World Academy of Science, Engineering and Technology 49 2011

Hole Configuration Effect on Turbine Blade

Cooling
A.Hasanpour, M. Farhadi and H.R. Ashorynejad

Abstract—In this paper a numerical technique is used to predict A study comparing total pressure losses between an uncooled
the metal temperature of a gas turbine vane. The Rising combustor endwall, a cooled endwall with cylindrical holes, and a cooled
exit temperatures in gas turbine engines necessitate active cooling for endwall with shaped holes was performed by Barigozzi et al
the downstream turbine section to avoid thermal failure. This study is
[10]. They reported that there was no significant difference
performed the solution of external flow, internal convection, and
conduction within the metal vane. Also the trade-off between the between the secondary flow fields with cylindrical holes
cooling performances in four different hole shapes and configurations versus conical diffuser shaped holes. They showed that for
is performed. At first one of the commonly used cooling hole low injection rates, the magnitude of the total pressure losses
geometry is investigated; cylindrical holes and then two other in the loss core of the passage vortex increased relative to the
configurations are simulated. The average temperature magnitude in uncooled case. Barigozzi et al[10] storied almost the same
mid-plan section of each configuration is obtained and finally the
performance in terms of film-cooling effectiveness for the two
lower temperature value is selected such as best arrangement.
different hole shapes at low mass flow rates, but reported
Keywords—Forced Convection, Gas Turbine Blade, Hole significantly better cooling performance by the shaped holes at
Configuration high mass flow rates. Note that their investigation featured
diverse hole patterns in each passage, which meant that it was
not a direct one-to-one comparison. Colban et al [12]
I. INTRODUCTION compared the film-cooling performance of cylindrical holes

I MPROVING the performance of gas turbines can be achieved

by increasing the turbine inlet temperature. This subject
requires highly effective cooling techniques to maintain the
and fan-shaped holes on a turbine vane endwall. They had the
same experimental setup as the current study, with two
passages having the same layout, one with cylindrical holes
temperature of the components in contact with the hot gases at and the other with fan-shaped holes. The missing element in
acceptable levels. Film-cooling has been widely used to the research that has been performed is the investigation into
protect gas turbine blades from hot gases by injecting the effect of difference holes configuration and shape on the
compressor bleed air through discrete holes in the blade cooling the gas turbine vane and comparing the result in clear
surface [1-3]. Protecting the blade surface by means of film- parameters.
cooling is tested by reducing the adiabatic wall temperature. II. NUMERICAL SIMULATION
Sieverding and Wilputte [7], Friedrichs et al [8] Kost and
Nicklas [9], and Barigozzi et al [10] have reported A.Model Geometry
aerodynamic benefits associated with end wall film-cooling in At first it has been modeled case 1(Fig.1) for validation.
the form of reduction of the size and strength of the passage Then other cases (2-4) are simulated and then compare them
vortex and a reduction in strength of the cross passage flow. with two parameters. One of them is steadying the temperature
However, there is an aerodynamic downside to film-cooling profile on the blade surface and the other is decreasing
injection, as Friedrichs et al [11] reported that film-cooling average temperature on the blade. When these traits are
ultimately increases the total pressure losses through the satisfied the failure possibility of gas turbine vane that
passage. They showed that the additional losses generated
happened because of thermal gradient and thermal stresses is
within the hole and the losses generated by coolant mixing
going to be les and les. The experimental data for validation
with the free stream outweigh the reduction in losses
of the this study for case 1, need to provide clear boundary
associated with a reduction in secondary flows.
conditions and cooling air flow rates of high accuracy. The
C3X transonic-turbine guide vane of Hylton et al [13] was
selected as the test case, for this work provides detailed
measurement of the external and internal convection and the
metal surface temperature. The experimental facility consisted
of a linear cascade of three C3X turbine vanes. The center
vane was cooled by air flowing through ten round flow
A. Hasanpour is MSc student at Faculty of Mechanical Engineering in the passages from the hub to the shroud.
Babol University of Technology, Babol, P. O. Bax 484, I. R. Iran (E-mail:
arman.hasanpour.87 @gmail.com).
M. Farhadi is assistant professor of Mechanical Engineering Department in
the Babol University of Technology, Babol, I. R. Iran. (Corresponding author
to provide phone: +98-111-3234205; E-mail: mfarhadi@nit.ac. ir).
H. R. Ashorynejad is MSc student at Faculty of Mechanical Engineering in
the the Noshirvani University of Technology, Babol, I. R. Iran
(h.r.ashorynezhad@gmail.com)

