IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS

Traditional methods for the design of radial-axial hydraulic turbines with

verification in CFD simulation
To cite this article: I Handal et al 2020 IOP Conf. Ser.: Mater. Sci. Eng. 779 012002

View the article online for updates and enhancements.

This content was downloaded from IP address [Link] on 23/07/2020 at 08:54

Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

Traditional methods for the design of radial-axial hydraulic

turbines with verification in CFD simulation

I Handal1,2, V Tkachuk1, A Petrov1and A Protopopov1

1
Bauman Moscow State Technical University

2
E-mail: handal.ignacio15@[Link]

Annotation. In the modern world, the use of hydrodynamic power plants is one of the
fundamental ways of generating electricity. The main part of such power plants is the hydraulic
turbine impeller, therefore, for the developer, the main task is to design it correctly. In this
work are used the traditional methods for designing radial-axis hydraulic turbines and test their
operability with CAD and CFD programs. For that was created a hydro turbine model that will
be installed in El Salvador. Subsequently, in the future, this hydroturbine will be optimized.

Introduction
Hydraulic turbines have been used for over a century and are based on reliable technology. However,
growth and competition in the energy market requires an increase in electricity production at lower
costs, which leads to the need to improve equipment for hydropower plants. Therefore, it is important
to increase the efficiency of hydroelectric power plants, namely the efficiency of turbines.
The efficiency of the turbine can be increased by changing the parameters of the geometry of the
turbine. Sequential tests are needed to increase turbine efficiency. Although it is possible to predict the
characteristics of the turbine using model tests in the laboratory; time and budget constraints,
prototype constraints contributed to the use of CFD tools for turbine optimization.
Each hydropower project requires a different turbine design. Preliminary design is based on
experimental methods, traditional design procedures, and the hydrodynamic theory of machines.
Preliminary model is subjected to CFD analysis to evaluate hydraulic performance. CFD analysis
allows to determine undesirable flow conditions, such as flow separation, cavitation zones, which
cannot be predicted in traditional design methods. This is an iterative process leading to optimal
turbine performance. It should be noted that the maximum efficiency of the turbine is achieved not
under design conditions, but under partial load. The efficiency of the turbine is slightly lower than the
maximum efficiency at the design pressure and the estimated flow: this case is called the optimal
operating condition of the turbine throughout the entire operation. It is worth noting that
approximately 96% of hydraulic losses occur in the blades of the radial axial turbine. Therefore, the
main design work is therefore focused on the rotor blades. Minor changes in the solid turbine model,
especially in the rotor, result in optimized turbine geometry with optimized efficiency and flow
dynamics. Problematic areas can be changed in order to achieve the target efficiency at the design load
and design [Link] related to the design and CFD simulation of hydrodynamic turbines are widely
covered in the literature [1–16].

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

A radial-axial turbineis applicable for a wide range of head and flow rates and may be preferable to
a rotary-blade hydraulic turbine due to its compactness; or bucket hydroturbine due to its higher
efficiency at a better efficiency [Link] those factors make the radial-axial turbine profitable, and it is
the focus of this study. It has been selected to design a turbine expected to work under the following
parametrers.
Table [Link] work parameters

Head Flow Rate, Rotor rotational

Theorical Efficency, % Specific speed,ns
pressure, м m3 /s speed, rpm
72 50.5 300 93 293

The radial-axial turbine design begins with the impeller, and then the guide vanes, fixed vanes and
spiral casing are developed. On this article it has been only descrived the process to obtain the impeller
geometry, nevertheless the rest of the elements have been calculcated accordingly. After obtaining the
theoretical geometry of the turbine, using the CFD software, it is necessary to compare the efficiency
obtained at different opening angles of the guide vane.

Method
Hydraulic profile diagram, Bovet method [1]
Theodore Bovet (1964), based on hydraulic turbines already built, determined a number of speed
factors to obtain the hydraulic profile of a new turbine.
A specific dimensionless speed is used
1/2
Q
nsync  
n0     (1)
 2 gH 3/4
Where n: rad / s, Q: m3 / s, H: (m), g: (m / s2)
For radial axial turbines the specific number of revolutions is usually within the range of
0.1 < n0 <0.8. Figure 1 shows the variations of the profiles according to their n0 which serves as a
guide for the layout of the hydraulic profiles.

