Proceedings of the ASME 2010 10th Biennial Conference on Engineering Systems Design and Analysis

ESDA2010
July 12-14, 2010, Istanbul, Turkey

ESDA2010-24744

DESIGN OF A FRANCIS TURBINE FOR A SMALL HYDRO POWER PROJECT IN

TURKEY

Kutay Celebioglu Gizem Okyay Mehmet Yildiz

Galaksi Design and Consulting Co. Dept. of Civil Engineering Dept. of Mechanical Engineering
Ltd. Middle East Technical Middle East Technical University
METU Technopolis Ostim University 06531 Ankara, Turkey
06531 Ankara, Turkey 06531 Ankara, Turkey E-mail: [email protected]
E-mail: E-mail: [email protected]
[email protected]

ABSTRACT computational tools are powerful and useful in solution and in

Many hydro power plants, both in small and large scales, visualization; they shorten the design process considerably. In
are being constructed in Turkey. The total potential of these this work a methodology is developed for the design of a
projects reach to 216 billion kWh of feasible energy. However complete turbine.
a method was not yet developed for the design of hydraulic
machinery equipment in Turkey.
NOMENCLATURE
In order to accomplish the hydraulic turbine design H Water head
without any prior information than the design parameters of Q Flow discharge
the hydraulic project, a methodology is developed. This P Power
methodology involves the use of computational tools and it is n Rotational speed
applied for small hydro projects. This methodology is a n s Specific speed
parametric design-optimization procedure which consists of ω Angular speed of rotation
parametric geometry modeling, computational fluid dynamics g Gravitational acceleration
analysis and structural verification. η Turbine efficiency
ηR Runner efficiency
Cp Pressure recovery factor
FS Factor of safety
INTRODUCTION V1 Velocity at draft tube inlet
Turbine equipments constitute the essential part of a hydro V2 Velocity at draft tube outlet
power project from the energy production point of view. Within HL Head losses
this concept Francis type water turbines are important due to
their large range of application in water discharge and head. In
this work a methodology is developed for a complete design of
the mechanical constituents of a Francis turbine in a small METHODOLOGY
scale hydro power project. Based on the water head and discharge values specific to a
hydro power project, general dimensions of the turbine are
Design technique varies depending on the tools in use. determined by in-house computer codes without any other
State-of-the-art techniques using computers are superseding prior information. Initial design geometry depends mostly on
experimental and trial and error methods. Recent

1 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 empirical formulas and on theoretical hydraulic and hydraulic procedure is followed to obtain an optimum working condition
machinery principles [1]. System followed is shown in Fig 1. of the turbine.

Water head and discharge values are specific to a hydro

power project. These two variables are the only input for the 1. Parametric solid modeling
turbine design operating condition. The type of turbine in Parametric models are constructed considering parameters
question is a horizontal shaft Francis turbine. affecting the turbine efficiency and the flow behavior. A
parametric model facilitates the construction of an optimized
model without redesigning the whole geometry [2]. Parametric
solid modeling is a valuable tool because it allows rapid
Inputs: H, Q change of geometry and shortens optimization process
considerably. A solid model is constructed for each turbine
component and the parameters are successively modified after
Inhouse computer codes each CFD and structural analyses. While, for butterfly valve,
different solid model design alternatives are modeled
simultaneously and the disk with the best structural
Solid model performance is selected to be analyzed in CFD.
Not Problems in
adequate CFD results,
structural CFD (coarse mesh) modification
safety of parameters

CFD (fine mesh)

