Turbine Blade Cooling by Air Using Different Methods
Turbine Blade Cooling by Air Using Different Methods
Turbine Blade Cooling by Air Using Different Methods
2, 2011
Najlaa A. Hussain
Dept. Mechanical Eng.
Univ. Babylon
ABSTRACT
The determination of temperature distribution within a turbine blade is
important in regards to avoid excessive high metal temperatures and temperature
gradients. The prediction of temperature distribution was achieved through the use of
the finite difference technique, developed to solve the problem with the transformation
and grid generation techniques. Steady-state two-dimensional conduction heat transfer
equation was applied to the internal nodes in mesh to get the temperature distribution,
where at the surface nodes (outer, inner), forced convection heat transfer equation was
applied. The solution of the problem was established using computer program, which
serves to solve the equations by using Gauss-Siedel iterative method. In the present
study, three cases are used for thermal analyses of the blade, which requires the
specification of external and internal boundary conditions. It is also found from the
results obtained that the best cooling method to the turbine blade have been obtained
for the third case when the blade cooled by impingement and film cooling method. It
was found in the film cooled blade that the blade temperature decreases about (170 K)
cooler than the blade cooled without the film cooling method. The results of thermal
analyses obtained were compared with the finite element results. The comparison
shows that the finite difference results give a good agreement with the finite element
results with a small difference was calculated to be (3.5%) as maximum value and it
fall to (0%) in most regions.
.
.
.()
.(Gauss-Siedel)
.
.( 0%) ( 3.5%)
220
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
INTRODUCTION
In order to increase the thermodynamic efficiency of a gas turbine engine, a
high turbine inlet temperature is required, [1]. Therefore, the high temperature passing
on the turbine blades causes thermal stress to the turbine blades materials. To reduce
the effect of thermal stress upon the materials of blade, cooling of blades is required,
[2]. The air cooling methods used in the turbine blade can be divided into the following
categories, [3]:
1. Convection cooling.
2. Impingement cooling.
3. Film cooling.
4. Full-coverage film cooling.
5. Transpiration cooling.
The purpose of the present study is to find the temperature distribution in aircooled turbine blades using transformed finite difference equations using body-fitted
coordinate system. The work in this paper is divided into two steps, first generating a
grid and secondly solving the transformed equations in the new computational
coordinates.
There are numbers of numerical, experimental work and theoretical researches
in this field to get the best means for improving the increase of the turbine blades
operating conditions such as Walker, M.J.B., [4], 1989, introduced a master thesis that
involves the determination of temperature distribution within a turbine blade for three
configurations of cooled blades (circular, elliptical cooling holes, and cooling with
impingement and trailing edge ejection). The determination was by the use of finite
element method and for each type, the analysis was demonstrated with different
specific boundary conditions and heat transfer correlations. The results are acceptable
in terms of their accuracy. This was determined by applying the procedure to a test case
and comparing the results with the reference results. Jaleel J.M., [5], 2001, illustrated
the finite difference method with suitable approximation for curved boundary to
determine the temperature distribution within air-cooled turbine blade utilizing circular
cooling holes and verified the effectiveness of the irregular finite difference method on
curved non-uniform shapes such as turbine blade. Kuder D.S., [6], 2003, submitted a
master thesis that involves the prediction of temperature distribution in gas turbine aircooled blades from hub to tip. It was achieved through the use of control volume
technique. The control volume mesh deals with irregular boundary. Unsteady state two
dimensional heat transfer was used to get the temperature distribution. The results of
thermal analyses obtained in two case studies in which cooling passages were taken as
(15) circular holes, with different diameters and as (3) channel passages. Her results
show that the surface temperature distribution with channel cooling passages is (20%)
lower than the surface temperature distribution with circular cooling passages. Marie et
al, [7], 2003, presented a paper to determine the flow and temperature fields around an
inlet guide vane numerically by CFD method. The governing equations are solved by
3D finite-volume Navier-Stockes method. In particular the outer surface temperature,
heat transfer coefficient distributions and static pressure distributions are also
presented. Ken et al, [8], 2000, investigated experimental tests in a low-speed wind
tunnel cascade to determine the film-cooling effectiveness of the film-cooling hole
geometry on turbine airfoils. The results of full coverage film-cooling and heat transfer
characteristics of rotating serpentine flow passage with and without angled turbulence
221
Arkan K. Al-Taie
Najlaa A. Hussain
promoter had been presented in the full paper. Stefan L.F., [9], 2001, have done an
experimental evaluation of both standard and prototype blade designs under real base
load conditions by using optical pyrometry. Pyrometry is a valuable tool for the quality
assurance, since the temperature distribution of each individual blade is determined
carefully. This paper describes the application of a newly developed high-resolution
pyrometer to the latest prototype, the V84.3A2 60Hz 180 Mw gas turbine. Thus, using
new pyrometer probes in conjunction with a continuous data acquisition mode allows
reliable, highly resolved blade surface temperature measurements, where errors and
potential risks for the gas turbine are minimized at the same time.