Fig. [Link] of the hydraulic profiles.

2
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

To draw the hydraulic profile of the turbine composed of curves i and e, the following equations are
used. It is important to highlight that each of the calculated measures are referred to the basic dimension
ρ2e shown in Figure 3, said measurement is taken as a unit equal to the radius of point 2e [2].

Fig. 2. Limit curves of the volume occupied by the blades.

Fig. 3. Typical dimensions of the rotor flow channel

3
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

Fig. 4. Boundaries of the curves inside and outside the hydraulic profile

As shown in Figure 4, the internal i and external curves e of the hydraulic profile are defined by the
following equation:

y  x x x
 3.08* 1   1   (2)
ym    
Ym values for the inner curve:
0.16
ymi  oi  0.7  (3)
no  0.08

i  3.2  3.2  2  n0  n0 (4)

And for the outer curve:

 e  2.4  1.9  2  n0  n0 (5)

Applying the equation for point 2e, we have:

ye  x  x  x 
 3.08* 1  2e  2e 1  2e  (6)
yme  e  e  e 

Where it is necessary to determine x2e and y2 e .

For x2 e , a constant value is taken that is independent of no :

x2e  0.5 (7)

For y2 e :

y2e  oe  1 (8)

For oe and other quantities that take into account the geometry of the hydraulic profile, the
following values are recommended:
0.493
oe  2/3  Forn0  0.275 (9)
no

4
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

oe  1.255  0.3n0  Forn0  0.275  (10)

b0  0.8  2  n0  n0 (11)

Using equation 9 or 10, oe is calculated, this value is taken into equation, and y2e is calculated,
and using equation 8, we obtain the relation y2e :
yme
y2 e
yme  (12)
y2 e
yme

me  oe  Yme (13)

Determining the actual impeller size [3]
As mentioned above, all dimensions are relativeto  2 e , this measurement will provide the real size
of all measurements of the internal and external curves, and its value is calculated, giving a convenient
value for the flow coefficient 2e expressed by the equation 14.
Q 1
2 e   (14)
n(2e ) u2e
2

If you substitute u2e  2n2e , the following equation is obtained

Q
2 e  ( ) 1/3 (15)
22 n 2e

According to turbines previously built and tested, it is known that the value 2 e , to guarantee
optimum performance, it is between 0.26 and 0.28, so the average value is selected 2e  0, 270.
Similarly, you can determine the true value 1i , in this case, the pressure coefficient u1i is used,
which is calculated as follows.
H
1i  (16)
(u1i ) 2 / 2 g

If u1i is replaced again and 1i = 1.72 is provided based on the experience in equation 16, the
following equation is obtained for 1i
1/2
60  2 gH 
1i    (17)
2n  1.72 
Designing a radial-axial impeller blade in a potential meridional flow (Bowersfeld — Vosnesensky
method) [4]
The method of designing the impeller blade in a potential flow was proposed by Bowersfeld in
1913, and then developed by I. N. Voznesensky. This method has been widely used in engineering
practice for designing impellers of radial-axial hydraulic turbines, since it allows reasonably and fairly
simply to determine the spatial surface of the impeller blade for given conditions.
The method proposes that the flow in the meridionalsection of the rotor is potential; therefore,
before calculating the blade, it is essential to construct the potential flow and determine the field of
constituent meridional speeds in the rotor flow tract obtained by the Bovet method.
In addition, you need to determine the law of change for the product of the variation of the
tangential velocity by the radius cur along some of the border flow lines. It is recommended to give the

