THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS

345 E. 47th St, New York, N.Y. 10017 96-GT-181

The Society shall not be responsible for statements or opinions advanced in papers or discussion at meetings of the Society or of its Divisions or
Sections, or printed in its publications. Discussion Is printed only If the paper is published in an ASME Journal. Authorization to photocopy
material for internal or personal use under circumstance not falling within the fair use provisions of the Copyright Act is granted by ASME to
libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Repotting Service provided that the base fee of
$0.30 per page is paid directly to the CCC, 27 Congress Street, Salem MA 01970. Requests for special permission or bulk reproduction should be ad-
dressed to the ASME Tedmical Publishing Department

Copyright C 1996 by ASME All Rights Reserved Printed in U.SA

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

A NOVEL TECHNIQUE FOR THE INTERNAL BLADE COOLING

III 1 11 1 MIMI 11 11 I 111

B. Glezer, H. K.Moon, T. O'Connell
Solar Turbines Incorporated
San Diego, California

ABSTRACT A0 inlet jet area

Development of an adequate air cooling system for the thermally D,„ diameter of leading edge
highly loaded leading edge and tip of the blade, that is cost effective and also D, hydraulic diameter
relatively insensitive to manufacturing tolerances and operating environment heat transfer coefficient
continua tote one of the major challenges in advanced gas turbine design. axial distance between bottom of the vortex chamber and exit
Extensive studies on the convective (including impingement) and film Port
cooling techniques produced remarkable progress in achieving a high thermal conductivity
cooling effectiveness level for turbine airfoils. However, in the case of length of test section
turbine blades, application of these techniques has severe limitations. momentum flux
Highly effective impingement cooling needs to be combined with film dis- Nu, average Nusselt number based on hydraulic diameter of cylindrial
charge of the spent air to avoid a negative impact of crossflow on internal test section, /IDA
heat transfer and also provide additional thermal protection of the surface Nu, avenge Nusselt number based on inlet slot, Jr ilk
downstream of die discharge holes. Noticeable aerodynamic penalties, stress pressure
concentration and significant increase in manufacturing cost limit applica- radial distance
tion of blade film cooling, particularly for moderately high operating radius of jet
temperatures. radius
Search for a highly effective robust design of internal airfoil cooling outer radius
which can delay the use of film cooling resulted in the creation of a new Re, Reynolds number based on hydraulic diameter of cylindrical test
technique which is described in this paper. This technique is based on section
generation of a swirling flow structure in the blade internal leading edge Re, Reynolds number based on inlet slot, V os/v
pnvir Significant heat transfer augmentation can be achieved when the inlet slot width
cooling air is delivered into the leading edge plenum tangentially to the inner x, y, z spanwise, streamwise, and normal coordinates
concave surface. The best results can be obtained when the swirling flow temperature
is allowed to move radially, creating a three-dimensional screw-shaped flow T, initial temperature of test section
in the plenum. T„ plenum air temperature
The presented results of the flow and heat transfer studies performed U,V,W velocity components in the x,y,z directions
for the practical range of Reynolds numbers for the internal flow show that V, mean velocity
the leading edge screw-shaped cooling technique provides internal heat 0 swirl number
transfer rate comparable with impingement coupled with film discharge of shear stress
the spent air, is mom effective than impingement with cross flow and is dynamic viscosity
almost five times higher than heat transfer in the smooth channel. swirl angle
kinematic viscosity
NOMENCLATURE 6 boundary layer momentum thickness
11 cooling effectiveness
A area

Presented at the International Gas Turbine and Aeroengine Congress & Exhibition
Birmingham, UK — June 10-13,1996
Subscripts • Concerns about engine durability associated with potential of the
s static plugging of the small film holes, especially for long-term operation in
t total an industrial environment
w wall
ex exit For aeroagines, all these concerns are easily outweighed by the
ch chamber requirement of very high cooling effectiveness, which justifies a large
in inlet amount of cooling air and the application of blade film cooling. For a