It has been shown from above researches reported here, that their mainly study
concentrate on the methods to determine temperature distribution within the blade
material. These methods are: analytical, experimental and numerical methods such as
control volume, finite element and finite difference method. More progress is needed in
present study by using finite difference method with body fitted coordinate system is
made to predict the temperature distribution in turbine blade.
PROBLEM FORMULATION
The problem specification was taken from a well documented reference source
containing the following necessary data:
1- Definition of the blade geometry shown in Figure (1), The blade incorporates an
insert which serves as the jet plenum, which is perforated to form array of jet orifices,
Figure (2), is giving rise to mid-chord impingement arrays which cool the pressure and
suction surfaces and an initial jet orifice row at the blade leading edge. The jet orifices
have a diameter (d), a chord-wise spacing of (x n ) a spanwise spacing of (y n ) and the
insert is separated from the blade wall by a distance (z n ). The cooling air, after passing
through the orifices and impinging on the blade surface is constrained to flow in the
chord wise direction and is discharged at the trailing edge flow from the upstream jets
in the array, therefore impose a crossflow on those located downstream, [4].
2- Specifying the boundary conditions. Along the external profile (hot gas side), the
heat transfer coefficient varies in the manner shown in Figure (3) with the hot gas
temperature taken to be uniform and equal to (963K) around the blade external profile
for all cases, these are described by Walker, [4 ]
Blade wall
Zn
S.S. Array
Insert
xn
Trailing Edge Slot
Leading Edge
Row
d
Chordwise
yn
Spanwise
P.S. Array
mj
Chordwise Direction
Initial Crossflow
Insert
222
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
-60
-50
-40
-30
-20
-10
10
20
30
40
50
60
P.S.
L.E
S.S.
Distance Along Blade Surface X (mm)
Figure (3) The Variation of the External Heat Transfer Coefficient,[4]
GRID GENERATION
Grid generation technique can be roughly classified into three categories:
1. Complex variables methods.
2. Algebraic methods.
3- Differential equation techniques.
Only the third category is suitable for the present case because of the complex
shape. In this technique, elliptic PDEs has been used to generate grids.
This choice can be better understood by considering the solution of a steady
heat conduction problem in two dimensions. The solution of this problem produces
isotherms, which are smooth and non-intersecting.
Thompson, [10], worked extensively on using elliptic PDEs to generate grids.
This procedure is similar to that used by Hoffman, [11], and transforms the physical
plane into the computational plane where the mapping is controlled by a Poissons
equation. This mapping is constructed by specifying the desired grid points (x,y) on the
boundary of the physical domain. The distribution of points on the interior is then
determined by solving the following equations,
xx yy P x, y
xx yy Qx, y
(1)
(2)
223
Arkan K. Al-Taie
Najlaa A. Hussain
( x, y )
( x, y )
(3)
(4)
x J 1 y
y J 1 x
(5)
x J 1 y
y J 1 x
Where J 1 is defined by equation (6)
J 1 1 x y x y
(6)
(7)
ay 2by cy J Py Qy
2
(8)
Where,
a x2 y2
b x x y y
c x2 y2
For P=Q=0, equations (7and 8) become in terms of central finite difference, [12]
x i 1, j 2 x i , j x i 1, j
x i 1, j 1 x i 1, j 1 x i 1, j 1 x i 1, j 1
2bi , j
ai , j
4
2
xi , j 1 2 xi , j xi , j 1
ci , j
0
2
y
y
y
y 2 yi , j yi1, j
y
2bi , j i 1, j 1 i 1, j 1 i 1, j 1 i1, j 1
ai , j i1, j
yi , j 1 2 yi , j yi , j 1
ci , j
0
2
(9)
(10)
Both of equations (9) and (10) can be solved easily by the Gauss-Siedel iterative
method,
To start the solution, an initial distribution of x and y coordinates of the grid
points within the physical domain must be provided by using an algebraic model, and
this coordinates (x, y) are used as initial values to solve both of equation (9) and (10).
