See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/305864502

Improvements of film cooling effectiveness by using Barchan dune shaped

ramps

Article  in  International Journal of Heat and Mass Transfer · August 2016

DOI: 10.1016/j.ijheatmasstransfer.2016.07.066

CITATIONS READS

12 422

2 authors:

Wenwu Zhou Hui Hu

Shanghai Jiao Tong University Iowa State University
39 PUBLICATIONS   243 CITATIONS    233 PUBLICATIONS   2,109 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Plasma Icing Control View project

Gas turbine cooling View project

All content following this page was uploaded by Wenwu Zhou on 07 October 2017.

The user has requested enhancement of the downloaded file.

 International Journal of Heat and Mass Transfer 103 (2016) 443–456

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Improvements of film cooling effectiveness by using Barchan dune

shaped ramps
Wenwu Zhou, Hui Hu ⇑
Department of Aerospace Engineering, Iowa State University, 2271 Howe Hall, Room 1200, Ames, IA 50011, USA

a r t i c l e i n f o a b s t r a c t

Article history: In the present study, an experimental investigation was conducted to explore a novel film cooling design
Received 12 January 2016 with Barchan-dune-shaped-ramp (BDSR) concept to augment the effectiveness of film cooling injection
Received in revised form 17 July 2016 from discrete holes. Inspired by the unique shape of Barchan dunes commonly seen in deserts to prevent
Accepted 18 July 2016
sand particles on the ground from being blown away by the oncoming airflow, the BDSR concept was pro-
posed to enable the coolant streams exhausted from coolant injection holes to stay more firmly on the
surface of interest for improved film cooling performance. During the experiments, while coolant streams
Keywords:
were injected from discrete circular holes on a flat test plate at an injection angle of 35°, a row of BDSRs
Film cooling
Barchan-dune-shaped ramps
were mounted on the test plate at either downstream or upstream of the coolant injection holes for
Cooling effectiveness quantification improved film cooling effectiveness. While a high-resolution Particle Image Velocimetry (PIV) system
PIV measurements was used to conduct detailed flow field measurements to quantify the dynamic mixing process between
PSP technique the coolant jet streams and the mainstream flows over the test plate, a Pressure Sensitive Paint (PSP)
Mass transfer analog technique was used to map the corresponding adiabatic film cooling effectiveness on the surface of inter-
est based on a mass-flux analog to traditional temperature-based cooling effectiveness measurements.
The cooling effectiveness data of the BDSR design were compared quantitatively against those of a con-
ventional film cooling design without BDSR (i.e., baseline case) under the same test conditions in order to
evaluate the effects of the BDSRs on the film cooling effectiveness over the surface of interest. The effects
of the locations and height of the BDSRs in relation to the coolant injection holes on the film cooling effec-
tiveness at different blowing ratios (i.e., the coolant-to-mainstream mass flux ratios) were examined
quantitatively based on the PIV and PSP measurements. The detailed flow field measurements were cor-
related with the measured film cooling effectiveness distributions to elucidate the underlying physics in
order to explore/optimize design paradigms for better film cooling protection of turbine blades from
harsh environments.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction continuously seeking to improve film cooling efficiency without

increasing the consumption of coolant flow.
Because of wide-spreading application of gas turbine engines in The film cooling effectiveness is found to be affected by a series
power and transportation industries, there is a great incentive to of parameters [1,2], including mass flux ratio (i.e., blowing ratio, M),
maximize the performance of turbine engines for both economic momentum flux ratio (I), coolant to mainstream density ratio (DR),
and environmental reasons. Efficiency improvements of gas tur- turbulence intensity levels of the oncoming mainstream flows, and
bine engines, even if only by marginal increments, will lead to sig- geometric shapes of the coolant injection holes. Film cooling, fun-
nificant value of savings. Thermodynamic analysis reveals that the damentally, is a jet-in-cross-flow (JICF) problem. As described in
thermal efficiency and power output of a gas turbine can be greatly Fric & Roshko [3], the dominated flow features of a JICF include
enhanced through a higher turbine inlet temperature. Therefore, shear-layer-related vortices formed at the leading edge of the jet,
means to protect the hot components of gas turbines from corro- horseshoe vortices around the jet, counter-rotating-vortex (CRV)
sion and melting are very essential, especially the implementation pair in the jet, and wake vortices downstream of the jet. The per-
of film cooling technique. While coolant bleeds represent energy formance of a film cooling design is closely related to how firmly
loss which should be minimized, gas turbine designers are the coolant stream can stay attached to the surface of interest after
exhausted from the coolant injection hole. A major problem of film
⇑ Corresponding author. cooling is that the coolant stream tends to detach from the surface
E-mail address: huhui@iastate.edu (H. Hu). of interest due to the vortex induction from the CRV structures in

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.07.066
0017-9310/Ó 2016 Elsevier Ltd. All rights reserved.
444 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

