Home Search Collections Journals About Contact us My IOPscience

Numerical prediction of a bulb turbine performance hill chart through RANS simulations

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2012 IOP Conf. Ser.: Earth Environ. Sci. 15 032007

(http://iopscience.iop.org/1755-1315/15/3/032007)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 193.198.8.211
This content was downloaded on 08/06/2017 at 22:17

Please note that terms and conditions apply.

You may also be interested in:

CFD analysis of a bulb turbine and validation with measurements from the BulbT project
T C Vu, M Gauthier, B Nennemann et al.

Modern challenges for flow investigations in model hydraulic turbines on classical test rig
C Deschnes, S Houde, V Aeschlimann et al.

Draft tube flow phenomena across the bulb turbine hill chart
P Duquesne, R Fraser, Y Maciel et al.

Experimental investigation of the draft tube inlet flow of a bulb turbine

J Vuillemard, V Aeschlimann, R Fraser et al.

Flow simulation and efficiency hill chart prediction for a Propeller turbine
T C Vu, M Koller, M Gauthier et al.

Numerical analysis of a measured efficiency hysteresis on a bulb turbine model

S Houde, A Carrier, J D Buron et al.

General overview of the AxialT project: A partnership for low head turbine developments
C Deschnes, G D Ciocan, V De Henau et al.

Part 1 Experimental study of the pressure fluctuations on propeller turbine runner blades during
steady-state operation
S Houde, R Fraser, G D Ciocan et al.

Flow field study in a bulb turbine runner using LDV and endoscopic S-PIV measurements
S Lemay, R Fraser, G D Ciocan et al.
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

Numerical prediction of a bulb turbine performance hill chart

through RANS simulations
V Gunette1, S Houde1, G D Ciocan1, G Dumas2, J Huang3 and C Deschnes1
1
Hydraulic Machines Laboratory, LAMH, Faculty of Science and Engineering, Laval University,
Qubec, Canada.
2
Numerical Fluid Mechanics Laboratory, LMFN, Faculty of Science and Engineering, Laval
University, Qubec, Canada.
3
Renewable Energy Technologies, CanmetENERGY, Natural Resources Canada, Ottawa, Canada.

E-mail: [email protected]

Abstract. Within the framework of an international research consortium on low-head hydraulic

turbine flow dynamics, the predictive behavior of Reynolds Averaged Navier-Stokes (RANS)
simulations of the efficiency () hill chart of a bulb turbine is investigated. The paper presents the
impacts of the blade tip gap and the hub gaps on performance predictions.

1. Introduction
Bulb units are installed on low head sites that may undergo significant head variations. For this reason,
they are designed to be double regulated. Because of their wide operating range, the study of bulb
turbine based on Computational Fluid Dynamics (CFD) presents interesting computational challenges.
In 2011, the Hydraulic Machines Laboratory (LAMH) of Laval University initiated a research project
named BulbT, within the framework of an international research consortium [1], to characterize both
experimentally and numerically the flow within a bulb turbine (Figure 1).

RNG RHG RTG

Figure 1. General overview of the BulbT turbine (left) and different runner geometries (right) in order to study
the effect of the blade tip gap (RTG) and the hub gaps (RHG) in comparison to the runner with no gap (RNG).

The first part of the paper presents the methodology used to numerically evaluate the efficiency ()
hill chart based on different heads (H), guide vane angles () and runner blade angles () using a
simplified runner geometry. The second part focuses on studying the effect of the blade tip gap and the
gaps at the hub on the evaluation of the turbine performance using Reynolds Averaged Navier-Stokes
simulations (RANS) and a classical turbulence modelling approach.

2. Methodology

Published under licence by IOP Publishing Ltd 1

26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

2.1. Geometry
The BulbT turbine model has 4 runner blades with a throat diameter (D) of 340 mm. The distributor
includes 16 guide vanes. The draft tube is a two-part structure with a conical section (a half angle of
6.66) and a transition section. All the geometries originated from the engineering drawings of the
model, except for the runner, which was retro-engineered from the actual model. It has been measured
with a laser based coordinate measurement machine to account for geometrical differences. Indeed, as
Nicolle et al. [2] have shown, small differences in runner blade geometry may yield significant
differences in numerical performances. Since the final simulations used only one runner passage, tests
were performed to select a runner blade geometry that represents adequately the hydraulic behavior of
the complete runner. In order to study the effect of the gaps on the turbine performances, three runner
geometries have been generated (Figure 1): 1 runner with no gap (RNG), 2 runner with only the
blade tip gap (RTG) and, 3 runner with only the hub gaps (RHG).