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

moderate TAIT of 1120-1200°C, a maximum effort should be made to
INTRODUCTION avoid blade film cooling until all means of augmentation of internal heat
transfer are exhausted.
Effects of Engine Application on the Blade Cooling Concept A multitude of straight, anted, chevron and "broken" trip-snips (ribs),
Selection pin-fins and other more exotic geometries intended to trip the boundary
Advancements in high efficiency gas turbine engines with the layer in the internal passage, have been applied or considered over the last
exception of recuperative industrial engines are often measured by gradually decade (Han and Duna, 1995). Unfortunately, some of the promising
increasing level of turbine inlet temperature and a correspondingly rising intricate vortex generating geometries which were studied in the up-scaled
optimal compressor pressure ratio. For advanced aeroengines turbine inlet research models cannot always be reproduced in actual size blade castings
temperature above 1650°C and compressor pressure ratios above 40 are or are very sensitive to the manufacturing tolerances, particularly when
becoming a reality. A high specific power, which is a primary objective for small internal holes and sharp edge features are required.
this type engine is vitally obtained together with high efficiency. Jet impingement cooling, which has been proven to be highly effective
However, anerne operating conditions require frequent engine inspections for turbine nozzles, and where spent air is removed toward the trailing edge
and continuous health monitoring. The leading requirement for industrial through the passage between airfoil inner wall and insert, is usually not
engines is bag-term durability without frequent inspections and overhauls. practical for the blades without immediate film discharge of the spent air.
This factor, as well as frequent selection of the upper limit for the Without that, the presence of crossflow results in a significant reduction of
compressor discharge pressure to be well below the available gas pipeline the impingement cooling effectiveness (Kercher and Tabakoff, 1978).
fuel supply pressure, lead to a moderate turbine rotor inlet temperature This brief summary indicates that development of more effective
(TR1T). The TRIT limit for a modem medium size industrial gas turbine internal cooling techniques, which can counterbalance high external heat
is established to be in the range of 1120- 1200°C. High thermal loads loads for the blade leading edge and tip section is required, particularly for
combined with large inertial and dynamic forces result in the turbine blade moderate turbine inlet temperatures, when the application of film cooling
being the most critical component for engine endurance. Highly effective can be avoided. This paper describes a non-traditional blade cooling
blade cooling techniques are required for long term endurance to satisfy the technique based on the application of a screw type vortex dominated flow
bulk metal temperatures of about 820°C required by creep-rupture life in the internal channel of the blade leading edge.
characteristics for advanced blade alloys. For these bulk metal temperatures
at a TRTT of 1200°C and compressor discharge temperatures close to SCREW COOLING CONCEPT DEFiNMON
450°C, the required cooling effectiveness parameter, as defined below has Most of the previous studies of confined vortex flows were performed
to be close to 0.5: with the objective to intensify near wall heat transfer for heat exchangers or
to improve fuel/air mixing characteristics of combustion devices. Most of
these studies were conducted for configurations which included separate
_ Tg - Tm - flow swirling devizes, such as turning vanes in the annular channel, twisted
11 -
Tg - Tc tapes or a central core with spiral passages along the cylindrical channel.
None of these flow-swirling configurations is publicly known to be applied
to turbine blades with the exception of the internal vortex matrix which was
where: first introduced during the 1993 Turbo Expo Congress in Cincinnati by
Mashproekt Gas Turbines (Ukraine). This innovative configuration,
Tg - average blade inlet relative gas temperature however, requires quite complicated precision casting of internal matrix ribs
Tm- average blade bulk metal temperature connecting the pressure and suction sides of the blade and results in an
Tc • blade entry coolant temperature increased weight of the blade.
A few years ago the authors developed a concept based on cooling air
A combination of the highest airfoil external heat transfer rate coupled supplied tangentially to the inner wall of the leading edge channel through
with wake induced turbulence and unfavorable area ratio between inner and a slat or a row of discrete holes parallel to the channel axis. The main intent
outer surfaces make the blade leading edge the most difficult to cool. was to address the challenge of finding a more effective internal cooling
Deviation of the stagnation point during off-design operation complicates technique for the blade leading edge, which is insensitive to manufacturing
the problem even more. tolerances, that is inexpensive and robust for long-term operation. Testing
Cooling techniques based on film protection of the leading edge of the early model, with almost tangential impingement jets, which also
("shower head") provide adequate cooling effectiveness, but produce a induded an air discharge path through the tip of the channel to simulate the
number of negative effects: turn into the blade tip passage (Fig. I), clearly indicated that formation of a
screw-type vortex flow in the channel produces significant heat transfer
• Increased local stress concentration augmentation and provides gnat potential for the blade leading edge
• Additional aerodynamic penalties cooling. Three distinct vortex or screw-cooling configurations which are