5
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

dependency cu r (s) in the form of a straight line or concave curve. It is necessary to assume a
theoretical value of hydraulic efficiency.
The main equation that allows by means of numerical integration to find the shape of the blade is
r 2   cu r
  S (18)
r 2 cm
Using this equation, the angle ΔΘ in the plane corresponding to the given parameters and to the
randomly selected segments ΔS in the current line in the meridional plane of the blade is determined.
The sum of the drag angle in the plane Θ=Σ ΔΘ must be within the limits accepted for rotors of similar
specific speeds. The procedure is graphical analytical and it’s described as follows in order to
complete the next table.
Table 2. Tabulated results for drag angle calculation

r cur Cm r 2   cu r Δs ΔΘ Θ
Point r2 r2ω r2cm (8)prom
(m) (m2/s) (m/s) r 2 cm (m) (rad) (rad)

1 2 3 4 5 6 7 8 9 10 11 12

 Draw the meridional section of the impeller hydraulic tract on a sufficiently large scale. Plot
construction flow lines of the potential flow (not less than 5 lines). Draw the entrance contour of the
blade (In Figure 5 A1A6) and identify the intersection point of the exit contour with the impeller band.
The recommendations for placement of the input contour and point B1 can be determined by analogy
with other impellers of the same specific speed. Then line A1B1 is divided into equal sections of length
8–10 mm.

Fig. 5. Hydraulic turbine profile and inlet and outlet contours

 The graph cu r (s) is constructed for the first current line between limits A1 and B1. For this, the
values cu r for the input contour are determined using the main turbine parameters
gH
 Cu r in   Cu r out  (19)


6
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

It can be assumed that ( cu r ) out=0, which corresponds to a normal speed at the exit, and therefore
gH
 Cu r in  (20)

The input and output contours are in meridional planes and are vortex lines along which
cu r = const. Therefore, the result of the previous equation for the input contour will be the same for
all its points and for all the points of the output contour it will be 0.
Having chosen a suitable scale, the cu r (s) curve is arbitrarily determined, as shown in Figure 5. On
the S axis the initial coordinate (point 0) corresponds to point A1 on the input contour. The section Δs
of the flow channel drawing is transferred to the graph and consequently for each of them is found the
cu r value.

Fig. 6. Variants of dependencies cur = f (s)

 The cm (s) graph representing the change in the meridional velocity along the current line,
obtained from the construction of the potential flow, is constructed. From the starting point of the
coordinates on the s axis for each current line, the point on the input contour is overlaid (A1, A2, ..., A6)

Fig. [Link] of meridional velocities along streamlines

 The table previously seen for the first current line is completed. In space 2, the radius values are
transferred at points A1, a1, б1 ... Б1, making a direct measurement of the plane of the hydraulic tract.
In space 3 the cur values corresponding to these points of the cur (s) graph are placed. The spaces

7
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

4,5,7,8 are completed by means of calculations. In space 6, by the data of the graph cm (s). In space 9,
an average is made between the data in space 8 before and after between lines. In space 10 there are
the values Δsbetween points A1, a1, б1 ... Б1 at real scale. The values to be placed in space 11 result
from multiplication (9) x (10). In spaces 12 and 13, the values Θ of the angles in radians and degrees
in summation are transferred, measuring from the input contour (from point A1) using equation 18.
 Line A1Б1 is constructed in the plane. The horizontal line 0A1 corresponds to the meridional
section in which the entrance contour is located. The meridional surface on which point a 1 is rotated
with respect to point 0A1 at an angle ΔΘa, which is taken from the [Link] the radius ra, the point is
cut on the line 0a1 and the points A1 and a1 are joined. Subsequently, the angles ΔΘб1, ΔΘв1 are placed
and the points corresponding to the radius are cut in the lines, the polyline which represents the current
line A1 Б1 in the plane is found. The central angle between the lines 0A1 and 0Б1 is the drag angle of
the blade.
Fig. [Link] and meridional coordinates of the
surface of the turbine blade

 The calculation is completed for the following current lines. The task is to find sections ΔS A2,
ΔΘa2that, after carrying out the calculation of the table, must obtain the same angles ΔΘ a, ΔΘб, ΔΘвas
those previously obtained in the first line and are contained within the drag angle.