the jet flow, which leads to deteriorated film cooling performance. University. While coolant streams was injected from discrete circu-
Extensive experimental and numerical studies have been con- lar holes on a flat plate at an injection angle of 35°, a row of BDSRs
ducted in recent years to explore new film cooling designs for were mounted on the test plate at either downstream or upstream
improved film cooling effectiveness. For example, Goldstein et al. of the coolant injection holes. During the experiments, a high-
[4] reported a great enhancement of film cooling effectiveness by resolution stereoscopic Particle Image Velocimetry (PIV) system
using fan-shaped holes. Their flow visualization results revealed was used to conduct detailed flow field measurements to quantify
that the coolant jet exhausted from a fan-shaped hole would stay the dynamic mixing process between the coolant jet streams and
close to the wall surface, unlike the coolant stream injected from the mainstream flows for the test cases with BDSRs mounted onto
a conventional circular hole which would detach from the surface the test plate, in comparison to a baseline case (i.e. without the
of interest at high blowing ratios. Zaman et al. [5] experimentally BDSRs mounted on the test plate). A Pressure Sensitive Paint
studied an inclined jet flow interacting with a vortex generator, (PSP) technique was used to map the corresponding adiabatic film
and showed that the vortex generator would produce a vorticity cooling effectiveness on the surface of interest, based on a mass-
cancellation effect that can postpone the jet lifting off from the flux analog to traditional temperature-based cooling effectiveness
wall surface, thereby, better film cooling performance. By conduct- measurements. The effects of the location and height of BDSRs
ing a large eddy simulation (LES) study on a film cooling flow with on the film cooling effectiveness over the surface of interest at dif-
a micro-ramp vortex generator, Shinn et al. [6] found that the ferent blowing ratios were examined quantitatively based on the
strength of counter-rotating vortex (CRV) structures in the coolant PIV and PSP measurements. The detailed flow field measurements
jet flows would be reduced substantially due to the existence of the were correlated with the measured film cooling effectiveness
vortex generator, resulting in a better film cooling effectiveness distributions to elucidate the underlying physics in order to
over the surface of interest. More recently, An et al. [7] conducted explore/optimize design paradigms for better film cooling protec-
an experimental study to investigate the film cooling performance tion of turbine blades from harsh environments.
behind a crescent block, and showed that the film cooling effec- It should be noted that recent advances in additive manufacture
tiveness would become better within a certain range of blowing (i.e., 3D-printing) technology have greatly broadened the horizon
ratios due to the existence of the block. Na & Shih [8] performed of film cooling designs. Many innovative-but-traditionally-unma
a numerical study on a new film cooling configuration with a ramp chinable film cooling designs (e.g., the innovative film cooling
placed upstream of a coolant injection hole, and found that the design with BDSR to be explored in the present study) can now
backward-facing ramp can greatly enhance the film cooling effec- be manufactured easily as an integrated part at engine scale, which
tiveness over the surface of interest. Sakai et al. [9] investigated greatly promotes the development of film cooling technology.
the combined effects of internal ribs and downstream bumps There are also several advantages for the proposed BDSR concept,
(semicircular, hemispherical and cylindrical bumps) on film cool- in comparison with conventional film cooling designs. For exam-
ing effectiveness. Their measurements showed that the film cool- ple, the BDSR design can provide an alternative choice to increase
ing effectiveness of the studied configurations would be strongly the injection angle of cooling streams and make coolant streams to
affected by both the rib orientations and bump shapes. While most stay attached to the surfaces of interest more firmly. It would also
of the ramp/bump designs suggested in those previous studies are offer a flexible option to place BDSRs at either upstream or down-
shown to be effective in improving film cooling performance, such stream of coolant injection holes. While most of turbine blades of
ramp/bump designs with sharp edges could cause high skin fric- modern gas turbines are coated with thermal barrier coatings
tion, thereby, increased aerodynamic losses. More comprehensive (TBCs), it is possible to utilize TBCs to make tiny BDSRs on the sur-
reviews about different film cooling designs can be found in the faces of blades through an additive manufacture method to
articles of Goldstein [10] and Bunker [11]. improve film cooling effectiveness for an even better protection
Barchan dunes, which are commonly seen in deserts, are fea- to the turbine blades from overheating.
tured by a crescentic platform and a streamlined shape along the
wind flow direction in both of the windward side (convex face)
and leeward side (concave face) along with two gradually- 2. Experimental setup and test models
widened horns pointing downstream [12,13]. When an airflow
passes over a Barchan dune, a pair of symmetric vortices are found 2.1. Experimental models and test rig
to form in the wake of the dune [14,15]. Induced by the arched
shape of the dune, the formed vortex pair can generate a continual The experimental study was conducted in a low-speed,
suction along the axis of the horns, creating a ‘‘quiet” wake flow open-circuit wind tunnel located at the Department of Aerospace
behind the dune. As a result, sand particles can settle down on Engineering of Iowa State University. The wind tunnel has
the ground behind the dune, instead of being blown away from an optically-transparent test section with a cross section of
the ground by the oncoming airflow. It should also be noted that 200 mm
125 mm and is driven by an upstream blower. With
the streamlined shape of Barchan dunes has also been found to honeycombs and screen structures installed ahead of contraction
be able to minimize the friction losses of the airflow as passing section, the wind tunnel can supply uniform low-turbulence
over the dunes. Inspired by the unique feature of Barchan dunes oncoming flow into the test section. The turbulence intensity of
to prevent sand particles on the ground from being blown away mainstream airflow in the test section was found to be about
by the oncoming boundary layer airflow, a new film cooling design 1.5%, as measured by using a hotwire anemometer.
with Barchan-dune-shaped-ramp (BDSR) concept is proposed here In the present study, all test models including test plate and
to enable the coolant streams exhausted from coolant injection BDSRs were made of a hard plastic material and manufactured
holes stay more firmly over the surface of interest for improved by using a rapid prototyping machine (i.e., 3-D printing) that built
film cooling performance. the models layer-by-layer with a resolution of about 25 microns.
In the present study, an experimental investigation was con- Fig. 1 shows the film cooling design with BDSR concept. A row of
ducted to explore a novel film cooling design with Barchan- 5 discrete circular holes with a same diameter of D = 4.0 mm were
dune-shaped-ramp (BDSR) concept for improved film cooling per- designed to inject coolant streams onto a flat plate at an injection
formance by taking advantage of the unique flow structures angle of 35°. While the entry length of the coolant injection holes is
induced by Barchan dunes. The experimental study was conducted 4D, the spanwise pitch between the adjacent holes was designed
in a low-speed, open-circuit wind tunnel available at Iowa State to be 3D. The axial centerlines of the coolant injection holes
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 445

Fig. 2, the thickness of the turbulent boundary layer was estimated,

which was found to be d99  1:4D, and corresponding momentum
thickness h  0:15D (corresponding Reynolds number of
Reh  960), and shape factor H  1:32.
It should be noted that, the size and shape of the Barchan dunes
existing in nature are found to vary in a wide range, due to the
environmental diversity (i.e., wind strength, moisture, sand supply
and so on). The geometry of the BDSR models used in the present
study was generated from a stereoscopic topographic map of an
idealized Barchan dune reported in the previous studies
[12,14,16,17]. As shown schematically in Fig. 1, five BDSRs were
affixed onto the upper surface of the test plate at either down-
stream or upstream locations of the coolant injection holes by
using a layer of 80 lm-thick double sided tape. During the exper-
iments, the distance between the trailing edge of slip face and cen-
ter of coolant hole, L, were adjusted for different test cases. The
effects of the height of the BDSRs on the film cooling effectiveness
over the test plate was also investigated by using two types of
BDSRs, which shared similar geometry, but different height (i.e.,
H = 0.3D and 0.5D, respectively).