2.2. Simulation strategies

The flow from the intake inlet to the draft tube exit is simulated. The objective is to get the
performance curves at different unit speed (N11), (), and (). It is accepted that monophasic RANS
simulations using eddy viscosity turbulence models, such as k- or SST, provide adequate
performance predictions for low and medium head turbines with fixed runner blade angle [3]. For bulb
turbines, the published results using RANS models show good predictions for runners opening around
the best efficiency points and a significant degradation in the prediction quality compared with
experimental results at larger and smaller runner openings [4]. In the present study, RANS simulations
with k- model and wall functions are used, and the numerical performance parameters are evaluated
based on the IEC model testing standards [5].
outlet Mixing plane interfaces inlet

PT or Q
<Ps>=0
T.I.=1%
T/=1

Extension Draft tube Runner Guide vane Intake

Figure 2. Sketch of the BulbT turbine and simulation boundary conditions.

2.3. Boundary conditions

For the simulations with the RNG, a procedure based on iterations of the total pressure (PT) at the inlet
was used in order to get a total head (H) with the accuracy prescribed by the IEC standards. For
simulations RTG and RHG, the flow rate (Q) coming from the RNG simulations was imposed at the
inlet. Preliminary numerical tests have shown the turbulence level at the inlet has almost no effect on
the global performance of the turbine. Since the test bench includes an upstream settling tank to reduce
the inlet turbulence level, the turbulence intensity (T.I.) was set to 1% with an eddy viscosity ratio
(T/) of 1. At the outlet of the draft tube, a slip wall extension was added as illustrated in Figure 2,
and a static average pressure of 0 was set at the outlet of this component. Mixing plane interfaces were
used between rotary and stationary parts, and between the intake and the distributor. The full turbine
has been simulated at 12 guide vanes angles (-9< -opt < 27by increment of 3) and for three
runner blade angles (-opt = 0, 5.5and 11.5). Several N11 values have been simulated.

2.4. Numerical methodology

All simulations were performed with ANSYS CFX v13.0, a commercial Computational Fluid Dynamics
package (CFD) using an unstructured finite element based finite volume method. The spatial
discretization order is controlled in CFX through a specific code parameter called the blend factor

2
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

(beta). If (beta=0) then the scheme is pure 1st order upwind, if (beta=1) then it is second order centered.
For the current simulations, a first order spatial upwind integration with upwind correction (beta=0.75)
has been used. That choice was found to be a good compromise to obtain low residual values and a
good precision. The influence of the solver residual level on the results was studied during the mesh
convergence test phase. All the turbine components require maximum residual values lower than 10-4,
except for the draft tube, that requires 10-5. The components loss and torque were also monitored
during the simulations. The simulations were judged converged when the requirements on the residual
values were met and the loss and torque fluctuations were below 0.5%.

2.5. Spatial discretization

Grid convergence tests were made for each component of the bulb turbine in order to optimize the
number of elements in the three coordinate directions (X, Y, and Z) independently, based on the
predicted values of component losses and torque. The final hexahedral meshes were created using
ANSYS ICEM CFD HEXA v13.0. For the intake (1.2X106 elements), the convergence test revealed that
a finer resolution in the axial direction is required only in the vicinity of the two piers. The distributor
(6.5X105 elements) appears to be sensitive to the resolution in the axial direction around the guide
vanes and their wakes. For the three different runner geometries (Figure 1), all refinement directions
had significant impact on the results. The convergence test revealed that 4X106 elements are required
for RNG, 5.5X106 for RTG and 4.9X106 for RHG. For the draft tube, converged loss values were
obtained with a 300k elements mesh at the best operating condition. The final mesh used four times
that number to have a safety margin at other operating conditions. Y+ values of 300 are sufficient at
the walls except at the guide vanes, runner blades and draft tube walls that are approximately at 100.

3. Results

3.1. Efficiency hill chart with no runner gaps

Figure 3. Efficiency of the full turbine in function of the flow rate for different values of blade angle at
N11/N11,opt : 0.75 (left) and N11/N11,opt : 1 (right). the presence of the gaps is not considered (RNG).

Since the experimental efficiency hill chart of the BulbT turbine is not yet available, the accuracy of
the numerical predictions presented in Figure 3 cannot be formally assessed. Nevertheless, good
confidence has been gained since the numerical simulations yield performance results quite consistent
with published data for similar bulb turbines (e.g. Gehrer et al. [4]). Since it was known beforehand
that the -opt = 0 was close to the turbine best efficiency condition, the on-cam curve of Figure 3 (in
red) behaves as expected, i.e. decreasing with increasing . Figure 4 shows the effect of N11 and unit
flow rate (Q11) variations for a constant . Once again, the numerical predictions are consistent with
the general behaviour hydraulic turbines with a maximum value reached for one combination of Q11
and N11.

