
declare. https://doi.org/10.1063/5.0004389

Acknowledgments Andreadis, K. M., Schumann, G. J. P., & Pavelsky, T. (2013). A

simple global river bankfull width and depth database. Water
The authors thank the second call joint R&D projects within Resources Research, 49(10), 7164-7168.
the framework of the agency regional R&D → i, for the financial https://doi.org/10.1002/wrcr.20440
support it provides given to the project “Development of a
hydro-turbine propeller-type kinetics for power generation
ANDRITZ Hydro (2012). Andritz atro - hydrodynamic screws.
electrical”. Additionally, the authors acknowledge the
financial support from the University of Antioquia [Online; accessed October 5, 2020].
(Sustainability Strategy 2020-2021. ES84190067). https://mienergiasolar.es/documents/oi-atro-hydrodynamic-
screws-en.pdf

Vol. 21, No. 2, April 2023 192

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

ANSYS Inc. (2018a). Ansys fluent theory guide, release 19. Gullberg, P., Sengupta, R., & Horrigan, K. (2013). Transient fan
modelling and effects of blade deformation in a truck cooling
ANSYS Inc. (2018b). Ansys fluent user’s guide, release 19. fan installation. In Vehicle Thermal Management Systems
Conference Proceedings (VTMS11): 15-16 May 2013, Coventry
Bienz, C., Larsson, T., Sato, T., & Ullbrand, B. (2003). In Front of Technocentre, UK (p. 219). Elsevier.
the Grid–CFD at SAUBER PETRONAS F1 Leading the https://doi.org/10.1533/9780857094735.5.219
Aerodynamic Development. In 1st European Automotive CFD
Conference, Bingen, Germany. Hawle, W., Lashofer, A., & Pelikan, B. (2012). Lab testing of the
Archimedean screw. Proceedings of the Hidroenergia
Bouvant, M., Betancour, J., Velásquez, L., Rubio-Clemente, A., Conference, Wroclaw, Poland, 23–26, May 2012.
& Chica, E. (2021). Design optimization of an Archimedes screw
turbine for hydrokinetic applications using the response Hobeika, T., & Sebben, S. (2018). CFD investigation on wheel
surface methodology. Renewable Energy, 172, 941-954. rotation modelling. Journal of Wind Engineering and Industrial
https://doi.org/10.1016/j.renene.2021.03.076 Aerodynamics, 174, 241-251.
https://doi.org/10.1016/j.jweia.2018.01.005
Chitrakar, S., Solemslie, B. W., Neopane, H. P., & Dahlhaug, O.
G. (2020). Review on numerical techniques applied in impulse Hopfe, N., Caldas-Pinto, P., & Kurth, G. (2013). Funnel Flight
hydro turbines. Renewable energy, 159, 843-859. Controller for a TDR 6-DoF Simulation Model. IFAC Proceedings
https://doi.org/10.1016/j.renene.2020.06.058 Volumes, 46(19), 125-130.
https://doi.org/10.3182/20130902-5-DE-2040.00084
Dellinger, G., Garambois, P. A., Dellinger, N., Dufresne, M., Terfous,
A., Vazquez, J., and Ghenaim, A. (2018). Computational fluid Kozyn, A., & Lubitz, W. D. (2017). A power loss model for
dynamics modeling for the design of Archimedes Screw Archimedes screw generators. Renewable Energy, 108, 260-273.
Generator. Renewable Energy, 118, 847–857. https://doi.org/10.1016/j.renene.2017.02.062
https://doi.org/10.1016/j.renene.2017.10.093
Lisicki, M., Lubitz, W., & Taylor, G. W. (2016). Optimal design and
Dellinger, G., Simmons, S., Lubitz, W. D., Garambois, P. A., & operation of Archimedes screw turbines using Bayesian
Dellinger, N. (2019). Effect of slope and number of blades on optimization. Applied Energy, 183, 1404-1417.
Archimedes screw generator power output. Renewable
https://doi.org/10.1016/j.apenergy.2016.09.084
Energy, 136, 896-908.
https://doi.org/10.1016/j.renene.2019.01.060
Madrid, M. E., Toro, J. M., & Ortega, S. (2015). Hydrodynamic
Franke, J., & Frank, W. (2008). Application of generalized screws as an alternative for small hydropower generation in
Richardson extrapolation to the computation of the flow across Colombia. WIT Transactions on Ecology and the
an asymmetric street intersection. Journal of Wind Engineering Environment, 195, 15-23.
and Industrial Aerodynamics, 96(10-11), 1616-1628.
https://doi.org/10.1016/j.jweia.2008.02.003
Menter, F. R. (1994). Two-equation eddy-viscosity
Go, G., & Ahn, H. T. (2019). Hydrodynamic derivative turbulence models for engineering applications. AIAA
determination based on CFD and motion simulation for a tow- journal, 32(8), 1598-1605.
fish. Applied Ocean Research, 82, 191-209. https://doi.org/10.2514/3.12149
https://doi.org/10.1016/j.apor.2018.10.023
Nieto, F., Hargreaves, D. M., Owen, J. S., & Hernández, S. (2015).
Gullberg, P. & Sengupta, R. (2011). Axial Fan Performance On the applicability of 2D URANS and SST k-ω turbulence
Predictions in CFD, Comparison of MRF and Sliding Mesh with model to the fluid-structure interaction of rectangular
Experiments, SAE Technical Paper, 2011-01-0652. cylinders. Engineering Applications of Computational Fluid
https://doi.org/10.4271/2011-01-0652 Mechanics, 9(1), 157-173.
https://doi.org/10.1080/19942060.2015.1004817