Required η reached
Structural check
Required FS
obtained

Manufacturing Figure 2. BLADE MERIDIONAL PROFILE AND AIRFOIL

PROFILES DEFINED FOR EACH SECTION.
Figure 1. DESIGN METHODOLOGY

Dimensions and parameters needed for the construction of

Initially the turbine specific speed, rotational speed and CAD models are generated by the help of in-house developed
power are determined according to head and discharge input computer codes. The main objective is to determine the
values. Dimensioning of the turbine components follows the dimensions of the runner and to design other components
choice of turbine speed and power. Specification of the turbine accordingly.
is listed in Tab. 1.
Rotational speed, runner diameter and wicket gate height
once decided, runner blade profiles are designed according to
Table 1. SPECIFICATION OF THE DESIGNED the initial assumptions such as zero outlet swirl. For the runner
TURBINE design meridional profile of the blade is determined based on
experience by the help of flow region visualized in blade
H (m) Q (m3/s) P (MW) n (rpm) η ns generator as shown in Fig. 2. The flow region is mapped to
164.84 2.0 2.94 1000 0.92 108 meridional section plane and a NACA profile is defined for
each meridional cross section [3]. Inlet and outlet flow
conditions and angles are provided by Euler equations.
Distribution of blade angles for each meridional section is
The methodology consists of the design of each defined by B-spline curves. This distribution is subject to
mechanical part of the turbine by use of state-of-the-art modification after CFD results to obtain a uniform pressure
parametric CAD modeling, CFD, structural modeling and distribution on the blade. Given the blade angles on each
analysis tools. Preliminary design is followed by simulations meridional section, blade profile is defined for each of them.
and a design loop is formed by interaction of design This method known as inverse design method allows a rapid
parameters with these simulation results. An iterative design optimization of turbine components [4].

2 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Guide vane design follows the design of runner blades. 2. CFD analysis methodology
Guide vane angle is determined according to the inflow angle CFD analyses are performed using commercial CFX v.11
to the runner. Also, the interaction between guide vanes and code. The main components considered in the fluid analyses
stay vanes are investigated even though the main function of are butterfly valve, spiral case, stay vanes, guide vanes, runner
stay ring is structural [1]. Therefore, the stay vane which and draft tube. Mesh type and number of nodes used for each
provides accurate incidence angle to guide vanes is designed in component is enumerated in Tab. 2. Mesh profiles of spiral
the CFD simulation stage. This is accomplished by changing case and runner blade are shown in Fig. 4 and Fig. 6
stay vane blade profile and blade angle and evaluating CFD respectively.
results consecutively. Stay vane, guide vane and runner blade
profiles are generated using a commercial blade generator code
[5].
Table 2. MESH CHARACTERISTICS OF EACH
TURBINE COMPONENT
Design of other components than the blade profiles are
accomplished using low order computer codes. Spiral case is Component Mesh type Number of elements
created for the chosen wicket gate height, inside diameter of
the spiral and predicted outflow angle of the spiral case. The Butterfly valve Tetrahedral 803242
aim is to distribute the flow uniformly around the stay vanes Spiral case Tetrahedral 209602
thus around the runner. This uniform distribution is obtained
Stay vane Hexahedral 240000 (x 12 blades)
by using the theoretical “law of constance of the velocity
moment” [6] through the spiral case. Water entering the spiral Guide vane Hexahedral 240912 (x 24 blades)
case has a moment of velocity around the rotation centerline of
Runner blade Hexahedral 238560 (x 15 blades)
the runner and using this principle the cross-sections defining
the spiral can be designed to ensure evenly distributed water Draft tube Tetrahedral 290224
around wicket gate. Unlike the theoretical design of the spiral
case, draft tube dimensions are based on the outlet of the
runner but also on the alignment of Francis turbine and on
experience data. Both spiral case and draft tube are constructed
of parametrized cross sections which are transformed to 3D
solid model as shown in Fig. 3.

Figure 3. PARAMETRIZED DRAFT TUBE GEOMETRY Figure 4. HEXADRAL MESH OF SPIRAL CASE

In the design of the butterfly valve, parametric model is Given mesh sizes are chosen for optimum simulation
mainly used for CFD and not subject to change significantly effort. For example, runner is the main component effecting
for structural purposes. This is because various design the efficiency of the turbine, so a plot for runner efficiency and
alternatives are investigated initially and the one giving the Hi-Ho are plotted against mesh size as shown in Fig. 5. Here
adequate design factor of safety is chosen. The safest Hi-Ho is the head difference from runner inlet to runner outlet.
alternative is to be analyzed hydraulically and parameters are There is a limit where the solution becomes mesh independent
modified accordingly without causing significant variation in and the solution remains nearly constant but the computational
structural strength. Flow coefficients and torque values are effort drastically increases. So the mesh size is decided such
estimated based on CFD results [7]. that the solution is mesh independent and sufficient for

3 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 engineering purposes but keeping the computation time Combined simulations of successive components are also
minimum. performed in order to check their interaction such as combined
and listed in Tab. 3.