The resultant mesh is shown in Figure (4).
224
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
(a)
(b)
From Figure (4) there is clearly a high degree of skewness in some regions of
the domain. This skewness will cause some difficulty and inaccuracy in the
computation of the normal gradients of the temperature at the surface.
So, Poissons equation was selected to achieve orthogonality of the grid lines at
the surface, these are described by Hoffmann, [11].
The source terms can be evaluated in a manner based on normal intersection
between the boundary and grid lines.
They have the form, showed by ref. [13]
Q ,
P , x2 y2
2
x
(11)
2
y
(12)
Substituting in equation (7) and (8) both of this equation can be written as:
a x x 2bx cx x
a y y 2by c y y
(13)
(14)
x x y y
y2
on b
(15)
x x y y
on b
(16)
y2
The parameters and are calculated at each mesh point in the boundary, then its
2
Arkan K. Al-Taie
Najlaa A. Hussain
TRANSFORMATION OF GOVERNING HEAT TRANSFER EQUATIONS:The next step in this work is the transformation of governing heat transfer
equations (conduction, convection). For steady-state, two dimensional, the basic
equation for temperature field can be written as:
2T 2T
0
(17)
x 2 y 2
The two partial derivatives will be transformed to a new computational domain
as follows:
2T
x2 T 2 x x T x2 T xx T xxT
2
(18)
x
2T
y
y2T 2 y y T y2 T yy T yy T
(19)
By substitution both of equations (18) and (19) in equation (17) this equation
becomes,
2
x
y2 T 2 x x y y T x2 y2 T
(20)
xx yy T xx yy T 0
This equation can be applied to the interior nodes to get the temperature
distribution in these nodes and can be solved by Gauss- Siedel iterative method, with
central difference approximation we get :
a1
b1
Ti , j 2 Ti 1, j Ti 1, j
Ti 1, j 1 Ti 1, j 1 Ti 1, j 1
2
c
d
Ti 1, j 1 1 2 Ti , j 1 Ti , j 1 1 Ti 1, j Ti 1, j
2
e1
2c
2a
Ti , j 1 Ti , j 1 12 12
2
(21)
For surface nodes (outer and inner) where convective boundary condition is
applied, the transformation applied to the distances between nodes along the blade in
order to be acceptable for using in determining the temperature distribution in these
nodes.
The distance along the blade as shown in Fig.(6) is :
s dx 2 dy 2
(22)
226
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
h ,T
h exex, T hh
S1
qconv.
i+1, qcond.1
i,j-1
qcond.2 i-1,j
i,j
qcond.3
S2
S1
i+1,
i,j+
qcond.3
i,j
qcond.2
qcond.1
qconv.
S2
S 1
S2
i-1,j
hin , Tc
(b)
(a)
(23)
For external surface nodes see fig. (6-a) the equation (23) becomes
S 1
S
S S 2
T 1 K 1 T 2 K 1
K
2
2
2
S S 2
T 1 hex 1
T Ti , j
(24)
1
a 2Ti , j 1 b 2Ti 1, j c 2Ti 1, j d 2Th
a 2 b2 c 2 d 2
(25)
Similarly, for the internal surface nodes see figure (6 b) the equation (23) becomes,
S 2
K
2
S
T 1 K 2
S S 2
T 2 K 1
S S 2
T 2 hin 1
Ti , j T
(26)
1
a3Ti , j 1 b3Ti 1, j c3Ti 1, j d 3Tc
a 3 b3 c 3 d 3
(27)
227
Arkan K. Al-Taie
Najlaa A. Hussain
equations (21), (25) and (27) with a boundary conditions and also the coordinates of the
outer, inner and interior points are taken from Figure. (5).
CONVERGENCE CRITERIE
The Gauss-Seidel iterative solution continues until specified convergence
criterion is met. For this purpose the total changes in the dependent variables are
evaluated as, these are described by Hoffmann, [11].
Error W
j j max 1
i i max 1
ABS W
i 2
j 2
n 1
i, j
Wi ,nj
(28)
Where W represents quantity of interest (T, x and y) and n represents the iterative level.