Fig. 1. The schematic of BDSR assisted film cooling design (D is the diameter of 2.2. Adiabatic film cooling effectiveness measurement by using PSP
coolant hole). technique

Adiabatic film cooling effectiveness g is traditionally expressed

intersect the upper surface of the test model at a distance of as,
Lleading = 112.0 mm from the leading edge of the test plate. Thus, T 1  T aw
the boundary layer develops for a length of 22D before the g¼ ð1Þ
T1  Tc
mainstream flow encounters with the coolant injection holes.
During the experiments, the test plate was flush-mounted to the where T 1 is the temperature of the main-stream, T aw is the adia-
bottom wall of the wind tunnel test section. batic wall temperature of the surface under inspection, and T c is
A stripe of 1.0 inch wide and 32 grit sand paper was employed the hole exit temperature of the coolant stream.
at the leading edge of test section to trip the oncoming boundary In the present study, instead of conducting temperature mea-
layer to ensure a fully developed turbulent boundary layer flow surements on the surface of interest (such as using thermocouples
over the flat plate. During the experiments, the mainstream flow [18], liquid crystal thermometry [19], infrared thermography [20]
velocity at the inlet of the test section was fixed at U1 = 25 m/s, or Temperature Sensitive Paint (TSP) [21]), the film cooling effec-
and the corresponding Reynolds number is Re = 1.8
105, based tiveness measurements were performed at an isothermal condition
on the distance between the leading edge of the test plate and by using a Pressure Sensitive Painting (PSP) technique [22] based
the coolant injection holes. Fig. 2 shows the measured boundary on mass transfer analogy. As described in Johnson et al. [23], PSP
layer profiles at the immediate upstream of the coolant injection based cooling effectiveness measurements can easily eliminate
holes for the baseline case (i.e., without BDSRs mounted onto the the concerns and implications associated with the effects of heat
test plate). Based on the measurement results as those shown in conduction through the test models on the adiabatic film cooling
effectiveness measurements. With the value of the Lewis number
(Le ¼ a=Ds ) being about 1.0, the thicknesses of the thermal bound-
ary layer and concentration boundary layer are of the same order,
the differential equations regarding heat and mass transfer can be
treated as analogous [21,24,25], which is the case of the present
study (Le  1).
For PSP experiments, the interested surface is coated with a
layer of oxygen-sensitive paint. This paint consists of lumino-
phores molecules hold together through gas-permeable polymeric
binder. The luminophores molecules in PSP paint emit light when
excited by certain UV light. However, the excited molecules will
return to ground state via a reduced or radiationless emission in
the presence of diatomic oxygen molecules. Such process is called
oxygen quenching [22] where the intensity of the photolumines-
cence is inversely proportional to the concentration of the local
oxygen. Consequently, the concentration of oxygen over the inter-
ested surface can be calculated based on the recorded light inten-
sity through a calibration procedure. To replace the temperature
terms with the concentrations of oxygen in Eq. (1), the adiabatic
film cooling effectiveness can be expressed as:





C o2 main  C o2 mix C o2 main  C o2 mix
g¼

 ¼
 ð2Þ
Fig. 2. Measured turbulent boundary layer profile of the mainstream flow over the C o2 main  C o2 coolant C o2 main
test plate.
446 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

For the choice of a coolant gas whose molecular mass differs were monitored by using a flow meter (Omega, FMA-1600A),
greatly from that of the mainstream, the film cooling effectiveness which were varied from 0.40 to 1.25. As a result, the corresponding
based on the measurements of the concentration of oxygen should momentum ratios (I ¼ qc V 2c =q1 V 21 ¼ M 2 =DR) were ranging from
be expressed as Eq. (3), as suggested by Charbonnier et al. [26], 0.10 to 1.02, and the velocity ratio (V r ¼ V c =V 1 ) changed from
where MW is the molecular weight ratio of the coolant gas to the 0.26 to 0.82.
mainstream gas. In the present study, as suggested by Johnson & Hu [28], a
1 spatially-averaging procedure with interrogation windows of
g ¼ 1  h     i ð3Þ 9
9 pixels and 50% overlap rate was performed for the acquired
po2 = po2  1 MW þ 1 PSP images, in order to minimize the measurement uncertainties.
air mix wall
Since the acquired PSP images had a magnification of 0.07 mm/
The pressure terms in Eq. (3) can be determined by using the
pix, the spatial resolution of the PSP measurement results would
recorded intensity of emission light, which is directly related to
be about 0.30 mm (i.e., 0.07D). It is well known that the uncer-
the partial pressure of oxygen. The mathematic function between
tainty of measured film cooling effectiveness by using PSP tech-
the normalized intensity and partial pressure can be obtained
nique is closely related to the PSP calibration process, quality of
through a dedicated PSP calibration procedure. Further technical
the acquired raw images, and the flow conditions of the main-
details of PSP technique and PSP calibration procedure can be
stream and coolant streams. Following up the work of Johnson &
found in published literatures of Wright et al. [21], Yang and Hu
Hu [28], the absolute measurement uncertainty of the present
[27], and Johnson et al. [23].
study was estimated to be within Dg ¼ 0:02 for the measured cen-
Figure 3 shows the schematic of the experimental setup used in
terline film cooling effectiveness. The relative measurement uncer-
the present study to map film cooling effectiveness distribution
tainty is estimated to be within 3% for g ¼ 0:6 and 5% for g ¼ 0:4.
over the test plate by using the PSP technique. As shown in the fig-
As for the laterally-averaged film cooling effectiveness, the abso-
ure, a UV LED light (LM2X-DM, ISSI) with wavelength of 390 nm
lute uncertainty would be within Dg ¼ 0:03, and the corresponding
was used as the excitation source for the PSP measurements. A
relative uncertainty is within 5% for g ¼ 0:6 and 7% for g ¼ 0:4 .
14-bit (2048 pixel
2048 pixel) charge-coupled device (CCD) cam-
These uncertainty values were estimated based on a Taylor Series
era (PCO2000, Cooke Corp.) fitted with a 610 nm long-pass filter
Method [29,30]. Further details about the uncertainty analysis of
was used to records the intensity of the photoluminescence light
the measured film cooling effectiveness by using PSP technique
emitted by excited PSP molecules. The PSP paint used in the study
can be found in Johnson & Hu [28].
was Uni-FIB provided by ISSI. This type of PSP paint has a low sen-
sitivity to temperature variation (0.5%/°C), making it an ideal can-
didate for the film cooling study. The experimental study was 2.3. Flow field measurements using the 2-D PIV and Stereoscopic PIV
conducted in a room (22 °C) where the temperature fluctuation (SPIV) technique
was found to be less than 0.5 °C.
During the experiments, while the airflow from the wind tunnel In addition to the PSP measurements, a high-resolution Particle
was used to simulate the hot gas flow in a typical turbine stage, an Image Velocimetry (PIV) system was also used in the present study
oxygen-free gas (i.e., CO2 with a density ratio of DR = 1.5 for the to conduct detailed flow field measurements to quantify the
present study) was supplied to simulate the ‘‘coolant” stream for dynamic mixing process between the coolant jet stream and main-
the PSP measurements. The oxygen-free coolant gas, which was stream flows over the test plate. Fig. 4 shows the schematic of the
supplied by a pressurized CO2 gas cylinder (99.99% purity), passed experimental setup used for the PIV measurement. During the
through a long pipeline and a constant temperature thermal bath experiments, the mainstream airflow and the coolant gas stream
before finally entering the plenum chamber underneath the test (i.e., CO2 for the present study) were seeded with 1 lm oil
plate. The blowing ratios (i.e., mass flux ratios, M ¼ qc V c =q1 V 1 ) droplets generated by using droplet generators. Illumination was
provided by a double-pulsed Nd:YAG laser (NewWave Gemini
200) emitting two pulses of 200 mJ at the wavelength of 532 nm
with a repetition rate of 10 Hz. By using a set of mirrors and optical
lenses, the laser beam was shaped into a thin light sheet
with a thickness of about 1.0 mm in the measurement region.