2
 •

Figures 3 and 4 show the radial distribution of static pressure,

tangential velocity and axial velocity in both types of devices at a distance
of about two diameters of the chamber from the tangential inlet Similar
experimental data is described by Lewellen (1971). Opwa's experiments
Section A-A were performed for the range of Reynolds numbers from 4100 to 8300,
TIP based on the diameter of the tangential jet. No experimental data were

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

presented for the velocities and radial pranue distribution in the vicinity of
the tangential ix entry, which is the most critical for the subject of this
paper. These data are required to determine the spacing between the
BLADE — consequent discrete tangential jets to provide e'recharge" of the swirling
LEADING (aigular)monienmm, increase of the tangential velocity component, thin
EDGE down the boundary layer, and intensify the heat transfer along the concave
wall.

FIG. 1 - INITIAL SCREW-COOING

MODEL
CONFIGURATION

402* for the blade leading edge and tip cooling are described in the end
of this section. Details of the flow visualization and heat transfer
measurement technique using a thermochromic liquid crystal are described
later in the paper.
Limited previous studies of this complicated three-dimensional flow
phenomain usually related to the so called vortex chamber, were performed
primarily for centrifugal separators. Ogawa (1992) provided a compre-
hensive analytical and experimental sudy of two types of vortex chambers:

• Separator with a single tangential flow jet entering the chamber near
its top and air discharged axially also through the top (Fig. 2a)

• Screw type wirier pert with the same air entry arrangement combined
with air discharged axially through the hole in the bottom (Fig. 2b)

1101111.03314

FIG. 3 - GEOMETRY DEFINITION AND VELOCITY

DISTRIBUTIONS IN THE VORTEX
Er•••••Vo CHAMBER

The flow phenomena in these type of devices are primarily defined by

the relation between angular and axial momentum fluxes (coordinates in
Hg.3). Initial momentum flux of the tangential jet (Equation I) is
MP. 7 transformed into angular momentum flux of the flow in the chamber
(Equation 2) and axial momentum flux (Equation 3).
'ft
(1)
p Ao Vo 2 ro

Mx = 2np f U V R 2 dR (2)
a. Returned Flow Type b. Axial Row Typo es
MINILODI

FIG. 2 - TWO TYPES OF VORTEX CHAMBERS My s. 2np f Vy 2 R dR (3)

3
 5 Axial Flow Type
= 193 rnm
4 — • = S--
P 4
70 My
0 10 20 30 40 50 60 70 3
•

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

>4'20 2
10
1
°O 10 20 30 40 50 60 70
1030
0 0.5 1.0
171-1
Moonlike!
Curves d
OGAWA
Vona

237

(a) Static Pressures (b) Velocities

1110211404M

FIG. 4- DISTRIBUTIONS OF THE STATIC PRESSURE AND

FIG. 5 - DISTRIBUTION OF THE SWIRL PARAMETERS
THE VELOCITIES
2
tifx Rtn
- Ao .
Ogawa suggested using the ratio of total nondimentional angular Ach R.
momentum flux (Equation 4) to non-dimensional axial momentum flux tify Ro R:
(Equation 5) as a characteristic of vortex flow in the chamber, which is
known as a swirl parameter (Equation 6).
A single tangential inlet results in some circumferential nonuniformity of the
2n f angular velocity in the chamber, particularly for a low momentum jet For
Mx - U V R 2 dR
Ao Vo 2 r. (4) a relatively high jet swirling momentum circumferential nonuniformity
disappears at a plane located approximately 114 of the chamber diameter
downstream of the jet. The following relation between swirling angle y and
swirling parameter was recommended: tan y = 1.18 4A76. No data was
7n available for the region immediate downstream of the tangential jet, where
Jaz - Vy 2 R 2 dR (5) the entire flow passes through the boundary layer with the radial velocity
Ac Vo 2 ro 2 Re component equal to zero. Further downstream, where the vortex is fully
developed a radial distribution of tangential velocity component outside of
boundary layer fits a law of constant circulation U R = const. This
He provided an axial distribution of the swirl parameter for both devices observation will be important later when the local heat transfer in the vortex
(Figs. 5a and b), indicating insignificant change of the swirl parameter for chamber is considered.
the screw type swirling pipe along the axis. Schuldn and IChalatov (1982) Another device (Hg. 6), whose configuration resembles the blade
used swirling angle y, defined as tan y =-;/t y, where ; and s ale axial screw-cooling concept the most, was briefly described by Kermbrock and
projections of rotational and axial shear stresses, in combination with Meghreblian, (1961). Very limited data on the flow study were provided.
jet/chamber areas ratio as main characteristics of the swirling flow. For the A number of attempts to solve analytically the quasi-cylindrical continuity,
plane tangential jet entering chamber at 90 to the chamber radius, they momentum and energy equations for compressible turbulent confined
defined an initial swirl parameter (geometry is defined in Hg. 3) as follows. vortex flow (Hall, 1966) which were based on "eddy" approach (Ragsdale