Fig. [Link] and meridional coordinates of the surface of the turbine blade

8
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

Mathematical model. In order to obtain the hydrodynamic modeling, the following equations were
used:
Equation of continuity of the liquid medium:

u x u y u z
   0, (21)
x y z

Where u i — time-averaged projections of fluid velocities on the corresponding axes;

Equation of the change in the amount of motion averaged over time:
 u u  p   v
 i  uj i     Tij  ui u j  , (22)
     
 t x j  xi xi

Where u i p — averaged speed and pressure;

Tij   2sij — viscous stress tensor for incompressible fluid;
v

1  ui u j 
sij     — strain rate tensor;
2  x j xi 
ui u j — Reynolds stresses.
The Reynolds system of equations is open due to the presence of unknown Reynolds stresses. The
system is closed using the k-ω SST turbulence model.
Terms of computer simulation. In the process of thecomputer simulation, a three-dimensional
model of the flow part of the turbine was created, and physical models and parameters of the
computational grid were set for it. The density value remained unchanged throughout the experiment.
For the boundary conditions, the inlet and outlet pressure values were determined to create a pressure
drop equal to the nominal pressure (see Table 3). The rotation speed is preliminarily assigned based on
the initial turbine design conditions. Impeller torque and flow rates measured due to pressure drop
were measured.
Table 3. Initial conditions
Parameter Value
Input stagnation 711125,0 Pа
Outlet pressure 0,0 Pa
Rotational speed 300 rpm

To build the grid, the following construction models were selected:

• Generator of polyhedral cells;
• surface mesh generator;
• Prismatic layer generator.
The grid parameters are shown in Table 4:
Table 4. The parameters of the computational grid
Parameter Value
Base size 50 мм
Prismatic Stretching 1.3
Prismatic layer thickness, relative to the base size 50%
The number of prismatic layers 5
Number of Cells Generated on Impeller 7541332
The number of generated cells on Inlet 5096089
The number of generated cells on Outlet 1024312

9
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

In the impeller, the mesh was thinned 2 times to reduce the calculation error. The resulting mesh in
the cross section of the flow part is shown in Fig. 4:

Fig. 10. The resulting mesh in the cross section of the flowing part

Results
Applying the Bovet method the geometry of the hydraulic tract is obtained using the main design
parameters of the turbine Q= 50 m3/h; H = 72.5 m; P = 33 MWand applying the Bowersfeld —
Vosnesenskymethod, the geometry of the blade is obtained.

Fig. 11. Dimensions and Coordinates obtained

The spiral chamber, the distributor and the discharge tube take their dimensions based on the
nominal diameter of the impeller. The geometry of the simulated turbine is constructed using CAD
programs [5]. The CFD program is used to obtain various parameters of the turbine, changing the
angle on the distributor. The results calculated are shown in Table 5.

10
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

Fig. [Link] model of the turbine composition

Table 5. Parameters obtained

Distributor
Flow Rate Head Torque Power impeller Power fluid
opening Efficiency
(m3 / s) (m) (Nm) (MW) (MW)
(grades)
26 28.93 71.75 378062.63 11.88 20.36 58.80%
30 43.6 69.66 627163.53 19.70 29.79 64.00%
33 46.3 69.89 693619.86 21.79 31.74 68.00%
35 48.55 69.66 723487.34 22.73 33.18 68.40%
38 51.83 69.26 747819.34 23.49 35.22 66.00%
40 53.77 69 758076.03 23.82 36.40 65.00%
43 57.48 68.4 762232.88 23.95 38.57 62.00%
46 60.36 67.85 744574.59 23.39 40.18 58.00%

70,00%
68,00%
66,00%
Efficiency %

64,00%
62,00%
60,00%
58,00%
56,00%
20 25 30 35 40 45 50
Distributor opening
(grades)

Fig. 13. Efficency graph

11
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

70
60
50
Flow Rate
(m3 / s)

40
30
20
10
0
20 25 30 35 40 45 50
Distributor opening
(grades)

Fig. 14. Flow Rate graph

30,00

25,00
Power impeller

20,00
(MW)

15,00

10,00

5,00

0,00
20 25 30 35 40 45 50
Distributor opening
(grades)