Fig. 3. Experimental setup used for film cooling effectiveness measurements with
PSP technique. Fig. 4. Experimental test rig for PIV measurements.
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 447

The illuminating laser sheet was aligned along the flow direction of the SPIV measurements in the cross planes normal to the main-
mainstream, bisecting the coolant hole in the middle of the test stream flow direction were obtained from a sequence of 1,000
plate, to conduct 2D PIV measurements in the streamwise plane frames of instantaneous PIV measurement results. The uncertainty
(i.e., within X-Y plane) at first. Then, the laser sheet was rotated level for the PIV measurements is estimated to be within 3% for the
90 degrees to perform stereoscopic PIV (SPIV) measurements at instantaneous velocity vectors, while the uncertainties for the
different downstream locations behind the exits of the coolant measured ensemble-averaged flow quantities such as vorticity dis-
injection holes to reveal the flow structures in the cross plane tributions are about 10%.
normal to the mainstream flow direction (i.e., in the Y-Z planes).
In the present study, two high resolution 14-bit high-resolution
digital cameras (PCO2000, Cooke Corp.) were used for the SPIV 3. Measurement results and discussions
image acquisitions. The two cameras were arranged with an angu-
lar displacement configuration of about 45° in order to get a large 3.1. Verification of measured film cooling effectiveness by using PSP
overlapped view. With the installation of tilt-axis mounts, laser technique
illumination plane, the lenses, and the cameras were adjusted to
satisfy the Scheimpflug condition. The cameras and double- While a number of experimental studies have been conducted
pulsed Nd:YAG laser were both connected to a Digital Delay Gen- in recent years by using PSP technique to achieve quantitative
erator (Berkeley Nucleonics, Model 565) to control the timing of measurements of adiabatic film cooling effectiveness, PSP tech-
the laser pulses and image acquisitions. A general in-situ calibra- nique is still a relatively new technique for film cooling studies,
tion procedure was performed to obtain the mapping functions in comparison to conventional temperature-based measurement
between the image planes and the object planes for the SPIV mea- methods. It is necessary to validate the reliability of the PSP mea-
surements. A target plate (150 mm
150 mm) with 0.5 mm surements in order to make PSP technique as an effective experi-
diameter dots spaced at intervals of 1.0 mm was used for the in- mental tool for turbine blade film cooling studies. With this in
situ calibration. The mapping function used in the present study mind, a comparative study was conducted to provide a quantita-
was a multi-dimensional polynomial function, which is fourth tive comparison of the measured film cooling effectiveness by
order for the directions parallel to the laser illumination plane using PSP technique against those derived from temperature-
(i.e., Y-axis and Z-axis directions), and second order for the direc- based measurements of the previous studies [32–34] under same
tion normal to the laser sheet plane (i.e., X-axis direction). The or comparable test conditions. It is worth noting that, the spanwise
coefficients of the multi-dimensional polynomial were determined periodicity of the coolant streams were first checked in the present
from the calibration images by using a least-square method as study, and very similar coolant coverages were observed for the all
used by Hu et al. [31]. the coolant injection holes. Therefore, only the measurement
For the 2D PIV measurements in the streamwise plane, instan- results behind the coolant injection hole in the middle of the test
taneous flow velocity vectors were obtained by using a frame-to- plate are shown here for conciseness. It should also be noted that,
frame cross-correlation technique to process the acquired PIV the centerline cooling effectiveness profiles given in the present
images with an interrogation window size of 32 pixels
32 pixels. study refer to the measured film cooling effectiveness values along
An effective overlap of 50% of the interrogation windows was the centerline of the coolant injection hole, while the laterally-
employed in PIV image processing, which results in a spatial reso- averaged cooling effectiveness is determined by averaging the
lution of 0.4 mm (i.e., 0.1D) for the PIV measurement results. measured film cooling effectiveness along the spanwise direction
Similar process methodology was also used for SPIV image pro- over one full period of the hole spacing (i.e., in the region of
cessing. The instantaneous 2D velocity vectors were then used to 1.5 6 Z/D 6 1.5) behind the coolant injection hole. As shown in
reconstruct all three components of the flow velocity vectors in Fig. 5, the measured centerline cooling effectiveness profile of the
the laser illuminating plane by using the mapping functions present study was found to agree with those of the previous
obtained through the calibration procedure. After the instanta- studies with temperature-based measurements well over the
neous flow velocity vectors (ui , v i , wi ) were determined, the distri- entire compared region (i.e., up to the downstream location of
butions of the ensemble-averaged flow quantities such as mean X/D > 30). For the laterally-averaged film cooling effectiveness,
velocity (U,V,W), spanwise vorticity (xz ) for the 2D PIV measure- while the measurement results of the present study was still found
ments in the streamwise plane and streamwise vorticity (xx ) for to be consistent with the archived measurement data of the