1089
 World Academy of Science, Engineering and Technology 49 2011

0.8
NuD = Cr (0.022 Pr 0.5 Re D ) (1)
The constant Cr in Eq.1 accounts the thermal entrance
effects for the cooling channels [13]. The hot gas flow path
details and cooling channel geometries and the coolant
conditions are specified in Table 2and Table 3 respectively.
The vane material is ASTM type 310 stainless steel [15],
which has the constant density (ρ) of 7900 kg/m3 and the
specific heat (Cp) of 586.15 J/kg.K. The thermal conductivity
is specified to vary linearly with temperature over the range
(from York [16]) in the simulations, as given by:
Fig. 1 C3X Vane with Cooling Holes k = 0.020176 T + 6.811 [W/m-K] (2)

In their studies the heat transfer measurements were

obtained by measuring the internal and external boundary
conditions of the test piece at thermal equilibrium and solving
the steady-state heat conduction equation for the internal
temperature field of the test piece [13]. Temperature
distributions were directly obtained from the normal
temperature gradient at the surface. The cooling holes are
approximately centered on the curvature line of the vane,
except for the two holes near the leading edge. The cross Fig. 2 Computational Domain Fig. 3 C3X Grid
section of this vane is shown in Fig.1 with the placement and
size of the cooling holes marked by white circles. Detailed
geometric parameters are listed in Table 1.

TABLE I
GEOMETRY PARAMETERS FOR C3X VANE
Throat (mm) 32.9
Vane spacing (mm) 117.7
Suction surface arc (mm) 177.8
Pressure surface arc (mm) 137.2
chord (mm) 144.9 Fig. 4 LE Enlargement Fig. 5 - TE Enlargement
Once this topology was generated, it was a relatively TABLE II
straightforward process to produce a grid. The topology HOT GAS PATH FLOW DETAILS
needed to be generated by hand, which turned out to be very Case Ma Ma Re T0 P0
time consuming for a complex geometry like that of C3X. (LE) (TE) (K) (bar)
Among the features that were desirable for viscous grids is 1 0.17 0.88 1.9∗106 796 3.2

that the grids are clustered along the solid boundaries. Grid
TABLE III
generator, Gambit TM was used with the topology to generate COOLING CHANNEL’S PROPERTY
the final grid. N 1 2 3 4 5 6 7 8 9 10
T K) 387 388 371 376 355 412 367 356 406 420
B. Numerical Methodology φ(mm) 6.3 6.3 6.3 6.3 6.3 6.3 6.3 3.1 3.1 1.98
The present computations employed periodicity conditions
to replicate the multiple vane passages in the experiment, and A super-block numerical grid was used in the present work
therefore only one vane was included in the domain. All other to allow the highest quality in all regions with the fewest
parameters were taken from the experiment set-up data. The numbers of cells (40000 cells). A two-dimensional grid was
inlet is located one chord length upstream of the leading edge, first created in Gambit TM from FLUENT INC. The present
where the turbulence level was measured in the experiments. simulations employed two variants of the popular, two-
The outlet is located one chord length downstream of the equation k-ε model for comparison purposes. The first model
trailing edge. Meshes are created for the hot gas path and the was the “standard” k-ε (S-kε) model originally proposed by
solid vane. The flow was assumed to be fully developed at the Launder and Spalding [17]. The second model was a
hole inlets at the hub of the vane, as there were long tubes realizable k-ε (R- k-ε) turbulence model, which is documented
feeding the channels in the experiment. The coolant exited the by Shih et al. [18]. The R-kε model satisfies the so called
top of the vane to atmospheric pressure. The heat transfer realizability constraints for the Reynolds stresses, specifically
coefficients for the internal flow in the cooling holes were requiring positivity of the Reynolds normal stresses and
obtained from the correlation [13, 14]: satisfaction of Schwarz’s inequality for the shear stresses. This