4
 nabulent flows can have a strong influence on both flow and heat transfer
phenomena in the vortex chamber. Tani (1961) demonstrated experi-
mentally that secondary Corder tyre voiliCCS exist along the concave
surfaces even for well-developed turbulent boundary layer.
Established by Shultz-Grunov (1973), the limits of stability for the
boundary layer on concave surface (Hg. 8) are heavily affected by the
thickness of the boundary layer (6) Corder number G = (Vo 6/v) (6/R)" is

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

now a commonly used characteristics of the flow behavior. According to
Ligrani and Niver (1988), the centrifugally induced instabilities accelerate
the flow just outside of the boundary layer and push this flow towards the
concave surface displacing slower-moving fluid near the wall. This is
Schematic Illustration of he Nuclear Rocket Engine expected to result in significant heat transfer augmentation.
Very limited heat transfer data for the vortex chambers are available.
TANGENTIAL
VELOCITY.
13
12 8/R0• 0.02
11

—STATIC
PRESSURE

■
7
Neal cold 0.06
6
FIG. 6 -VORTEX CONTAINMENT FOR THE
GASEOUS FISSION ROCKET 5 0.10
4
3
0 0.1 02 0.3 0.4 0.5 0.6 0.7
a& 2xEirAg
010211.00•61

FIG. B- LIMITOF THE STABILITY FOR THE BOUNDARY

LAYER ON THE CONCAVE WALL SURFACE

Schukin (1970) summarized the results of the previous studies in the

correlation

Nu = .186C, Re' Pr',

where
e hDo Re e VoDo 0.76( Ao r (Aar"
•10211407111 C1
1 Ach Ach
FIG. 7 - SCHEMATIC ILLUSTRATION OF
GORTLER VORTICES IN THE
BOUNDARY LAYER ON THE For a long (lid =39) chamber with a single tangential inlet, area ratio
CONCAVE WALL SURFACE Ao/Ach = 0.5 and Reynolds number in the range of 3 x 10 to 2 x 10', this
publication reammends a formula Nu =0.417 Re'. For a chamber of Lid
1961, Kinney 1966) to reduce turbulent equation to an equivalent laminar =13.3 with two axisymmetric narrow rectangular tangential slot entries of
problem, could not describe satisfactorily the flow phenomena. Additional Ao/Ach = 0.75 and a range of Reynolds number between 2 x 1 ft, and 8 x
complication comes from the de-stabilizing effect of centrifugal forces 10' he recommended a formula N, = .044 Re". The results of this study
acting hi the boundary layer along the concave wall, leading to formation of indicated that beat transfer in the swirling flow can be 15-20 times higher
the standing wave of vortices first observed by Corder (1940). Axes of than in the axial pipe flow.
Gortkr voices (E.g. 7) are parallel to the mainstream flow along the curved A few vortex or saew type swirling arrangements could be considered
wall with the vortical disaubance propagating normally to the principal flow (Fig. 9) for application in the turbine blade leading edge and tip cooling
directions. The presence of Corder vortices in laminar and probably (Glezer, et al, 1993):

5
 (c)

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

View A-A View A.A
cab==ma._
View A-A

FM. 9 SCHEMATIC VIEW OF SCREW-COOLING (CONFIGURATIONS a, b,

• Tangential swirling flow supplied along the channel (Configuration

1) (Fig. 9a)

• Three-dimensional flow structure with a partial tangential swirling

flow supply along the channel (Configuration 2) (Fig. 9b)

• Twodimensional axially uniform flow stricture using tangential inlet

slot with dividers and coupled with discharge (film) holes
(Configuration 3) (Fig. 9c)