Figure 5. RUNNER EFFICIENCY AND HEAD

DIFFERENCE VERSUS NUMBER OF MESH
ELEMENTS
Figure 6. TETRAHEDRAL MESH OF RUNNER BLADE

CFD analysis is performed for spiral case with mass flow

rate given as inlet boundary condition. CFD results indicate
the necessary adjustments in spiral case geometry to obtain Table 3. SIMULATIONS OF INTERACTIONS
uniform flow distribution over the radial outlet section of the
spiral. Modifications of spiral cross sections are also necessary Spiral case + Stay vanes
to obtain the required flow angle at the spiral exit in order to Stay vanes + Guide vanes
prevent improper angle of attack to stay vanes.
Guide vanes + Runner
Since a complete simulation of the whole turbine would Runner + Draft tube
require very large computation time and effort, each
component is simulated separately for validation of the initial
design. According to these basic simulation results, parameters
are adjusted for a correct flow behavior and suitable velocity Steady state simulations are performed, including the
and pressure distributions. rotational components where a frozen rotor connection is used
as mesh connection method. In this method transient
Guide vane and stay vane positions are adjusted based on simulations are avoided because of their large computational
the CFD results. Exact inflow and outflow angles through stay time requirements. A steady state simulation with multiple
and guide vanes leads to the computed inflow angle to turbine frames of reference is used instead [9]. Stationary frame of
runner. reference is assigned to stator parts. Rotational frame of
reference is assigned to the runner defining its rotational
Considering the head loss from the butterfly valve up to speed. This method requires the interface to be uniform. In
runner, pressure inlet and mass flow outlet boundary order for the guide vane wakes to mix out and give a uniform
conditions are used in runner simulations. This configuration flow at the inlet of the runner, the minimum radius ratio of
is suggested to give better results for rotating machinery guide vane trailing edge to runner leading edge should be 1.04.
applications [5]. Blade NACA profile, leading edge and [10,11] In this design, a ratio of 1.09 is used.
trailing edge angles are modified according to subsequent
simulation results until the required head and efficiency is Rapid simulations for the determination of overall
obtained through the runner. Consequently adequate blade geometry of components are performed using coarser mesh.
loading on suction and pressure sides of the blade and Parts other than the runner blade are simulated using standard
elimination of low pressure zones are attained at the turbine k-ε model for simplicity. For the runner, in order to reduce the
design operating condition. design time span, initial simulations are performed with coarse

4 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 mesh and upwinding scheme in k-ε turbulence model. For fine investigated before performing CFD simulations. The reason is
mesh simulations k-ε and shear stress transport methods are that the worst operating condition on the butterfly valve is the
used as turbulence models to see the effect on the flow closed position with pressure rise due to water hammer. With
separation at the runner inlet. A high resolution advection the help of parametric modeling, geometry is changed quickly
scheme is selected for fine mesh cases. Finer mesh is used for and various design alternatives are analyzed for the design
precise adjustment of blade angles and cross sections as higher pressure. Then, with respect to the result of the FEM analysis
accuracy is requested for these corrections. the models which do not have required structural adequateness
are eliminated.
As other turbine components, draft tube is also simulated
separately for the validation of the initial design. Draft tube
pressure recovery efficiency and flow behavior ability is
investigated. Draft tube efficiency is described by pressure
recovery factor given in Eqn. 1, as the role of the draft tube is
to decrease the water velocity while increasing the pressure at
the connection to tail water.