1090
 World Academy of Science, Engineering and Technology 49 2011

model has been shown by several researchers to reduce the

excessive and non-physical production of turbulent kinetic
energy characteristic of the standard k-ε model in areas of high
irrotational strain. This strain condition occurs at the
stagnation point on the leading edge of a turbine airfoil and
downstream of the leading edge on the suction side. Walters
and Leylek [19] documented the superior performance of the
R-kε model in a simulation of a turbine airfoil cascade. The
two-layer near-wall model of Wolfstein [20] was employed to
resolve the flow all the way to the wall, including the viscous
a- Case 2 b- Case 3
sub-layer. The present simulations were run using the
FLUENTTM 6.3.26 code from FLUENT INC., Inc. In the
Fig. 6 Cases grid
fluid zones, the steady, timed-averaged Navier-Stokes
equations were solved, and pressure-velocity coupling was III. RESULTS AND DISCUSSIONS
achieved with a pressure-correction algorithm. In the solid
A. C3X Model
zone, the Fourier equation for heat diffusion was solved. All
The result of the simulation are presented and analyzed in
equations were discretized with second-order accuracy. At the
this section. Two simulations are discussed with different
fluid-solid interfaces, an energy balance was satisfied at each
turbulence models, the S-k-ε and R-k-ε. At first the result for
iteration, such that the heat flux at the wall on the fluid side
case1 (Figs.1-5) are shown (in Figs.7-12) and validate with
was equal in magnitude and opposite sign to the heat flux on
experimental results [13].Then the temperature profiles for
the solid side. The temperature of the boundary itself was
other configurations and arrangements (case 2, 3) (Fig.6) are
adjusted during each iteration to meet this condition. Thus, all
presented (Fig.15). In Fig. (7) for suction side (-1≤x/s≤0), the
fluid-solid interfaces were fully coupled. A final solution was
pressure falls very rapidly from the stagnation line toward the
confirmed only after grid-independence was established. The
throat, reaching a minimum value of static pressure by x/s = -
hanging node adaption technique was employed to retain the
0.25. A mild adverse pressure gradient follows this minimum,
high quality of the background grid. A converged solution was
and after x/s = -0.5, a soft acceleration arises to the trailing
then obtained with the adapted grid.
edge. On the pressure side (0≤x/s≤1), the pressure stays almost
C.Hole Configurations constant near P0 from the LE to about x/s = 0.5 and then falls
off with further distance toward the TE. The calculations
Most often, the holes on the blade tip are arranged on the
reveal good agreement with the data of Hylton et al.
mid camber line of the tips due to manufacturing reasons.
[13],(Figs,7 and 8). Figure 8 shows the heat transfer
However, the simple camber line film-hole arrangement may
coefficient normalized by href (1135 w/m2.K) along the vane
not yield a good design for blade tip film cooling applications
surface by Realizable k-ε (R-kε) and standard k-ε model (S-
since the film cooling jets do not provide adequate coverage
for the high heat transfer regions on either the plane or kε) and contrast with experimental data. Results generally are
squealer blade tips. For the plane tip case, Kwak and Han [21, small percent higher compared to the experimental data. The
22] observed than the high heat transfer region was located on difference between experimental and computational data near
the pressure side of the blade tip. By this reasons in this paper the leading edge can be because of acceleration after
other configurations and arrangements of cooling holes were stagnation point and FULENT software cannot predict the
simulated. At first, for validation with experimental results behavior of flow at this point, thoroughly.
(Hylton et all [13]), the conventional film cooling holes E xperim ent
(Figs1-5) was modeled and then other well-known hole 1
P eresent W ork

configurations were studied (Fig.6)([3] and [23]). After

0 .9
modeling the different hole shapes a comparison between each
other of them are done. The compression parameters are 0 .8

steadying the temperature profile on the blade surface and

P stat/ Po

0 .7
decreasing average temperature. The first parameter can be
0 .6
judge by temperature couture and the second by the average
temperature. By this comparisons between cases (2 and 3), the 0 .5

selection of best hole arrangements could be done. 0 .4

0 .3

-1 -0.5 0 0.5 1
X /S

Fig.7 Experimental and calculated data of Pstat / P0

1091
 World Academy of Science, Engineering and Technology 49 2011

The stream lines of hot gas flow on blade surface are shown
S - κε
E x p e rim e n t in Figure 11. Figure 12 shows the stagnation point
1 .6 R - κε
enlargement and stagnation point illustrated with bold point.
1 .5
1 .4 It is clear that mainstream near the stagnation point has dual
1 .3 actions. When the main flow across the vane, it had to be
1 .2
1 .1 stopped. A part of this flow goes to the upper than stagnation
1 point and makes the suction side and scantling stream of hot
h/h ref