TEST APPARATUS AND EXPERIMENTAL PROCEDURE

An initial qualitative study of the flow structure was performed using
the helium filled soap bubble flow visualization technique (Greta, 1992)
T-C CAMERA
with thin light sheets for Illumination in the plane of interest. More detailed
measurements of three-dimensional flow phenomena using hot wire NOY
anemometers and Laser-Doppler velocimetry systems are underway. This WORKSTATION
study, as well as local heat transfer measurements, were performed in un- N021141011
sealed transparent acrylic models simulating the practical geometry and
FIG. 10 SCHEMATIC VIEW OF TEST APPARATUS
Reynolds numba range in the channel. The local heat transfer coefficients
were measured with the thermochromic liquid crystal (TLC) technique and 158 rum deep in the streamwise direction. In order to have a close to
described by Metzger and Larson (1986), and Vendula et al (1988) using uniform static pressure air supply, the plenum was sized to 76.2 x 76.2 mm
a transient heat transfer solution for the semi-infinite plate. abng the section length. A constant length of the flow path. 350 mm, was
Schematic of the test rig is shown in Rg. 10. The test apparatus maintained from the plenum to the test surface. The test was conducted for
consisted of the test section, air-supply system, thermal data acquisition five different air flowrates corresponding to Reynolds number range
system, and image processing system. The surface of the test section was between 3,000 and 20,000.
coated with the TLC. Compressed preheated air supplied to the distributing The test inner surface of a semi-circular section was coated with the
plenum was directed through a narrow channel to the concave test surface TLC (Hallarst, R35C1W). Then the test surface was spray-painted black
simulating the internal passage of a turbine blade leading edge. (Hallcrest, BB-G1) to provide an optimum contrasting visual background
Temperature and pressure data were acquired by a Daytronic and Micro for the liquid crystal color display. The air was controlled with a dome
Vax system, and thermal image data were processed with a Silicon Graphics pressure regulator and the flow rate was monitored by a turbine flow meter.
Indy workstation. Automatic temperature controlled power input supplied the heated air. A
The entire test section was constructed from a 6.3 mm thick pneumatic three-way divater valve was used to bypass the test section until
transparent acrylic to allow viewing from the outside. The cylindrical the heated air reached the desired parameter. The "time zero" light was
chamber (Mg. 11), which simulates the turbine leading edge cooling included in the test set up to synchronize a Micro Vax and Daytronic data
passage, had an inner diameter of 50 mm and was 375 mm long in the axial acquisition system with the thermal imaging system. The image processing
(lengthwise) direction. The rectangular slot channel, which provided a system included a Y-C camera (Cohu 82100) and Silicon Graphics Indy
tranwase jet tangential to the test surface, was 4.0 mm wide, 350 mm long workstation (R4000SC). The acquisition program was implemented to

6
 • 04 19 5.10 • 10.15 a; 15.20 • 2345 02533 • 331.5 • 35-C • 4345 B4s.63 • 5655 • 5.540

45 03 641 3 40 533 Leo 431 3f3 33' 2, 221

yIO

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

(a) Local Nusseft Number Distribution for Re, = 6356
Configuration 1

AIFI
SUPPLY I
*-218 mm
• 04 05.10 • 1015 01533 • 2425 02533 • 70-35 • 35-C • 43-45 04550

HEWN
60.8 mm -

74$ 6.93 641 503 533 454 4.34 353 311 2 PO 2_78 176 1.24 072 0.21

TEST SURFACE
View A-A
—mans
(b) Local Nussett Number Distribution for Re,= 9964
FIG. 11 - SCHEMATIC VIEW OF TEST SECTION Configuration 2

acquisition system with the thermal imaging system. The image processing
system included a Y-C camera (Cohu 82100) and Silicon Graphics Indy
workstation (R4000SC). The acquisition program was implemented to 070-75 t25-32 030 34 073-40 •4043 CM 55

utilize the advanced features of the lady workstation which digitally records
57
the image of the surface pixels as they experience a color transition during 3:

the test.
Cahbration of the TLC was performed using the spectral peaks of light 37