1 Present study M=0.60

0.5 Present study M=0.60
Schmidt et al (1996) M=0.60
Laterally averaged effectiveness ηLA

Sinha et al (1991) M=0.60 Schmidt et al (1996) M=0.60

Centerline effectiveness ηCL

Pedersen et al (1977) M=0.52

0.8 Pedersen et al (1977) M=0.52
0.4

0.6
0.3

0.4 0.2

0.2 0.1

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
x/D x/D
(a) Centerline film cooling effectiveness (b) Laterally-averaged effectiveness
Fig. 5. Comparison of the measured film cooling effectiveness of the present study with those of the previous studies at the test conditions of M = 0.60, DR = 1.53, and P/D = 3.
448 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

previous studies reasonably well in general, the PSP based mea- the performance of film cooling greatly at relatively high blowing
surement results of the present study seems to be slightly greater ratios (i.e., M > 0.85), it is highly desirable to elucidate the underly-
than those of the temperature-based measurements, especially in ing physics pertinent to the improved film cooling effectiveness.
the region near the exit of the coolant injection hole (i.e., X/D < 7.0). With this in mind, a high-resolution PIV system was used to con-
duct detailed flow field measurements to quantify the dynamic
3.2. Effects of the blowing ratio on the film cooling effectiveness behind mixing between the mainstream and coolant stream flows over
BDSR the test plate. Figs. 7 and 8 shows the PIV measurement results
in the streamwise plane (i.e., X–Y plane as shown in Fig. 4) for
Figure 6 shows the measured film cooling effectiveness distri- the baseline case (left column) and the new configuration with
butions over the surface of the test plate for the new film cooling BDSRs (right column) at the blowing ratios of M = 0.40 and
design with BDSR at different blowing ratios (i.e., mass flux ratios), M = 1.00, respectively. The instantaneous vorticity distributions
in comparison with those of the baseline case (i.e., the case without shown in Figs. 7(a) and 8(a) reveal that, as exhausted from the
BDSR). As shown clearly in Fig. 6(a), while the effectiveness of the coolant injection holes, the coolant jet streams would mix with
film cooling over the test plate was found to be much better for the the mainstream flows intensively over the test plate. Due to the
test cases with BDSR than that of the baseline case at relatively Kelvin–Helmholtz instabilities in the shear layers between the
large blowing ratios (i.e., at M > 0.85), the existence of the BDSR coolant streams and mainstream flows, unsteady vortex structures
has very limited or even detrimental effects on the film cooling were found to be generated and shed periodically in the shear lay-
effectiveness at relatively low blowing ratios (i.e., at M < 0.85). As ers over the test plate. As expected, the appearance of the vortex
revealed clearly in Johnson et al. [23], at relatively low blowing structures was found to be closer to the surface of the test plate
ratios (e.g., M  0.4), the coolant streams out of the coolant injec- at relatively low blowing ratio of M = 0.4, in comparison to that
tion holes would remain attached to the surface of the test plate at relatively high blowing ratio of M = 1.00.
to provide a nice protection to the surface of interest. As shown As shown clearly in Fig. 7(b), the coolant streams exhausted
in Fig. 6, the existence of the BDSR would actually cause slightly from the coolant injection holes would stay attached to the surface
detrimental effects on the film cooling effectiveness over the test of the test plate in general at the relatively low blowing ratio of
plate at the relatively low blowing ratio of M = 0.40, especially in M = 0.40 for both the baseline case and the test case with BDSR.
the region near to the exits of the coolant injection holes. However, After a careful inspection of the flow features for the two compared
at relatively high blowing ratio of M P 0.85, the regions with high cases, the streamlines exhausted from the coolant injection hole
cooling effectiveness values were found to become much narrower for the baseline case seem to be able to stay slightly closer to the
and shorter for the baseline case, indicating the high momentum surface of the test plate, indicating a slightly better film cooling
coolant jet streams separated from the surface of plate. Under such protection to the surface of interest at the low blowing ratio of
conditions, the existence of the BDSR was found to cause much M = 0.40, in comparison to the case with BDSR. The experimental
better coverage of the coolant stream over the surface of the inter- observation of the flow pattern is found to be in good agreement
est in both spanwise and stream-wise directions, in comparison with the measured film cooling effectiveness distributions
with those of the baseline case without BDSR. It suggests that described above.
the existence of the BDSR is very beneficial to improve film cooling However, the situation was found to change completely at the
effectiveness over the surface of interest, even with the same relatively high blowing ratio of M = 1.00. As revealed clearly in
amount of coolant fluid exhausted from the same coolant injection Fig. 8, after exhausted from the coolant injection hole, the stream-
holes. lines of the coolant jet stream for the baseline case were found to
The effects of the BDSR on the film cooling effectiveness over separate from the surface of the test plate and penetrate into the
the surface of interest were revealed more quantitatively from mainstream flow at the blowing ratio of M = 1.00, causing poor film
the quantitative comparison of the centerline and laterally- cooling protection over the surface of interest. While a similar
averaged cooling effectiveness profiles between the test cases with trend was also observed for the case with BDSR, the separation
the BDSR and the baseline case, which are shown in Fig. 6 and penetration of the coolant jet stream were found to become
(b) and (c) at relatively low blowing ratios (i.e., M 6 0.85) and rel- much mild due to the existence of the BDSR on the test plate. A
atively high blowing ratios (i.e., M > 0.85), respectively. It can be more careful inspection of the flow features around the BDSR
seen clearly that, at the relatively low blowing ratio of M = 0.40, reveals that a downwash flow was found to be generated at the
the absolute values of the measured cooling effectiveness (i.e., for top of the BDSR to suppress the take-off of the coolant jet stream
both centerline and laterally-averaged profiles) for the test cases from the surface of the test plate. This is believed to be closely
with BDSR were found to be slightly lower than those of the base- related to the suction effect induced by the BDSR, which will be
line case, which is consistent with the spatial contour distributions discussed further in the later section. As a result, the film cooling
shown in Fig. 6(a). The film cooling effectiveness for the configura- effectiveness over the surface of interest was found to become
tion with BDSR was found to start to exceed that of the baseline higher for the BDSR case than that of the baseline case at the blow-
case as the blowing ratio increases to M = 0.85. As the blowing ratio ing ratio of M = 1.0, which was revealed quantitatively from the
becoming greater than 0.85 (i.e., M > 0.85), the film cooling perfor- measured distributions of the film cooling effectiveness given in
mance for the configuration with BDSR was found to completely Fig. 6.
surpass the baseline case with an augmentation of 40–150% in
the laterally-averaged cooling effectiveness values. Such measure- 3.4. Stereoscopic PIV measurements for the BDSR case and the baseline
ment results indicate that the BDSR would significantly improve case
the performance of film cooling over the surface of interests, espe-
cially for the lateral-averaged cooling effectiveness, at relatively As described above, film cooling, fundamentally, is a jet-in-
high blowing ratios. cross-flow (JICF) problem. The dominated flow features for a JICF
include shear-layer-related vortices formed at the jet leading edge,
3.3. PIV measurements for the test cases with BDSR and baseline cases horseshoe vortices around the jet, counter-rotating-vortex (CRV)
pair in the jet, and wake vortices downstream of the jet [3]. As
While the measured film cooling effectiveness results given shown in the PIV measurement results given in Fig. 8, the coolant
above reveals clearly that the existence of the BDSR would improve jet stream tends to separate from the surface of interest and
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 449