The convergence criteria is set as error < error-max where error-max is a specified
input. In the present study, the error-max is taken as 0.001
The number of nodes that the temperature distribution will be predicted on it is
(496) with (I=124 and J=4), and hence, 496 equations must be solved to get the
temperature distribution on the blade. To solve these equations, Gauss-Siedel iterative
method can be used to get the solution for this system of equations with the help of
computer program which serves to give the temperature distribution with 500 iteration.
CASE STUDIES:Analysis No. (1):
Turbine Blade Cooling with Impingement-and Trailing Edge Ejection Based on
the Turbulent Flow Between Two Parallel Plates
The insert is perforated by (120) jet orifices and is separated from the hot blade
surface by a distance of (1mm). Each of the (120) jet orifices are of (1mm) and
arranged into (12) rows of (10) orifices each, i.e., each row consists of (10) orifices in
the span wise direction. The rows are arranged so that (6) rows cool the blade suction
surface and (5) rows cool the pressure surface as shown in figure (2). For this (11) rows
the chord wise spacing (x n ) is equal to (5) jet orifice of diameters (5mm), while the
spanwise jet spacing (y n ) is equal to (8) orifice diameters (8mm). This jet impingement
array will be identified as F (5,8) where (5) refers to the chord wise spacing, and (8) to
the span wise spacing and the letter (F) to the source of the heat transfer correlation
used to describe the array by Florschuetz [14] and [15]. The remaining row is located at
the leading edge. The trailing edge region consists of (0.6mm) wide slot which is
(80mm) high, (in the spanwise direction). The cooling air mass flow rate is taken as
28.710-3 Kg/s, all of which passes through the insert and through the jet orifices to
form the cooling jets but only half of it, 14.3510-3 Kg/s, passes through the trailing
edge slot
For the internal surfaces three distinct regions can be considered;
(A) The Leading Edge Region
Chupp et al, [16], investigated the heat transfer at a simulated internal leading
edge region of a turbine blade cooled by a single row of impinging air jets. The results
of this experimental work consist of two heat transfer correlations. One for the jet
stagnation region and one for the area around the jet stagnation region. These two
correlations will be used here to characterize the heat transfer coefficients in this
region. The mass flow rate through the leading edge row, m i , is taken as 2.39210-3
228
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
Kg/s, which is one twelfth of the total coolant mass flow rate, m c . From Ref. [16] the
spanwise averaged Nusselt Number for the jet stagnation region was correlated from
the experimental results to be;
0.8
0.4
d s d d
0.7 d
Nu stag 0.44 Re j exp 0.85
(29)
d yn d le
yn
Where Re j is the jet Reynolds number, d le is the leading edge diameter and
d s is the distance between the insert and the leading edge blade wall, where
Re j 13739, d 1mm, d Le 3.6mm, d s 2mm, y n 8mm
Hence from equation (29); Nu stag 58, hstag 1825W m 2 K
Around the stagnation region, Chupp et al, [16] established another correlation
to found the averaged Nusselt Number from experimental results to be;
Nu asg
d
0.63 Re
yn
0.7
j
0.5
d
d le
0.6
0.5
1.2
d s d d
exp 1.27
d yn d le
(30)
Z
G
(31)
Nu a A1 Re mj 1 A2 n c Pr 0.333
d G j
The calculated values of heat transfer coefficients are shown in table (1) in
terms of the distance of jet spacing (x) along the blade surface.
Along the blade wall, (the impingement surface), the heat flux can be defined
as:
(32)
q ha Ts Taw
Where( Taw ) is the adiabatic wall temperature and can be defined in a non-dimensional
effectiveness, (E T ) as:
T Tj
ET aw
(33)
Ti T j
Where, T j : is the jet temperature, T i : is the temperature of the initial cross flow, Ti T j
The driving fluid temperature along the array (T a ) can be taken as (T aw ) hence
once (E T ) is known (T a ) can be found. Florschuetz [14] estimates the values of (E T ) for
229
Arkan K. Al-Taie
Najlaa A. Hussain
an F(5,8) array and represented in table (1) together with the calculated values of (T a ),
where (T j ) was taken as (378K) and (T i ) as (477K).