3
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

Figure 4. Efficiency of the bulb turbine versus the flow rate for different values of N11/N11,opt at a constant blade
angle -opt=0. The presence of the gaps is not considered (RNG).

3.2. Impact of the blade tip gap

3.2.1. Blade tip vortex. Simulation results indicate that the inclusion of the runner gaps does not affect
the intake and distributor losses. Hence, the present analysis focuses only on the runner and draft tube
flows. The major difference between RTG and RNG is produced mainly by the blade tip vortex
illustrated in Figure 5. It is created near the leading edge of the blade and evolves at the suction side
near the casing [6]. Figure 5 shows also that this vortex affects locally the pressure distribution on the
suction side of the blade.
RTG RNG RTG RNG

Figure 5. Streamlines (left) and pressure distribution (right) at the suction side of the blade near the casing for
the case including the blade tip and the case without any gap at the operating condition: -opt = 5.5, N11/N11,opt
= 0.75, Q11=1.72.

3.2.2. Impact of N11. The results presented in Figure 6 show that the difference in terms of global
performance predictions between RTG and RNG decreases when N11 increases for a constant . Thus,
for high N11 values, the effect of the blade tip gap on appears to vanish.

Figure 6. Comparison of efficiency versus N11 at a constant blade angle -opt=0and for the cases RNG versus
RTG.

The numerical results indicate that the position of the stagnation line on the blade surface moves
toward the suction side with increased N11 for constant (Figure 7). This observation matches the
expected inlet velocity triangle behaviour.

4
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

N11/ N11,opt=0.75 N11/ N11,opt=1.25

Figure 7. Pressure distribution on the blade surface and wall shear surface streamlines for N11/N11,opt = 0.75 (left)
and N11/N11,opt = 1.25 (right) at a constant blade angle: -opt=0and considering the blade tip gap (RTG).

A modification of the stagnation line position implies a variation in the effective angle of attack on
the blade. Thus, increasing the N11 value, for constant , leads to decreasing the blade angle of attack.
According to the aerodynamic theory on induced drag, the intensity of the wing tip vortex decreases if
the angle of attack decreases [7]. This phenomenon is illustrated in Figure 8 with contours of the
vorticity components in the tangential direction of the blade. Comparing the tip vortex for N11/N11,opt
of 0.75 and 1.25, it becomes evident that its size is decreasing with increasing N11. This observation
tends to support the conclusion stemming from Figure 6.
N11/ N11,opt=0.75

N11/ N11,opt=1.25

Figure 8. Vorticity contours in the tangential direction of the blade for the case N11/N11,opt = 0.75 (up) and the
case N11/N11,opt = 1.25 (down) at a constant blade angle: -opt=0.

3.2.3. Impact of . Figure 9 illustrates the impact of the gap on and P11 for constant values of N11 and
. The difference between RTG and RNG appears constant on the studied range of . The simulations
also indicate that there is almost no displacement of the position of the stagnation line on the runner
blade leading edge within the studied range (-3< -opt < 6). Hence, following the analysis
presented above linking the angle of attack of the blade to the tip vortex, it appears coherent that the
overall intensity and shape of the vortex is not significantly affected by a variation of (Figure 10).

Figure 9. Comparison between RNG and RTG simulations in terms of efficiency (left) and power (right) versus
the flow rate at a constant blade angle -opt=0and N11/N11,opt = 1.

5
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

-opt=-3

-opt=6

Figure 10. Vorticity contours in the tangential direction of the blade for the case -opt =-3(up) and the case -
opt =6(down) at a constant blade angle: -opt=0, N11/N11,opt = 1 and considering the blade tip gap (RTG).

3.2.4. Component loss analysis. The major difference in terms of losses between the RTG and RNG
geometries is inside the runner as illustrated in Figure 11. This difference in runner losses (HRTG-
HRNG)runner is relatively constant in the range of the studied Q11 ( 0.75% of the head). With the current
numerical methodology, the impact of the tip gap on the draft tube losses is low.

Figure 11. Runner and draft tube losses versus the flow rate for the case without any gap (RNG) and the case
including the blade tip gap (RTG) for -opt=0and N11/N11,opt = 1.