(a) Cooling effectiveness distributions for the baseline case (left) and the configuration with BDSR (right)
0.8 M=0.40, Baseline 0.6
M=0.40, Baseline
M=0.40, BDSR (H=0.5D) M=0.40, BDSR (H=0.5D)

Laterally averaged effectiveness ηLA

0.7 M=0.85, Baseline M=0.85, Baseline
M=0.85, BDSR (H=0.5D) 0.5
Centerline effectiveness ηCL

M=0.85, BDSR (H=0.5D)

0.6
0.4
0.5

0.4 0.3

0.3
0.2
0.2
0.1
0.1

0 0
0 5 10 15 20 0 5 10 15 20
x/D x/D

(b) The centerline and laterally-averaged cooling effectiveness profiles at M=0.40 and 0.85
0.8 M=1.00, Baseline 0.6 M=1.00, Baseline
M=1.00, BDSR (H=0.5D) M=1.00, BDSR (H=0.5D)
Laterally averaged effectiveness ηLA

0.7 M=1.25, Baseline M=1.25, Baseline

M=1.25, BDSR (H=0.5D) 0.5 M=1.25, BDSR (H=0.5D)
Centerline effectiveness ηCL

0.6
0.4
0.5

0.4 0.3

0.3
0.2
0.2
0.1
0.1

0 0
0 5 10 15 20 0 5 10 15 20
x/D x/D
(c) The measured centerline and laterally-averaged cooling effectiveness at M=1.00 and 1.25
Fig. 6. Comparison of the measured film cooling effectiveness between the new configuration with BDSR (H = 0.5D, L = 0.9D) and the baseline case at various blowing ratios.

penetrate into the mainstream flow due to the induction from the In the present study, a stereoscopic PIV (SPIV) system was also
strong CRV at relatively high blowing ratios, which leads to poor used to reveal the generation and evolution of various streamwise
film cooling performance over the surface of interest. As demon- vortex structures in the coolant jet stream for the test case with
strated by the work of Zaman et al. [5] and Shinn et al. [6], a BDSR mounted on the test plate, in comparison with those of the
promising method to improve film cooling effectiveness is to baseline case in order to gain further insight into the inherent
reduce the detrimental effects of the vortex induction of the CRV mechanism pertinent to the improved film cooling performance
by introducing a pair of anti-CRV to weaken the strength of CRV due to the effects of the BDSR. Fig. 9 shows some typical SPIV mea-
to make the coolant jet stream stay attached to the surface of surement results (i.e., ensemble-averaged measurement results) in
interest. the cross planes normal to the flow direction of the mainstream
450 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

(a) Instantaneous vorticity distributions

(b) Ensemble-averaged streamlines and velocity fields

Fig. 7. PIV measurement results for the baseline case (left column) and the case with BDSR (right column) at a relatively low blowing ratio of M = 0.40.

(a) Instantaneous vorticity field

(b) Ensemble-averaged velocity field

Fig. 8. PIV measurement results for the baseline case (left column) and the case with BDSR (right column) at a relatively high blowing ratio of M = 1.00.

(i.e., within Y–Z plane) at the blowing ratio of M = 1.00. As shown additional pair of streamwise counter-rotating vortices would be
clearly in Fig. 9(a), a pair of streamwise counter-rotating vortices generated behind the BDSR, besides the CRVs induced by the cool-
(CRV) were found to be generated behind the coolant injection hole ant jet stream. The existence of the counter-rotating streamwise
as the salient flow structure for the baseline case, where the near- vortex pair behind the two horns of a Barchan dune was also
wall coolant flow was found to eject away from the surface of the observed by Zheng et al. [35] and Omidyeganeh et al. [36]. Since
test plate by an upwash flow formed between the CRV pair. the rotation directions of the streamwise vortices induced by the
The SPIV measurements given in Fig. 9(b) reveal clearly that, BDSR were found to be opposite to those of the CRVs in the coolant
with a BDSR mounted in front of the coolant injection hole, an jet stream, they are referred as anti-CRVs. The anti-CRVs could
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 451

(a) Baseline case in the X/D=3.0 cross plane (b) BDSR case in the X/D=3.0 cross plane.