For the pressure surface, the heat transfer coefficients are calculated according
to equation (31) for which only Re j and G c /G j will change and the results are listed in
table (2). The value of (E T ) reflects the influence of the higher temperature of the initial
cross flow from the leading edge and because the ratio of the initial cross flow to the
impingement jet array flow, M, has increased from 0.083 to 0.1, different (E T ) values
exist for the pressure surface. These values, together with the calculated values of (T a )
based on a (T i ) of (477K) and (T j ) of (378K) as before, are shown in table (2).
Table (1) Result of Suction-Surface Heat
Transfer
Coefficients
and
Coolant
Temperatures Obtained From Equations (31)
and (33)
x(mm)
h a (W/m2K)
ET
T a (K)
x(mm)
h a (W/m2K)
ET
T a (K)
05
5 10
10 15
15 20
20 25
25 30
1581
1531
1514
1516
1540
1559
0.258
0.220
0.198
0.166
0.120
0.100
403.5
399.8
397.6
394.5
389.9
387.9
05
5 10
10 15
15 20
20 25
1601
1550
1533
1536
1552
0.278
0.219
0.182
0.135
0.106
405.5
399.7
396
391.4
388.5
230
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
231
Arkan K. Al-Taie
Najlaa A. Hussain
5
6
7
10
11
12
Film Cooling
are ordered with a spacing of (5) jet orifice diameters in the chord wise direction and
(8) in the span wise direction. One half of the spent cooling air from these jets flows
rearwards towards the trailing edge and exits the blade through the trailing slot as in
analysis (1). The remaining spent cooling air exits the blade through a series of film
cooling ejection holes distributed throughout the blade wall. There are in total (468)
film cooling holes of (0.52mm) diameter. These are arranged in (12)rows each
containing (39) holes each. (7) of the rows are spaced along the suction surface while
(5) are spaced along the pressure surface. Within each row the holes are spaced a
distance of (2mm) apart. The film cooling holes consist of two rows with an ejection
angle of 640, rows 1 and 8 and 10 rows with an ejection angle of 450.
Specification of Boundary Conditions
The specification of the boundary conditions can be sub-divided into two main
areas:
1-The internal boundary conditions, which will be dealt with in a similar manner to that
used in analysis (1), as the internal cooling arrangement and geometry remains
unaltered from this analysis.
2-The external boundary conditions, which must take into account the presence of the
ejected cooler film of air over the outer surface of the blade.
Internal Boundary Conditions
The boundary conditions at the leading edge are unaltered from analysis (1),
Likewise the heat transfer coefficients for the suction and pressure surfaces jet arrays
also remain unchanged
However the value of T a , the array driving temperature does change because the
temperature of the initial cross flow, T i is reduced due to the effect of film cooling in
reducing the heat transfer to the blade. The trailing edge slot conditions also change
because the temperature of the cooling air entering the slot, T sin is reduced again
because of the decrease in heat transfer to the blade.
Suction Surface-Coolant Temperatures
The value of T i , from a series of preliminary runs was determined to be (465K)
which is less than the value obtained in analysis (1) due to the protection obtained from
the film cooling which reduces the heat transfer to the blade and consequently to the
cooling air.
This value of T i is then used in equation (33) to obtain the values of (T a ) as
shown in figure(13) where (E T ) is obtained from analysis (1)
Pressure Surface Coolant Temperatures
232
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
Again the initial cross flow temperature, T i is reduced to (465 K) and using the
same values of E T as for analysis (1), T a is determined and the results shown in
figure(13)
The Trailing Edge Slot
The reduction in the heat addition to the cooling air along the jet arrays at the
leading edge due to the film cooling reduces the temperature of the coolant entering the
slot and hence alters the slot boundary conditions. The boundary conditions are based
on the parallel plate flow with the following required parameters;
Tsin 423 K , Tsout 428 K , Tslot 425.5 K
The calculated values of h slot along the slot are presented in table (4)
Table (4) Results of Trailing Edge Slot- Heat Transfer Coefficients Obtained from
Equation (35)
x(mm)
h slot (W/m2K)
1
1673
1.5
1404
5
1282
15
1265
External Boundary Conditions
The ejection of cooling air through the film cooling holes results in a change in
the conditions existing along the outer profile of the blade. This section seeks to specify
the new boundary conditions that now exist as a result of the film cooling used on this
blade.
An analysis of the boundary layer with film cooling results in the specification
of the heat flux to the blade surface from the hot gas as;
q hext Taw Ts
(36)
Where q is the heat flux to the blade.
hext is the heat transfer coefficient without film cooling , i.e.; as given by
Figure (3).