Fig. 15. Power on the Impeller Graph

Fig. 16. Pressure distribution in the hydraulic circuit

12
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

Fig. 17. Speed distribution in the hydraulic path

Fig. 18. The presence of cavitation in the impeller

Conclusion
As a result of the experiment, it can be concluded that a Radial Axial Turbine was designed with the
following specifications: Net height of 72.5 m, installation flow 50 m3 / s. For this point of operation,
an overall efficiency of 93% is projected. The behavior of the parameters is that expected for an axial
radial type turbine, in which when the distributor is opened, the flow is increased, having a constant
head, reaching an efficiency of 68.4% at an opening of 35 degrees and a flow rate of 48.55 m3 / s
The traditional design for Radial Axial Turbine impllers design, specifically the Bovet and
Bowersfeld — Vosnesensky Methods, can be applied in the design of a functional, achieving an
acceptable performance according to the efficiency results obtained from the 3D numerical
simulations. Nevertheless, considering that the process is iterative, the results cannot be expected to be
optimal at the first attempt, so it is necessary to repeat the process until the best results are [Link]
conclusion, we must add that the obtained efficiency has a lower value than that of analogues, since
this model has not yet been optimized in CFD packages. However, the power value is acceptable,
which means that further modeling should lead to an increase in efficiency based on optimization
methods of the flow part.

Literature
[1] Bovet , T. Contribucion of the study of trace of blading of a reaction turbine of the Francis type
(1963) New York ASME Winter Meeting 1961, pp. 47–79.

13
Hydraulics IOP Publishing
IOP Conf. Series: Materials Science and Engineering 779 (2020) 012002 doi:10.1088/1757-899X/779/1/012002

[2] Mataix, C. ―Hyraulic Hydromachines‖ (2003) Editorial ICAI

[3] Vasiliev Yu. S. Samorukov I. S. Khlebnikov. S. N. Basic energy equipment of hydro power
plants. Composition and selection of basic parameters (2002) Saint Petersburg State
Polytechnical University
[4] Topazh, G. I. The Basics of the Working Process and Calculation of Hydro Turbines (2011)
Saint Petersburg State Polytechnical University, pp. 139–151
[5] De Andrade Correia, J. A. Mechanical and Hydraulic Design for a Tubular Francis Turbine
(2006)Simon Bolivar University
[6] Busyrev, A.I. Topazh, G.I. Selection of basic parameters and elements of the flowing part of
reactive hydroturbines (2006) Saint Petersburg State Polytechnical University
[7] Laubah D.V. Design of the flow part and the rotation mechanism of the vanes of the guide vane
for the Myatlinskaya HPP (2017) Saint Petersburg State Polytechnical University
[8] Barlit V.V. Hydraulic turbines (1977) Editorial Union ―Vishcha Shkola‖, pp. 360
[9] Kovalev, N. N. ―Hydroturbines‖ (1971), Editorial ―Mashinostroienie‖, Leningrad.
[10] Morales, M. P. Francis Turbines Design in C++ Builder (2000) Costa Rica University, pp. 125–136
[11] Mathematical modeling of the mechanisms of volumetric hydraulic machines. S. Semenov and
D. Kulakov 2019 IOP Conf. Ser.: Mater. Sci. Eng. 492 012042
[12] The effect of design parameters of the closed type regenerative pump the energy characteristics.
N Isaev 2019 IOP Conf. Ser.: Mater. Sci. Eng. 492 012026
[13] Hydrodynamic modelling of the impact of viscosity on the characteristics of a centrifugal pump.
V Tkachuk 2019 IOP Conf. Ser.: Mater. Sci. Eng. 589 012007
[14] The development of the theory of calculation of the hydrodynamic coupling. V Lomakin 2019
IOP Conf. Ser.: Mater. Sci. Eng. 492 012012
[15] N Egorkina and A Petrov 2019 IOP Conf. Ser.: Mater. Sci. Eng. 492 012015
[16] A Petrov 2019 IOP Conf. Ser.: Mater. Sci. Eng. 492 012036

14