V 21 V 22 V2
C p= ( − −H L ) / ( 1 ) (1)
2g 2g 2g

Apart from its efficiency the flow behavior is also Figure 7. VARIOUS DESIGN ALTERNATIVES FOR THE
important. In the design operation condition, runner outlet DISK OF THE BUTTERFLY VALVE
swirl is nearly zero and the flow tends to detach from solid
boundaries [6]. Flow separation is prevented by adjusting the
draft tube cone angle.
CFD results transferred to structural analysis constitute an
Differing from the above methodology, the safety check of important part in order to understand the effect of the loads
the butterfly valve is done prior to CFD simulations. Only the due to dynamic flow conditions which was derived from CFD
valve disks with adequate strength to resist very high design computations of turbine parts at different operating conditions.
pressure are chosen and they are modeled with their housing. Forces caused by mechanical properties of the machinery are
Fluid is modeled inside the butterfly valve housing around the also considered. Stress distribution over all parts is calculated
disk. The inlet pipe length is kept limited to decrease and necessary modifications to the geometry are made via the
computational effort by giving developed velocity profile at the same parametric model. After creating a 3-D solid model as an
inlet of the valve. Loss coefficients are determined for varying input to the CFD operator, flow patterns and static pressure
disk angles from fully opened up to nearly closed position. distribution are calculated. The net unit forces (pressure and
shear stress components) on surfaces computed in CFD
Off-design conditions are also investigated against critical analysis is mapped to the simulation domain for the structural
loading conditions. One of the problems in Francis type analysis. Following this routine optimum structural and flow
turbines is the high hydraulic thrust. Use of balancing holes geometries are obtained. The materials are chosen accordingly
and optimization of runner sealing geometry are efficient to meet the safety conditions for the turbine.
solutions to this loading and leakage problem [12]. Different
geometry alternatives are simulated rapidly in CFD and the An example to CAD model, CFD and structural analysis
solution giving the best leakage performance is chosen. interaction is guide vane component. Guide vane profile and
angle is optimized with the CFD analysis. Then the related
parametrical model of the guide vane is created for this profile
points. With the boundary conditions set from the obtained
3. Structural analysis and safety check results of CFD analysis, the model of the guide vane is
Final hydraulic design of turbine components is decided analyzed using finite element tools. If the result of the
upon CFD results. Once the required runner efficiency and structural analysis is not within the safety margins, changes
overall turbine efficiency is reached, structural performance of are made on the model. So the profile of the guide vane should
each component is investigated. be changed. The profile that is modified due to the structural
reasons is optimized again with CFD. The final geometry is
As mentioned under the parametric modeling and CFD obtained when the CFD and structural analyses are obtained
analysis sections, the strength of the butterfly valve disk is within the required criteria.

5 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 RESULTS AND DISCUSSION almost uniform radial flow distribution is obtained at the outlet
Flow passes from the butterfly valve before entering the section of the spiral case except at the baffle vane as shown in
turbine. It is necessary to determine the flow behaviour and the Fig. 10.
loss coefficient of the valve at fully open condition in order to
determine the head at the turbine.
Table 4. BUTTERFLY VALVE COMPUTED LOSS
COEFFICIENTS
Valve opening angle (degree) Loss coefficient
Fully open position 0 0.47
10 0.53
20 1.80
30 4.12
40 9.19
50 19.67
60 68.06
70 323.50
80 4849.00
Closed position 90 -

Figure 8. FLOW AROUND BUTTERFLY VALVE DISK It is observed that the spiral case interaction with stay
vanes are as predicted, pressure distribution is periodic around
stay ring and the flow is directed radially toward guide vanes
with the designed angle of attack as shown in Fig. 11.
Flow around the butterfly valve disk is examined as shown
in Fig. 8 in order to determine any separating flow that will
cause a larger loss coefficient.