(c) Evolution of the vortex structures for the baseline case (d) Evolution of the vortex structures for the BDSR case
Fig. 9. Stereoscopic PIV measurement results for the baseline and the test case with BDSR.

countervail the detrimental effects of the CRVs in the coolant jet X/D = 1.3, 3.0, and 8.0, respectively). It can be seen clearly that,
stream. Consequently, the separation of the coolant jet stream the flow features for the baseline case were characterized with a
from the surface of the test plate would be postponed due to the pair of large-scale streamwise CRVs behind the coolant injection
existence of the BDSR, in comparison to the baseline case. As a hole and a pair of horseshoe vortices (HV) with much lower vortex
result, as shown clearly in Fig. 8(b), downwash flows were found strength at the two sides. As reported in many previous studies on
to be generated in the regions between the streamwise vortices, JICF [3,37–39], these streamwise vortex structures are the domi-
which would push the lifted coolant jet streams back to the surface nating flow features of an inclined jet flow. As shown clearly in
of interest for a better film cooling protection. Fig. 9(c), the streamwise vortex structures would be dissipated
In comparison with the baseline case, the existence of the BDSR gradually due to the viscous effects, as they travel downstream.
would also make a faster spanwise spreading of the coolant stream For the test case with BDSR, while the features of the vortex struc-
over the test surface. The distance between the CRVs in the coolant tures were found to be similar as those of the baseline case, the
jet stream was found to be elongated in the lateral direction due to strength of the anti-CRVs generated behind the BDSR were found
the existence of the BDSR, in comparison to the baseline case. The to be much stronger than that of the HVs for the baseline case.
flow features would result in a wider coverage of the coolant The anti-CRV was found to interact intensively with the CRVs in
stream over the surface of interest, as revealed clearly from the the JICF, which would suppress the lift-up of the coolant jet stream
measured film cooling effectiveness distributions given in Fig. 6. from the surface of interest.
It can also be seen that, a relatively low-velocity region (i.e., as Based on the SPIV measurements as those given in Fig. 9, a
indicated in light blue color in the Fig. 9(b)) was found to exist in sketch of the flow features behind the BDSR was conjectured,
the near wall region for the test case with BDSR, which indicates which is given in Fig. 10. The diagram vividly illustrates the unique
a ‘‘quiet” wake flow behind the BDSR. Such flow features were also flow features behind the BDSR, which can be used to reveal the
revealed clearly in the 2-D PIV measurements along the stream- underlying physics pertinent to the improved film cooling perfor-
wise plane, as shown in Fig. 8. mance for the case with BDSR at relatively high blowing ratios.
Figure 9(c) and (d) show the evolution of the ensemble- As shown in the diagram, the vortex structures behind the BDSR
averaged streamwise vorticity distributions for the two compared are highly three-dimensional, featured by two pairs of large-scale
cases at three typical streamwise locations (i.e., at the locations of vortices. The first pair of vortices were found to be in the horizontal
452 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

Fig. 10. Schematic of the vortex structures behind the BDSR.

plane (i.e., X–Z plane) behind the BDSR, as reported by Christensen the settings with the BDSR being placed either upstream or down-
et al. [15]. This pair of vortices could generate continuous suction stream of the coolant injection hole. With the BDSR located at the
along the ridges of the BDSR to make the coolant stream to spread upstream of the coolant injection hole (i.e., the distance L becom-
out more widely and uniformly over the surface of the test plate. ing negative), the film cooling performance for the test case with
The second pair of the vortices are the streamwise anti-CRVs gen- L = 0.9D was found to be better than that of the test case with
erated behind the horns of the dune, which are in the vertical plane L = 3.0D at both M = 0.85 and M = 1.25. It is because that the
normal to the flow direction of the oncoming mainstream flow (i.e., downwash flows and suction effects induced by the BDSR would
within Y–Z plane). The streamwise anti-CRVs were found to coun- become stronger as the BDSR was placed more closely to the cool-
teract the detrimental effects of the CRVs in the coolant jet stream ant injection hole.
to generate downwash flows to bring the separated coolant stream However, for the configuration with the BDSR located at the
back to the surface of interest. As a result, the coolant jet stream, downstream of the coolant injection hole, the test case with
after exhausted from the coolant injection hole, would more likely L = 3.4D was found to exhibit a higher film cooling effectiveness
to stay attached to the surface of interest and provide a better cov- than that of L = 2.7D case. It suggests that the film cooling effec-
erage over the test plate along the spanwise direction. In summary, tiveness would be slightly better if the BDSR was placed at a fur-
the augmentation of the film cooling effectiveness for the test case ther downstream location away from the coolant injection hole.
with BDSR as revealed in Fig. 6 is suggested to be caused by the By analyzing the film cooling effectiveness for all the compared
downwash effects induced by the streamwise anti-CRVs generated cases, the test case with L = 0.9D (i.e., BDSR located at the
behind the BDSR along with the dynamic interactions among var- upstream of the coolant injection hole with a smaller distance of
ious complex vortex structures generated in the wake flow of the L/D = 0.9D) was found to display the highest film cooling effective-
BDSR. ness at relative high blowing ratios, followed closely by the test
case of L = 3.4D (i.e., with the BDSR located at a further down-
3.5. The effects of BDSR location in relation to the coolant injection stream location away from the coolant injection hole).
hole on the film cooling performance Figure 12 shows the PIV measurement results in the streamwise
plane with the BDSR located at either upstream (left column, with
The film cooling for a gas turbine blade is usually achieved by L = 0.9D) or downstream (right column, with L = 3.4D) of the cool-
injecting coolant flow from arrays of coolant injection holes. In ant injection hole at the blowing ratio of M = 1.0 (i.e., the corre-
such a setting, the BDSR placed upstream of a coolant injection sponding velocity ratio being 0.66, V c =V 1 ¼ 0:66). As described
hole (i.e., the configuration as shown in Fig. 6) could be located above, since a shorter distance between the BDSR and the down-
at the downstream of the upstream coolant injection holes. While stream coolant injection hole would result in a faster laterally
the measurement results given in Fig. 6 indicate that the existence expanding and a stronger downwash effects induced by the BDSR,
of the BDSR at the upstream of a coolant injection hole can enhance the film cooling performance over the surface of interest was found
the film cooling effectiveness substantially at relatively high blow- to increase greatly when the BDSR was placed immediately
ing ratios, it is also meaningful to examine the effects of the BDSR upstream of the coolant injection hole (i.e., for the test case with
location in relation to the coolant injection hole on the film cooling L = 0.9D). As shown clearly in Fig. 12(a), due to the downwash
performance in order to explore/optimize design paradigms for effects induced by the BDSR, the streamlines of the mainstream
improved film cooling effectiveness with arrays of coolant injec- flow were found to bend downward slightly as passing over the
tion holes. Fig. 11 shows the measured film cooling effectiveness ridge of the BDSR. Consequently, the coolant jet streams were
as a function of the distance (L) between the BDSR and the coolant found be confined in the region closer to the surface of the test
injection hole at the blowing ratio of M = 0.85 (left column) and plate, in comparison to the case without the BDSR.
M = 1.25 (right column), respectively. It is apparent that there are As shown in Fig. 12(b), for the case with the BDSR placed at the
great differences in the measured film cooling effectiveness for downstream of coolant injection hole (i.e., L = 3.4D case), due to the
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 453