Vol. 21, No. 2, April 2023 193

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

Prakoso, A. P., Adanta, D., & Irwansyah, R. (2020). Approach for Tian, W., Mao, Z., & Ding, H. (2018). Design, test and numerical
a breastshot waterwheel numerical simulation methodology simulation of a low-speed horizontal axis hydrokinetic
using six degrees of freedom. Energy Reports, 6, 611-616. turbine. International Journal of Naval Architecture and Ocean
https://doi.org/10.1016/j.egyr.2019.11.127 Engineering, 10(6), 782-793.
https://doi.org/10.1016/j.ijnaoe.2017.10.006
Prakoso, A. P., Warjito, W., Siswantara, A. I., Budiarso, B., & Adanta,
D. (2019). Comparison Between 6-DOF UDF and Moving Mesh Tiwari, G., Kumar, J., Prasad, V., & Patel, V. K. (2020). Utility of
Approaches in CFD Methods for Predicting Cross-Flow PicoHydro CFD in the design and performance analysis of hydraulic
Turbine Performance. CFD Letters, 11(6), 86-96. turbines—A review. Energy Reports, 6, 2410-2429.
Retrieved from https://doi.org/10.1016/j.egyr.2020.09.004
https://akademiabaru.com/submit/index.php/cfdl/article/vie
w/3174 Wang, W. Q., Yin, R., & Yan, Y. (2019). Design and prediction
hydrodynamic performance of horizontal axis micro-
Qu, Q., Hu, M., Guo, H., Liu, P., & Agarwal, R. K. (2015). Study of hydrokinetic river turbine. Renewable energy, 133, 91-102.
ditching characteristics of transport aircraft by global moving
https://doi.org/10.1016/j.renene.2018.09.106
mesh method. Journal of Aircraft, 52(5), 1550-1558.
https://doi.org/10.2514/1.C032993
Wang, X., Luo, X., Zhuang, B., Yu, W., & Xu, H. (2011, January). 6-
Richardson, L. F. (1911). The approximate arithmetical DOF numerical simulation of the vertical-axis water turbine.
solution by finite differences of physical problems involving In Fluids Engineering Division Summer Meeting (Vol. 44403, pp.
differential equations, with an application to the stresses in a 673-678).
masonry dam. Philosophical Transactions of the Royal Society https://doi.org/10.1115/AJK2011-22035
of London. Series A, Containing Papers of a Mathematical or
Physical Character, 210(459-470),307–357.
https://doi.org/10.1098/rsta.1911.0009 Waters, S., & Aggidis, G. A. (2015). Over 2000 years in review: Revival
of the Archimedes Screw from Pump to Turbine. Renewable and
Roache, P. J. (1994). Perspective: a method for uniform Sustainable Energy Reviews, 51, 497-505.
reporting of grid refinement studies. https://doi.org/10.1016/j.rser.2015.06.028
https://doi.org/10.1115/1.2910291
Xing, T., & Stern, F. (2010). Factors of safety for Richardson
Roache, P. J. (1998). Verification and validation in
extrapolation. Journal of Fluids Engineering, 132(6).
computational science and engineering (Vol. 895, p. 895).
https://doi.org/10.1115/1.4001771
Sammartano, V., Morreale, G., Sinagra, M., & Tucciarelli, T. (2016).
Numerical and experimental investigation of a cross-flow water Yavuz, T., Koç, E., Kılkış, B., Erol, Ö., Balas, C., & Aydemir, T.
turbine. Journal of Hydraulic Research, 54(3), 321-331. (2015). Performance analysis of the airfoil-slat arrangements
https://doi.org/10.1080/00221686.2016.1147500 for hydro and wind turbine applications. Renewable
energy, 74, 414-421.
Siswantara, A. I., Budiarso, A. P. P., Gunadi, G. G. R., and Warjito,
https://doi.org/10.1016/j.renene.2014.08.049
D. (2018). Assessment of turbulence model for cross-flow Pico
hydro turbine numerical simulation. CFD Letters, (10)38–48.
Retrieved from Yu, H., Janiga, G., & Thévenin, D. (2016). Computational fluid
https://akademiabaru.com/submit/index.php/cfdl/article/vie dynamics‐based design optimization method for archimedes
w/3290 screw blood pumps. Artificial organs, 40(4), 341-352.
https://doi.org/10.1111/aor.12567
Tabib, M., Siddiqui, M. S., Rasheed, A., & Kvamsdal, T. (2017).
Industrial scale turbine and associated wake development-
Zhao, H., Kang, C., Ding, K., Zhang, Y., & Li, B. (2020). Transient
comparison of RANS based Actuator Line Vs Sliding Mesh
Interface Vs Multiple Reference Frame method. Energy startup characteristics of a drag-type hydrokinetic turbine
Procedia, 137, 487-496. rotor. Energy Conversion and Management, 223, 113287.
https://doi.org/10.1016/j.egypro.2017.10.377 https://doi.org/10.1016/j.enconman.2020.113287

Vol. 21, No. 2, April 2023 194

 J. Betancour et al. / Journal of Applied Research and Technology 181-195

Zitti, G., Fattore, F., Brunori, A., Brunori, B., & Brocchini, M. (2020).
Efficiency evaluation of a ductless Archimedes turbine:
Laboratory experiments and numerical simulations. Renewable
Energy, 146, 867-879.
https://doi.org/10.1016/j.renene.2019.06.174

Vol. 21, No. 2, April 2023 195