Figure 9. STREAMLINES INSIDE THE BUTTERFLY

VALVE AT DIFFERENT CLOSURE ANGLES
Figure 10. DISTRIBUTION OF RADIAL VELOCITY AT
THE SPIRAL OUTLET
Different closure angles as shown in Fig. 9 of the valve are
also simulated. Based on these simulation results butterfly
valve loss coefficients and discharges are determined at
different opening conditions; these values are given in Tab. 4. Loading obtained from the CFD analysis, the thickness of
Hydraulic torque values derived from these CFD simulations the spiral case is optimized for the material selected. For the
are added to other structural forces used in the structural design pressure, the results of the structural analysis are
analysis for a complete definition of operational conditions. obtained and the stress distribution is shown in Fig. 12.
Maximum deformation and maximum equivalent stress were
Spiral case circular cross sections are adjusted according found to be within the safety limits.
to the results of CFD simulations of spiral case alone. An

6 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Figure 13. RUNNER BLADE SUCTION SIDE
Figure 11. PRESSURE DISTRIBUTION AND VELOCITY
VECTORS AT THE MID SPAN OF SPIRAL CASE AND
Low pressure zones at the suction zone as shown in Fig.
STAY VANES
13 are susceptible to be affected by cavitation damage. As
shown in Fig. 14, the pressure distribution is uniform along
The lifetime of a turbine runner is shortened by cavitation. the water passage and the minimum pressure is held above the
It is important to predict the cavitation behavior of the turbine water vapor pressure.
by detecting problems in CFD simulations. It is essential that
the runner converts most of the water energy to rotation
uniformly. But it is also important that the minimum pressure
zones are greater than the value of the vapor pressure. The
vapor pressure is calculated according to the elevation of the
turbine runner and the minimum pressure in the turbine
runner is checked according to this value.

Figure 14. PRESSURE DISTRIBUTION ALONG BLADE

PASSAGE

Theoretical incidence angle of the water at the leading

edge is also corrected according to CFD simulation results.
Improper inflow angles would cause undesirable peak pressure
at the leading edge of runner blade thus causing an unbalanced
Figure 12. STRESS DISTRIBUTION ON THE SPIRAL blade loading [13].
CASE

7 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 as seen in Fig. 17. This has no significant effect on runner
overall efficiency for engineering purposes (η R,SST = 97.3, ηR, k-ε
= 97.4).

Figure 17. Y+ VALUES FOR RUNNER BLADE FINE

MESH
Figure 15. RUNNER BLADE LOADING AT MID SPAN
OF HUB AND SHROUD.
Hub and shroud geometries are important for leakage flow
behavior. Leakage flow is simulated around the runner through
Figure 15 shows the runner blade pressure distribution. It the gaps and it is found that the pressure is balanced at the
is shown that the blade loading is balanced: leading edge runner hub and shroud sides as shown in Fig. 19, thus
cavitation is prevented by a correct incidence angle and reducing the high hydraulic thrust problem.
trailing edge cavitation is prevented by keeping the lowest
pressure above the vapor pressure. [8]

This flow behavior around runner blade is verified by its

simulation with guide vane connection. Simulations of guide
vanes with runner blades permit the determination of runner
efficiency hill curve [9]. Furthermore the guide vane closure
angles at different operation conditions are determined based
on these simulation results.

Figure 18. INTERACTION OF RUNNER BLADES WITH

Figure 16. PRESSURE DISTRIBUTION AT BLADE GUIDE VANES
MIDSPAN GIVEN BY k-ε (LEFT) AND SST (RIGHT)
TURBULENCE MODELS

At the runner outlet, draft tube ensures a pressure recovery

Although two different turbulence closures, namely SST by decreasing the flow velocity towards the tail water
and k-ε, are used for the simulations, slight differences are connection. According to CFD results draft tube operates with
observed between the simulations of different turbulence a pressure recovery factor of 0.78 and without any flow
closures as shown in Fig. 16. The differences arise at about ¼ separation even at the most critical condition where no swirl
streamwise direction from the blade inlet at the pressure side. occurs at the runner outlet.
The reason can easily be seen when the y+ values are checked

8 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Figure 19. PRESSURE DISTRIBUTION ALONG
RUNNER BLADE AND RUNNER LEAKAGE AREA
Figure 21. EQUIVALENT STRESS ON DRAFT TUBE

After obtaining the optimized geometry, it is easy to make

related 3-D model for structural analysis with a proper
thickness for the selected material. After creating 3-D model, CONCLUSION
as a loading the pressure on the wall of the draft tube is In this work the design of all components of a hydro
obtained from the CFD analysis. The results of the structural turbine is accomplished by a methodology developed using
analysis are obtained as shown in the Fig. 21, deformation and computational tools. Solid models generated by parametric
stress values are in safety limits. modeling are optimized and verified by CFD and structural
analyses. The design procedure developed is applied to a
Francis type water turbine of a hydro power project in Turkey.