intensity for each of the three observed colon. A diffuse-light source was 7.0
mounted on the camera to eliminate the view-angle dependency of the TLC
color transition discussed by Herold and Wiegel (1980). TLC was (c) Local %melt Number Distribution for Re, = 8286
calibrated with a coupon at the representative view angle at the beginning Configuration 3
of each test. The experimental uncertainty for the present measurement of
heat transfer coefficient was within ±10%, using methods of Kline and FIG. 12- EXAMPLES OF HEAT TRANSFER DATA SETS
McClintock (1953) for single sample experiments. Repeatability of the data
was considerably better than ±10%. The supply air temperature was
70
optimized for a given air flow rate to eve a transient test duration well under
40 seconds to satisfy a semi-infinite thermal transient solution. The
(0
optimum supply air temperature was analytically predicted and verified
experimentally with four thermocouples.
ea
TEST RESULTS
Data presented in this section specify Nu and Re numbers basal on the 40
tangential slot width, S. accept for Hg. 16, where these data were readjusted
Pt
for the leading edge diameter to be compared with other cooling techniques.
Figure 12 presents examples of the local Nusselt number distribution
for the 120° circumference of the concave surface (±60°) from the
simulated leading edge stagnation point, obtained with the '11..0 technique. 20

Configuration I (Hs. 9a) with air supplied to the leading edge through
the full length slot, generating a screw-shaped column of vortices and then 10
exiting through the tip, presents a baseline of the screw-cooling geometry.
Figure 13 presents heat transfer measurement results for this configuration 0
with the Nussek numbers averaged circumferentially and plotted against the 0 1 2 3 4 5 7 8
normalized lengthwise distance from the bottom of the test section. Y/1)
4144n4.444
Although the Nusselt number generally decreases with the distance FIG. 13 - CIRCUMFERENTIALLY AVERAGED
from the injection slot for all five considered cases it changes more NUSSELT NUMBER DISTRIBU-
noticeably in the axial direction as shown in Fig. 13. It was determined that TION, CONFIGURATION 1

7
 the Nussek number VASS less than 7% circumferentially from its mid-point cooling method quite attractive from the manufacturing producibility point
This suggests the existence of a dominant vortex column. From Figure 13 of view.
it is clear that the flow in the test section can be divided into three major Figure 14 shows the redistributed, circumferentially avenged heat
regions: Ocy/D43, 4_5<y/D<6.8, and 6.8<y/Ek7.4. The first region =der oxfficients, where the peak heat transfer occurs at a half diameter
corresponds to the area where both tangential and the axial velocities (D) upstream of the blocked portion of the rectangular slot channel.
steadily increase. A highly swirling flow may have resided in the second Therefore, it was demonstrated that the local heat transfer can be adjusted
region, resuking in high heat transfer (Nu), and its plateau may indicate that to match the desired external heat transfer distribution. It was also noticed

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

the flow has already started making a turn toward the tip passage. There is that the heat transfer effect at the tip region, where the flow makes a 90°
no tangential air injection in this region. The dip (minimum point) at the turn was not as pronounced as before.
top, ego:chilly noticeable at high Reynolds numbers, may have been caused The case in which all the air supplied to the leading edge exited
by secondary vortex formed by a flow separation at the corner Although through the film cooling holes was investigated with Configuration 3 (Fig.
it may have been deflected somewhat, the screw-shaped flow very Rely 9c). hi order to realistically simulate the turbine airfoil leading edge, the test
impinges on the top of the channel. section was modeled in a semicircular shape. The air supplied tangentially
To evaluate the possibility of controlling the local heat transfer along to the leading edge through the discrete rectangular holes exits through the
the cooling channel, the test seakin was modified with the top section of the film cooling holes. The total area of the film cooling holes matched the
skit blocked off to intensify local angular momentum flux upstream of this rectangular slot area. In order to achieve a uniform pressure in the plenum,
section (Configuration 2) (Fig. 9b). Following the earlier described it was elongated by 75 mm compared to Configuration 2 at the entrance.
ironed= the cimumfaentially avenged Nusselt numbers are plotted in Fig. The test was conducted for five different flowrates (Reynolds numbers) and
14. The Nusselt numbers averaged over the test data field are plotted the results were reduced to Nusselt numbers based on slot width. The
against the Reynolds number (data set b) in Fig. 16. Nusselt number distribution was also averaged circumferentially and plotted
against the normalized lengthwise distance from the bottom of the test
section in Fig. 15. Also, the Nusselt number was averaged over the test data
00
field, -4.57s xiss4.57 and 0<y/13„<6.98, and plotted against the Reynolds
Rsr°560 number in Fig. 16.
70 "41/4'8864