Ts the blade wall surface temperature, Taw is the adiabatic wall temperature.
The adiabatic wall temperature can be non-dimensionalized as a film cooling
effectiveness, E f defined by;
T Taw
Ef g
(37)
Tg Tc 2
Where Tg is the gas temperature.
Tc 2 is the temperature of the coolant leaving the film cooling holes, as shown in
Figure (8).
the effectiveness (E f ) is defined for the 450 and 64o ejection angles as :
0.3901 1 0.46 R
(38)
0.8
(39)
Where,
233
x
Re c 2 c 2
R
1.2 Se
g
Arkan K. Al-Taie
Najlaa A. Hussain
0.25
(40)
The value of T g used in equation (37) varies from row to row because the films
from successive rows reinforce each other, i.e.; the film temperature upstream of a row
becomes the gas temperature of the next downstream row. Rows (1) and (8), which are
the first rows on the suction and pressure surfaces inject the coolant into the external
flow which has a temperature of 963K the second rows, i.e.; (2) and (9), inject their
coolant into the film provided by rows (1) and (8) which is now at a lower temperature.
This reduction in the effective T g continues for the other downstream rows also. The
effective T g used for each row is shown in table (5). This temperature is equal to the
value of T c film existing just upstream of the row. For example row (2) is 7.7 mm
downstream of row (1). For x=7.7 mm, E f64 can be calculated to be 0.2028 from which
T cfilm is 853.5 K, based on a T g of 963 K. This value of T c film is now used as the
effective gas temperature, T g for the calculations for the second row.
Table (5) Results of Effective Gas Temperature
Obtained from Equation (37)
Film Cooling Row
Effective T g (K)
Tc
Tg
1
963
2
853.5
3
752
Tc2
4
669
5
609
6
564
7
529
Fig. (8) Film Cooling Hole
8
963
9
847
10
742
11
663
12
603
RESULTS AND DISCUSSIONS
Prediction of blade temperature distribution was performed by using three
cases, hence three different heat transfer correlation were used.
In the three cases, the blade cooling was assumed to be by using a blade with
impingement cooling and trailing edge ejection. The reason behind this choice is that
this type of cooling has a wide range of use in practice.
Analysis No. (1):
The internal boundary conditions for analysis (1) can be summarized in
Figure (9), showing the values of the heat transfer coefficient and the fluid temperature
associated with each region of the blade.
The temperature contour pattern of Figure (10) shows the maximum temperature of
(662 K) at the leading edge region with falling temperatures along both the pressure
and suction surfaces followed by a rise in temperature towards the trailing edge. Closer
inspection of figure (10) shows that over the pressure surface, from the leading edge,
the temperature falls quite rapidly before leveling out and then increasing continuously
234
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
towards the trailing edge. However along the suction surface the fall in temperature
from the leading edge region is not great, but this reduction extends over a greater
length of the blade than that on the pressure surface. However the minimum
temperature on the suction surface is approximately (20 K) higher than the minimum
temperature on the pressure surface i.e., over the mid-chord regions of the blade, the
suction surface is hotter than the pressure surface. At the trailing edge the situation is
reversed with the pressure surface hotter than the suction surface.
These differences in temperatures between the two surfaces can be accounted for by a
number of factors:
1) The blade structure is hollow which reduces the connectivity between the pressure
and suction surfaces and hence reduces the tendency to smooth out temperature
differences between the two areas.
2) As a result of (1) above, the influence of the variation of external heat transfer
coefficient is much greater. For example, 8mm along the suction surface (8,S.S.), there
is a rapid fall in h ext which results the reduction in temperature along the suction
surface at (8,S.S.). After this point as h ext begins to fall again the temperature levels
also fall. Comparing this to the pressure surface where h ext falls very quickly from the
stagnation point to (5,P.S.), before rising slowly. The temperature variation is seen to
mimic this by falling continuously to (5,P.S.). The mid-chord regions also exhibit the
effect of h ext , with the decreasing temperature along the suction surface from (10,S.S.)
to (35,S.S.) reflecting the decreasing h ext while the rise in temperatures along the
pressure surface from (18,P.S.) to (28,P.S.) mimicing the rising h ext .