0 .9
0 .8 gas flow goes lower of this point and creates pressure side.
0 .7
0 .6
0 .5
0 .4
0 .3
0 .2
0 .1
0
-1 -0 .5 0 0 .5 1
x /s

Fig. 8 Experimental and calculated data of h / href

From the results it can be seen that the enough pragmatic
results are obtained by Realizable k-ε (R-kε) model. Therefore Fig. 11 Stream lines on the blade Fig. 12 Stagnation point
the R-kε result is plotted. The contour of Mach number on
vane surface is shown in Figure 9. Due to the shape of the B. Other Hole Configurations
airfoil, the flow sees strong acceleration along the suction side In this section the result of other configurations and
near the leading edge. The maximum Mach number in the arrangements of cooling holes are indicated. Because of
vane passage is about Ma=0.85 and it occurs at a position just comparison the result, Fig.13 shown the temperature profiles
off the SS about 20% of the distance from the LE to TE. The for the case2 and 3 (Fig-6) and the magnitude of average
Mach number remains quite low near the pressure side until temperature on the vane surfaces are listed in table 4.
the aft quarter of the vane, when the flow accelerates to the TE
mach number of about 0.8. Notice also that the maximum TABLE IV
Mach numbers in some of the cooling channels exceed AVERAGE TEMPERATURE ON THE VANE SURFACE
Case 2 3
Ma=0.75, meaning the internal flow is also compressible. Average Static Temperature (K) 658 664

Fig. 13 exposes the temperature profiles of case 2, 3 and

4(Fig-6). It is clear that in the case 2 streaming of blade
temperature are very steady. Perhaps it be seems that in
another case this state had been happened. But by respect to
the average temperature such as other parameter for
comparison is considered (Table 4) it is clear that this
parameter in case 4 is smaller than another configuration. By
this reason case 2 is selected such as best arrangement and
configuration between the hole shapes. It means that with this
hole shapes the vane resistance can be improved in relation to
Fig. 9 Contours of Mach number
thermal collapse and satisfied the blade cooling goal.