•
•
Les -: 1 684
Res
R4s =9 45
60
40 ege.
0 9
50 e: 441; 35 -
gio • % •nie x/ % gems
a etio je j1/40,9 %
Air .terses4fra% •
e 4•441ar cebtr; `s%
4°
.rA re : 30 -
t
• %A •
25
30
itrar 1 20 r ti• tertr% N.4%;
20 vis "NS
15
re%
10 trirrorsaH 0 R04 =4418
bt R.544158
Nei 10
1
A Res = 5460
0 0 Res a 6605
0 1 2 3 4 5 A 7 8
5 0 Res = 8286
WD imanowm4
FIG. 14• CIRCUMFERENTIALLY AVERAGED NUSSELT
3 7
NUMBER DISTRIBUTION, CONFIGURATION 2
The plots of the Nusselt numbers show a pattern of secondary vortex
0
WD 4 5 6
01121101916
effects downstream fiat the tangential injection slot. The reattachment of •
FIG. 15 • CIRCUMFERENTIALLY AVERAGED NUSSELT
the deflected tangential jet is apparent at high Reynolds numbers, resulting _
NUMBER DISTRIBUTION, CONFIGURATION 3
in a secondary apex of contours. However, the circumferential vortex
generated by the tangential jet is still the dominant source of heat transfer The plots of Nusselt number (Hg. 12c) distribution show the effect of
enhancement as seen in Fig. 14, such that the heat transfer varies over the dividers in the tangential slot. The circumferentially averaged Nusselt
axial (along the channel) direction rather than the circumferential direction. number shows a sinusoidal peaks corresponding to the location of the slots.
This study revealed an important characteristic of the screw-cooling The overall parabolic (symmetric) distribution seen in Hg. 15 is due to the
technique being not very sensitive to some deviation in the local width of end wall effect.
the tangential slot due to the dominating effect of the swirl, generated A comparison of the three tested configurations of screw-cooling
upstream of the particular location. This observation makes the screw- against impingement cooling both with and without crossflow and also

8
 1000 6. The relatively uniform heat transfer distribution of screw cooling along
900 the loading edge contour makes this concept more advantageous ova
900 impingement cooling which requires not only an (1011100t alignment
700
of the jets with the stagnation point but also provides inadequate
GOO
cooling to a blunt-nosed leading edge.
600
400 REFERENCES