ACKNOWLEDGMENTS
The computations are performed at Middle East Technical
University Hydraulics Laboratory under the supervision of Dr.
Ismail Aydin.

REFERENCES
[1] Raabe, J., 1985, Hydro Power: The Design, Use, and
Function of Hydromechanical, Hydraulic, and Electrical
Equipment, VDI-Verlag, Verlag des Vereins Deutscher
Ingenieure, Düsseldorf, pp. 50, 171-201, 234-244.

[2] Yildiz, M., Celebioglu, K., 2009, Parametric Modeling of

a Francis Turbine, Proceedings of Nuclear & Renewable
Energy Resources Conference 2009, September 28-29,
Ankara, Turkey, pp.402-407.

[3] Milos, T., Barglazan, M., 2004. CAD Technique Used To

Figure 20. VELOCITY CONTOURS IN DRAFT TUBE Optimize The Francis Runner Design, Scientific Bulletin
MIDDLE CROSS SECTION of the Politechnica University of Timisoara, Transactions
on Mechanics, Special Issue, The 6th International

9 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Conference on Hydraulic Machinery and Hydrodynamics,
Timisoara, Romania, October 21 – 22, 2004, pp. 125-
130.

[4] Wu, J., Shimmei, K., Tani, K., Niikura, K., Sato, J., 2007.
“CFD-Based Design Optimization for Hydro Turbines”.
ASME Journal of Fluids Engineering, Vol. 129, February,
pp. 159-168.

[5] ANSYS Workbench v.11 BladeGen, CFX User Guide.

[6] Krivchenko, G.I., 1986, Hydraulic Machines: Turbines

and Pumps, Mir Publishers, Moscow, pp.97, 102.

[7] Song, X. G., Park, Y. C., 2007, Numerical Analysis of

Butterfly Valve – Prediction of Flow Coefficient and
Hydrodynamic Torque Coefficient, Proceedings of the
World Congress on Engineering and Computer Science
2007, WCECS 2007, October 24-26, San Francisco,
USA.

[8] Okyay, G., Celebioglu, K., Aydin, I., Ger, M., 2009, Design
of a Francis Type Water Turbine Using Computational
Fluid Dynamics Methods, Proceedings of Nuclear &
Renewable Energy Resources Conference 2009,
September 28-29, Ankara, Turkey, pp.388-394.

[9] Keck, H., Sick, M., 2008, “Thirty Years of Numerical Flow
Simulation in Hydraulic Turbomachines”, Acta
Mechanica 201, pp. 211-229.

[10] Jansen, W., 1964, The Design and Performance Analysis

of Radial-inflow Turbines, NREC Report, 1067-1.

[11] Maki, H., Mori, Y., 1973, On the Study of the Flow
Through an Impeller of Mixed and Inward Flow Radial
Turbines , Bull JSME, 16(91), pp.81-92.

[12] Kurokawa, J., Kitahora, T., 1994. Accurate

Determination of Volumetric and Mechanical
Efficiencies and Leakage Behavior of Francis turbine
and Francis Pump-Turbine, Proceedings of XVII IAHR
Symposium, Beijing, China.

[13] Avellan, F., Introduction to Cavitation in Hydraulic

Machinery, Scientific Bulletin of the Politechnica
University of Timisioara Transactions on Mechanics
Special Issue, the 6th International Conference on
Hydraulic Machinery and Hydrodynamics Timisoara,
Romania, October 21-22, pp.11-22.

10 Copyright © 2010 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/12/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use