Fig. 10 Contours of Static Temperature

a- Case 2 b- Case 3
Fig. 13 Contours of Static Temperature

1092
 World Academy of Science, Engineering and Technology 49 2011

[13] Hylton, Milhec, Turner, Nealy and York.(1983) “Analytical and

Experimental Evaluation of the Heat Transfer Distribution Over the
IV. CONCLUSION Surface of Turbine Vanes”. NASA CR 168015.
A numerical methodology for conjugate heat transfer [14] Jatin Gupta, B.Tech ,(2009,”Applicarion Of Conjugate Heat Transfer
simulations has been applied. The present study was designed Methodology For Computation Of Heat Transfer On A Turbine Blade”,
to high-light the promise of this relatively new tool in gas ), A Thesis Presented in Partial Fulfillment of the Requirements for the
Degree Master of Science in the Graduate School of The Ohio State
turbine heat transfer design, wherein the individual heat University
transfer problems (external/internal convection, conduction in [15] Goldsmith, Waterman and Hirschhorn. (1961) “Handbook of
the metal) are coupled in a single simulation. The benefits of Thermophysical Properties of Solid Materials - Volume II: Alloys”. The
Macmillan Company, New York, USA.
the conjugate approach are inherently better accuracy and [16] York, D. W., and Leylek, J. H.,( 2003),”Three-Dimensional Conjugate
reduced turn-around time as compared to the common practice Heat Transfer Simulation of an Internally-Cooled Gas Turbine Vane”,
of decoupled simulations. The vane was cooled by air flowing ASME Paper No. GT2003-38551,.
[17] Launder and Spalding. (1972) “Lectures in Mathematical Models of
radially through ten smooth-walled channels. The predicted Turbulence”. Academic Press, London, England.
mid-span temperature distribution on the vane external surface [18] Shih, Liou, Shabbir and Zhu. (1995) “A New k-ε Eddy-Viscosity Model
was in reasonable agreement with experimental data when the for High Reynolds
[19] Number Turbulent Flows: Model Development and Validation.
realizable k- ε (RKE) turbulence model was used. The Computers and Fluids”. v. 24, no. 3, pp. 227-238.
standard k-ε model did not perform as well, and showed a [20] Walters and Leylek. (2000) “Impact of Film-Cooling Jets on Turbine
positive offset in predicted temperature as compared to the Aerodynamic Losses”. ASME Journal of Turbo-machinery, v.122, pp.
RKE model and the experimental data. There is room for 537-545.
[21] Wolfstein. (1969) “The Velocity and Temperature Distribution of One-
improvement in the turbulence modeling aspect of this Dimensional Flow With Turbulence Augmentation and Pres-sure
methodology, such as the implementation of a model that can Gradient,” International Journal of Heat and Mass Transfer. v. 12, pp.
accurately predict boundary layer transition. With the 301-318.
[22] Kwak and Han. (2002) “Heat Transfer Coefficient and Film Cooling
methodology validated, the computational results were fully Effectiveness on a Gas
analyzed. The highest metal temperature and largest [23] Turbine Blade Tip”. GT -2002-30194.
temperature gradients were located at the trailing edge of the [24] Kwak and Han. (2002) “Heat Transfer Coefficient and Film Cooling
Effectiveness on the
vane where the metal is very thin. The metal temperature [25] Squealer Tip of A Gas Turbine Blade”. GT-2002-30555.
everywhere was much closer to the temperature in the passage [26] Uwe kruger ,karsten kusterer,Gernot Lang, Hauke Rosch, 2000,Analysis
free stream than the coolant temperature. This was due to the of the influence of cooling steam conditions on the cooling efficiency of
a steam cooled vane using, conjugate calculation technique, Private
fact that the thermal resistance due to internal convection was publication-D-52070 Aachen-Germany, vol .5, pp.87-96.
much greater than the resistance of the external convection or
the conduction within the metal.

REFERENCES
[1] B.Dennis, I.Egorov, G.Dulikravich, S.Yoshimura, Optimisation of a
Large Number Coolant Passages Located Close to the Surface of a
Turbine Blade, ASMEPaperGT2003-38051, 2003.
[2] Zhihong Gao, DigantaP.Narzary, Je-ChinHan,2008, Film cooling on a
gas turbine blade pressure side or suction side with axial shaped holes,
International Journal of Heat and Mass Transfer 51,2139–2152
[3] Grzegorz Nowak, Wlodzimierz Wroblewski, 2009, Cooling system
optimization of turbine guide vane: 567–572, Applied Thermal
Engineering 29
[4] Kercher, D.M., 2003. Film-cooling bibliography: 1940–2002, Private
publication.
[5] Kercher, D.M., 2005. Film-cooling bibliography addendum: 1999–2004,
Private publication
[6] Bogard, D.G., Thole, K.A., 2006. Gas turbine film cooling. AIAA J.
Prop. Pow. 22, 249–270.
[7] Sieverding, C.H., Wilputte, P., 1981. Influence of mach number and
endwall cooling on secondary flows in a straight nozzle cascade ASME
J. Eng. Gas Turb. Pow. 103, 257–264.
[8] Friedrichs, S., Hodson, H.P., Dawes, W.N., 1995. Distribution of
filmcooling effectiveness on a turbine endwall measured using the
ammonia and diazo technique. ASME J. Turbo. 118, 613–621.
[9] Kost, F., Nicklas, M., 2001. Film-cooled turbine endwall in a transonic
flow field: Part I – Aerodynamic measurements, ASME 2001-GT-0145.
[10] Barigozzi, G., Benzoni, G., Franchini, G., Perdichizzi, A., 2005.
Fanshaped hole effects on the aero-thermal performance of a film cooled
endwall, ASME GT2005-68544.
[11] Friedrichs, S., Hodson, H.P., Dawes, W.N., 1997. Aerodynamic aspects
of endwall film-cooling. ASME J. Turbo. 119, 786–793.
[12] Colban, W., Thole, K.A., Haendler, M., 2006a. A comparison of
cylindrical and fan-shaped film-cooling holes on a vane endwall at low
and high freestream turbulence levels, ASME GT2006-90021.

1093