As can be seen from contours of the axial velocity at the runner exit (Figure 12), the tip gap
generates local topology variations. But, the mixing plane interface coupling the runner and the draft
tube and the time-averaged nature of the numerical solutions introduces simplifications to the velocity
profile at the draft tube inlet that influences the draft tube losses. Between the runner and the draft tube,
the mixing plane interface is smoothing out the local effect of the tip gap vortex since it relies on a
circumferential averaging of the velocity and turbulence quantities as illustrated in Figure 12c. Thus,
in using this kind of interface, it is difficult to assess the real impact of the blade tip gap on the draft
tube performance.
a b c

Mixing
RNG RTG planes

Figure 12. Axial velocity profiles at the outlet of the runner without any gap (a), including the blade tip gap (b)
and axial velocity profile at the inlet of the draft tube using mixing plane interfaces for the case with the blade tip
gap (c) at the operating condition : -opt=5.5, N11/N11,opt = 0.75 and Q11=1.72.

6
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

3.3. Impact of hub gaps

An analysis of the global performance parameters (Figure 13) shows that the impact of the hub gaps
(RHG) appears smaller than that of the blade tip gap (RTG).

Figure 13. Comparison between no gap, only the blade tip gap and only the hub gaps on efficiency in function
of the flow rate (left) and N11 (right) at a constant blade angle -opt=0.

Figure 14. Comparison between no gap, only the blade tip gap and only the hub gaps in terms of power and
losses versus the flow rate at a constant blade angle -opt=0and N11/N11,opt = 1.

Although the predicted between RNG and RHG is slightly different, the calculated torque
appears unaffected by the hub gaps (Figure 14 left). Thus, the variation associated with the hub gaps
is mainly coming from small differences in the runner and draft tube losses (Figure 14 right). The
comments made previously (section 3.2.4) about the validity of runner-draft tube coupling are still
applicable here, and the present numerical methodology would require further analyses to ascertain the
full impact of the hub gaps on the draft tube flow.

4. Conclusions
Three different runner geometries were created in order to compare cases with blade tip gap, with hub
gaps and without any gap. The blade tip gap study shows that the size and the intensity of this vortex
are affected by the variation of N11 but not significantly by the variation of . The hub gap study
predicts a lesser impact on the performance than the tip gap. Unfortunately, it is difficult to assess the
real impact of the gaps on the draft tube performance. In future work, it would be interesting to
perform unsteady simulations of the full runner and draft tube using a transient rotor-stator interface in
order to understand the real impact of the gaps on the draft tube performance predictions.

Acknowledgments
The authors would like to thank the participants on the Consortium on Hydraulic Machines for their
support and contribution to this research project: Alstom Power & Transport Canada Inc., Andritz
Hydro Ltd., Hydro-Qubec, Laval University, NRCan, Voith Hydro Inc. Our gratitude goes as well to
the Canadian Natural Sciences and Engineering Research Council who participated to the funding for
this research.

Nomenclature
Guide vane angle [] D Throat diameter of the turbine [m]
Runner blade angle [] H Total head [m]

7
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 032007 doi:10.1088/1755-1315/15/3/032007

Hydraulic efficiency opt Index referring to the point near the best
N11 Unit speed N11=ND/H1/2 efficiency point
P11 Unit power P11=P/D2H3/2 Vref Reference axial velocity Vref=4Qopt/D2
Vz Axial velocity component [m/s]
Q Flow rate [m3/s]
Q11 Unit flow rate Q11=Q/D2H1/2

References
[1] Deschnes C et al. 2010 General overview of the AxialT Project: a partnership for low head
turbine developments 25th IAHR Symp. on Hydraulic Machinery and Systems (Timisoara,
Romania)
[2] Nicolle J, Labb P, Gauthier G and Lussier M 2010 Impact of blade geometry differences
from CFD performance analysis of existing turbines 25th IAHR Symp. on Hydraulic
Machinery and Systems( Timisoara, Romania)
[3] Vu T C, Koller M, Gauthier M and Deschnes C 2011, Flow simulation and efficiency hill
chart prediction for a propeller turbine, International Journal of Fluid Machinery and
Systems 4 (2) 243-254
[4] Gehrer A, Benigni H and Kstenberger M 2004 Unsteady simulation of the flow through a
horizontal-shaft bulb turbine 22st IAHR Symp. on Hydraulic Machinery and Systems
(Stockholm, Sweden)
[5] IEC 60193, 1999, Hydraulic turbines, storage pumps and pump-turbines Model acceptance
tests, Second Edition 1999-11, Published by the international Electro technical Commission.
[6] Nilsson H and Davidson L, 1999, A numerical investigation of tip clearance flow in Kaplan
water turbines, Proc. of hydropower into the next century III (Gmunden, Austria)
[7] Birch D, Lee T, Mokhtarian F and Kafyeke F 2004 Structure and Induced Drag of a Tip
Vortex J. Aircraft 41(5) 1138-45