3) The cooler pressure surface can also be accounted for by the higher values of the
impingement heat transfer coefficients in this region. The variation of the array heat
transfer coefficients does affect the variation of the resulting temperature distribution
but because these array variations are much smaller than the variation of h ext their
influence on the temperature variations is consequentially less.
4) The trailing edge temperature distribution is more complex than the other blade
regions and reflects both the variations of h ext as well as the varying blade wall
thickness around the trailing edge slot. The general increase in temperature towards the
trailing edge can also be accounted for by the lower value of h slot when compared to the
impingement values, h a .
Analysis No. (2):
The internal boundary conditions for this analysis are presented in Figure
(11),While the external boundary conditions are the same as in analysis (1)
The results of the thermal analysis are presented in Figure (12) shows a pattern
very similar to the analysis (1), (5,8) array with maximum temperature of (688 K) at the
leading edge. The temperature levels fall over the mid-chord regions before rising
towards the trailing edge. For the purpose of comparing the (5,8) and (5,5) arrays, It is
clear that while the pattern remains very similar, the (5,8) case, with less jet orifices
produces lower blade temperatures than the (5,5) case which has more jet orifices. This
apparent contradiction occurs due to the reduction in Re j as the number of orifice
increase leading to a reduction in the levels of the heat transfer coefficient. For the
trailing edge region of the blade, the difference between the blade surface temperatures
is very small reflecting the diminishing influence of changes to the cooling
arrangement of the insert and also reflecting the similarity of the boundary conditions
along the trailing edge slot.
235
Arkan K. Al-Taie
Najlaa A. Hussain
399.8
1531
394.5
1516
397.6
1514
389.9
1540
387.9
1559
T
h
403.5
1581
378
2121
378
1825
405.5
1601
399.7
1550
396
1533
391.4
1536
388.5
1552
Tslot 458
378
2121
236
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
660
650
640
630
Y- axis (mm)
10
620
610
600
590
580
10
15
20
25
30
35
40
X- axis (mm)
570
560
550
399.8
1181
386.2
1241
389.8
1198
383.3
1310
540
381.5
1384
414.6
1205
378
1834
378
1773
421.2
1249
378
1834
399.7
1222
389.4
1239
385.8
1282
382.2
1350
Tslot 457.5
Y- axis (mm)
650
10
640
630
620
610
0
0
10
15
20
25
30
35
40
X- axis (mm)
600
590
580
575
570
Arkan K. Al-Taie
Najlaa A. Hussain
392.5
1516
395.2
1514
388.4
1540
397.1
1531
386.7
1559
T
h
400.4
1581
378
2121
402.2
1601
378
1825
389.7
1536
393.8
1533
397
1550
387.2
1552
Tslot 425.5
378
2121
Y- axis (mm)
10
0
0
10
15
20
25
30
35
40
X- axis (mm)
675
650
625
600
575
550
Analysis (1)
525
Analysis (2)
500
Analysis (3)
Ref. [4]
475
450
425
400
-1.00
-0.75
-0.50
-0.25
0.00
L.E.
P.S.
0.25
0.50
0.75
1.00
S.S.
X/C
Figure (15) Comparison between Blade Surface Temperature For the Three Analyses and Ref.[4]
238
Arkan K.
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
CONCLUSIONS
The main conclusions that can be drawn from this work are summarized in the
following points:
1- Three case studies are used for thermal analyses of the blade, which requires the
specification of external and internal boundary conditions. It is also found from the
results obtained that the best cooling method to the turbine blade have been
obtained for the third case when the blade cooled by impingement and film cooling
method.
2- The analysis procedure used gave a good results in terms of their accuracy when
compared with the results of other researchers.
3- Numerical finite difference was used successfully in obtaining temperature
distribution for complicated geometries as compared to the finite element results.
4- The body fitted coordinates method represents an efficient flexible tool for
treating difficult geometries.
5- An impingement-cooled blade can be dealt with a two-dimensional thermal
analysis through the use of heat transfer correlations, which provide spanwise
averaged values of the heat transfer coefficients.
6- The effect of increasing the number of cooling air jets, which impinge on the hot
blade surfaces, results in an increase in the blade temperature. This is because the
increase in number of jets of high intensity cooling regions is accompanied by a
reduction in the jet Reynolds number.
7- Several different heat transfer correlations can be used to describe the
impingement jet array areas as they all produce similar values of heat transfer
coefficients and hence similar blade temperatures.