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

tat Arsenievl-V., Mitriaev, LB., "Heat Transfer of the Tangential to the
Concave surface Jet in Crossflow", Teploenergetica, 1987, #11, pgs. 56-60.
Chupp, RE., Helms, H.E., McFadden, P.W., and Brown, T.R., 1968,
200 "Evaluation of
Internal Heat Transfer Coefficients for Impingement Cooled Turbine
Airfoils," AMA Papa No. 68-564.
M.V. and Brown, A., 1994, 'The Sotamwise Development of
100 Cartier Vortices in a Favorable Pressure Gradient", ASME Paper 94-GT-
10000 100,000 166.
Rap
1110114161 Glezer, B., 1992, "Some Aspects of Industrial Turbine Cooling
Development," International Symposium Heat Transfer in Turbornachinery,
FIG. 16 - COMPARISON OF DIFFERENT COOLING Athens. Greece.
TECHNIQUES Gleam B., Lin, T.. Mom, H.K, 1993, "An Improved Turbine Cooling
System". (patent pending).
against smcah and nirsnip augmented channel surface is presented in Fig. Gutter, H., 1940, "Uber Eine Dreidimensionale Instabilitat Laminmer
16. For impingement cooling the NusseIt number is calculated from the Grennhichten an Konkaven Wanda', "Ges d. Wiss. Gottingen, Nachr. Ad.
modified currelation of Chupp, et al. (1968). The heat transfer in a channel Math_ 2(1); translated as "On the Three-dimensional Instability of Laminar
augmented with namel trip strips around the entire circumference was Boundary Layers on Concave Walls," NACA TM 1375 (1954).
obtained from the correlation of Webb et al. (1971). Han, J.C. and Duna, S., 1995 Internal Convection Heat Transfer and
The aomperison shows that saew cooling techniques provides internal Cooling-An Experimental Approach". VKI Lecture Series 1995-05, Parts
channel heat transfer augmentation comparable with the impingement 1, a 111, IV.
coupled with film discharge of the spent air and is more effective than Herold, W. And Wiegel, D., 1980, "Problems of the Photographic
impingement with the cross-flow or trip strips. Documentation of Liquid Crystalline Thermographs, Advances in Liquid
Crystal Research and Applications, L. Bata, ed., Pergamon Press, Oxford,
CONCLUSIONS pgs. 1255-1259.
ICarter, D.M., and Tabakoff, W., 1970. "Heat Transfer by a Square
I. It was &migrated that significant heat transfer augmentation in the Array of Round Air Jets Impinging Perpendicular to a Flat Surface Including
inner passage of the leading edge of the blade can be achieved when the Effects of Spent Air," Journal of Engineering for Power, Trans. ASME,
air is delivered tangentially to the concave wall of the leading edge, Series A, Vol. 92, No.1, January 1970, pp. 73-82.
generating a three-dimensional vortex flow. 1Caretrock, J.L and Meghreblian, R.V., 1961, "Vortex Containment
for the Gaseous Fission Rocket", J. Aerospace Sci.. V.28, p. 710.
2. Investigation of three different leading edge internal cooling Kline, SJ. and McClintock. F.A., 1953, "Describing Uncertainties in
configurations based on vortex flow for the realistic range of Reynolds Single Sample Experiments", Mechanical Engineering, V.75, pp. 3-8.
numbas resulted in selection of a screw-cooling configuration, which Ligrani, P.M., and !giver, R.D., 1988, "Flow Visualization of Dem
demonstrated superior heat transfer rate in comparison with a smooth Vortices in a Curled Channel with 40 to 1 Aspect Ratio". J. Physics of
channel flow, channel with trip strips and even with impingement Fluids, V.31, No.12, pp. 3605-3612.
cooling. Meager, DE. and Larson, D.E., 1986, "Use of Melting Point Surface
Coatings for Local Convection Heat Transfer Measurements in Rectangular
3. Optimizanco of the screw-cooling configuration resulted in an average Channels with 90 Degree Turns," J. Of Heat Transfer, Trans. ASME, 108.
heat transfer rate comparable to impingement cooling over an arc of pgs. 48-54.
t60° from the stagnation point and also demonstrated the ability to Ogawa, A., 1993, Vortex Flow, CRC Press.
adjust local cooling effectiveness along the blade span. Lewellen. WS., 1971,"A Reveiw of Confined Vortex Rows," NASA
CR-1772, Hall P., 1986.
4. The manufacturing advantages of the screw-cooling technique, which Shultz-Grunow, F., and Hein, H., 1956, "Beitrag Zur Comae-
is less sensitive to fabrication tolerances than other highly effective Strutting, "ZFW 4, pgs. 28-30.
internal cooling techniques, make this concept specially attractive for Schukin, VIC, 1970, "Heat Transfer and Hydrodynamics of
industrial applications. Contained Rows in the Reid of Inertial Forces," Machinostroenie, Moscow.
Schukin, V.K. and IChalatov, A.A., 1982, "Heat Transfer. Mass
5. The screw-cooling technique reveals definite advantages over Trutt and Hydrodynamics of Swirling Flows in Axisymmeric Channels,"
impingement cooling for designs without film cooling discharge where 1982, Machinostroenie, Moscow.
cross flow along the leading edge channel is expected.

9
 Tani, I., 1962, -Production of Longitudinal Vorticies in the Boundary
Layer Along a Concave Wall," Of Geophys. Res. V.67 No.8, pg. 3075.
).

Vendula, RJ., Metzger, DE and Beard, W.B., 1988, "Effects o,

Anisotopic Conduction on the Detemnnation of Local Convection heat
Transfer Characteristics with Transient Test and Surface Coatings,"
Collected Papers in Heat Transfer 1988, ASME, Heat Transfer Div., Vol.
104, pgs. 21-28.

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1996/78750/V004T09A015/4215850/v004t09a015-96-gt-181.pdf by guest on 19 March 2024

Webb, R.L,, Eckert, E.R.G., and Goldstein, R.J., 1971, "Heat
Transfer and Friction in Tubes with Repeated-Rib Roughness,"). of Heat
and Mass Transfer, V.I4, pgs. 601-617, Pergamon Press.

10