8- Film cooling can be analyzed by varying the external boundary conditions of the
blade and results in an appreciable reduction in blade temperatures about (170 K)
cooler than the blade without film cooling.
REFERENCES
[1] El-Wakil, M.M., Power-Plant Technology, McGraw-Hill, Inc., 1985.
[2] Turner A.B., Long C.A., Childs P.R .N., Hills N.J., and Millward J.A., A Review
of Some Current Problems in Gas Turbine Systems, http:\ www.google.com, (Key
Words Gas Turbine), 1998.[Paper]
[3] Jack D. Mattingly, Elements of Gas Turbine Propulsion McGraw-Hill, Inc., 1996.
[4] Walker M.J.B., Turbine Blade Internal Heat Transfer, M.Sc. Thesis, Cranfield
Institute of Technology, 1989.
[5] Jaleel J. M., The Effectiveness of Finite Difference method on Curved Boundary
AirCooled Turbine Blades, Journal of Military College of Engineering, No.25, pp. 1932, 2001.
[6] Kudor D.S., Effect of Cooling Air Passage Shape on Gas Turbine Blade
Temperature Distribution, M.Sc. Thesis Al-Mustansiria University, Iraq, 2003.
[7] Marie L.H.,Eriksson L.E. and Suden B., Convective Heat Transfer on an Inlet
Guide Vane, http:\ www.google.com, (Key Words Turbine Blade Cooling),
2003.[Paper]
239
Arkan K. Al-Taie
Najlaa A. Hussain
[8] Ken I.T.and Takasago S.A., Contribution of Heat Transfer to Turbine Blades and
Vanes for High Temperature Industrial Gas Turbines, http:\ www.google.com, (Key
Words Turbine Blade Cooling), 2000.[Paper]
[9] Stefan L.F., Surface Temperature Mapping of Gas Turbine Blading by Means of
High Resolution Pyrometry, http:\ www.google.com, (Key Words Turbine Blade
Cooling), 2001.[Paper]
[10] Thompson J.F., Grid Generation techniques in Computational Fluid Dynamics,
AIAA Journal, Vol.22, No.11, November, pp. 1505-1522, 1984.
[11] Hoffmann K.A., Computational Fluid Dynamics for Engineers, 1989.
[12] Anderson J.D., Computational Fluid Dynamics, McGraw-Hill, Inc., 1995.
[13] Anderson D.A., John C. Tennehill and Richared H. Pletcher, Computational Fluid
Mechanics and heat Transfer, McGraw-Hill Company, 1984.
[14] Florshuetz L.W. Metzger D.E., Su C.C.,Isoda Y. and Tseng H.H., Jet Array
Impingement Flow Distribution and Heat Transfer Characterstics-Effect of Initial
Crossflow and Nonuniform Array Geometry, NASA CR 3630, 1982.
[15] Florshuetz L.W. Metzger D.E. and Truman C.R., Jet Impingement with
Crossflow-Correlation of Streamwise Resolved Flow and Heat Transfer Distributions,
NASA CR 3373, 1981.
[16] Chupp R.E., Helms H.E., McFadden P.W. and Brown T., Evaluation of Internal
Heat Transfer Coefficients for Impingement- Cooled Turbine Airfoils, Journal of
Aircraft, Vol.6, PP 203-208, 1969.
[17] Rohsenow W.M., Hartnet J.P. and Ganic E.N., Handbook of Heat Transfer
Fundamentals, 2nd ed. McGraw-Hill Book Co., 1987
NOMENCLATURE
Symbol
Description
Unit
A 1 ,A 2
mm
Ef
Gj
kg/m2.s
h ex
h in
W/m2.
K
W/m2.
K
W/m2.
h slot
i,j
Jacobian Transformation
Thermal Conductivity
W/m. K
240
Arkan K.
Nu
P
The Iraqi Journal For Mechanical And Material Engineering, Vol. 11,No. 2, 2011
Nusselt Number
Pr
Prandtle Number
q\\
Heat Flux
W/m2
S,S ,S
mm
Temperature
T aw
Physical Coordinates
mm
xn
Chordwise Spacing
mm
Physical Coordinates
mm
yn
Spanwise Spacing
mm
Zn
mm
Greek Symbol
,
Ejection Angle
Deg.
Abbreviations
P.S.
Pressure Surface
S.S.
Suction Surface
241