(a) Cooling effectiveness distributions at the blowing ratios of M=0.85 (left) and M=1.25 (right)
0.8 0.8
BDSR ( L=-3.0D), M=0.85 BDSR ( L=-3.0D), M=1.25
BDSR ( L=-0.9D), M=0.85 BDSR ( L=-0.9D), M=1.25
0.7 BDSR ( L= 2.7D), M=0.85 0.7 BDSR ( L= 2.7D), M=1.25
Centerline effectiveness ηCL

BDSR ( L= 3.4D), M=0.85 BDSR ( L= 3.4D), M=1.25

Centerline effectiveness ηCL

0.6 Baseline, M=0.85 0.6 Baseline, M=1.25

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0 0
0 5 10 15 20 0 5 10 15 20
x/D x/D
(b) Centerline cooling effectiveness profiles at M=0.85 (left) and M=1.25 (right)
0.6 BDSR ( L=-3.0D), M=0.85 0.6 BDSR ( L=-3.0D), M=1.25
BDSR ( L=-0.9D), M=0.85
Laterally averaged effectiveness ηLA

BDSR ( L=-0.9D), M=1.25

Laterally averaged effectiveness ηLA

BDSR ( L= 2.7D), M=0.85 BDSR ( L= 2.7D), M=1.25

0.5 0.5
BDSR ( L= 3.4D), M=0.85 BDSR ( L=3.4D), M=1.25
Baseline, M=0.85 Baseline, M=1.25
0.4 0.4

0.3 0.3

0.2 0.2

0.1
0.1

0
0 5 10 15 20 0
x/D 0 5 10 15 20
x/D

(c) Laterally-averaged cooling effectiveness profiles at M=0.85 (left) and M=1.25 (right)
Fig. 11. Measured film cooling effectiveness as a function of the distance between the BDSR and the coolant injection hole at the blowing ratios of M = 0.85 (left) and M = 1.25
(right).

Coanda effect (i.e., the phenomena in which a jet flow tends to 3.6. The effect of BDSR height on the film cooling performance
attach itself to a nearby surface and remains attached even when
the surface curves away from the initial jet direction), the sepa- Since the configurations with L = 0.9D and L = 3.4D were found
rated coolant jet stream was found to stay attached to the surface to exhibit better film cooling performance in comparison with the
of the BDSR. As a result, the separated coolant stream would be other test cases, these two configurations were chosen for further
able to re-attach to the surface of interest at the downstream loca- investigation to examine the effects of BDSR height on the film
tions behind the BDSR to provide a better film cooling perfor- cooling performance over the surface of interest. It should be noted
mance, as shown quantitatively in the measured film cooling that, all of the measurement results given above were obtained
effectiveness distributions given in Fig. 11. with the height of the BDSR being 0.5D (i.e., H = 0.5D). In order to
454 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

(a) BDSR at the upstream of the coolant hole (b) BDRS at the downstream of the coolant hole
Fig. 12. PIV measurement results with BDSR located at either upstream (i.e., L = 0.9D) or downstream (i.e., L = 3.4D) of coolant injection hole at the blowing ratio of M = 1.0.

evaluate the effects of BDSR height on the film cooling perfor- injection hole (i.e., the configuration with L = 3.4D), the film cool-
mance, a new set of BDSRs with a smaller height of H = 0.3D were ing effectiveness of the test case with a smaller BDSR height (i.e.,
designed and manufactured for the comparative study. Fig. 13 H = 0.3D) was found to be slightly better than that with greater
shows some typical measurements results in terms of the center- BDSR height (i.e., H = 0.5D case). It suggests that the smaller BDSR
line and laterally-averaged cooling effectiveness profiles. As shown could be a better choice for the configuration with the BDSR
clearly in Fig. 13(a), with the BDSR placed upstream of the coolant located at the downstream of the coolant injection hole.
injection hole (i.e., the configuration with L = 0.9D), the test case
with a higher BDSR (i.e., H = 0.5D case) was found to exhibit an 3.7. Area-averaged film cooling effectiveness
improved film cooling performance than that with a smaller BDSR
height (i.e., H = 0.3D case) at both the blowing ratios of M = 0.85 In the present study, a parameter, named the area-averaged
and M = 1.25. This is believed to be caused by the stronger down- film cooling effectiveness, is introduced, to compare the overall
wash and suction effects with a higher BDSR. For the configuration performance of the film cooling effectiveness for the different test
with the BDSR placed at the downstream of the of the coolant cases. The area-averaged film cooling effectiveness, gareaav eraged , is

0.8 0.6
H=0.3D, L=-0.9D, M=0.85 H=0.3D, L=-0.9D, M=0.85
H=0.5D, L=-0.9D, M=0.85 H=0.5D, L=-0.9D, M=0.85
0.7
Laterally averaged effectiveness ηLA

H=0.3D, L=-0.9D, M=1.25 H=0.3D, L=-0.9D, M=1.25

0.5
H=0.5D, L=-0.9D, M=1.25 H=0.5D, L=-0.9D, M=1.25
Centerline effectiveness ηCL

0.6
0.4
0.5

0.4 0.3

0.3
0.2
0.2
0.1
0.1

0 0
0 5 10 15 20 0 5 10 15 20
x/D x/D

(a) Measured cooling effectiveness with the BDSR placed at the upstream of the coolant injection hole (L=-0.9D)

0.8 0.6
H=0.3D, L=3.4D, M=0.85 H=0.3D, L=3.4D, M=0.85
H=0.5D, L=3.4D, M=0.85 H=0.5D, L=3.4D, M=0.85
Laterally averaged effectiveness ηLA

0.7 H=0.3D, L=3.4D, M=1.25 H=0.3D, L=3.4D, M=1.25

0.5
H=0.5D, L=3.4D, M=1.25 H=0.5D, L=3.4D, M=1.25
Centerline effectiveness ηCL

0.6
0.4
0.5

0.4 0.3

0.3
0.2
0.2
0.1
0.1

0 0
0 5 10 15 20 0 5 10 15 20
x/D x/D

(b) Measured cooling effectiveness with the BDSR at the downstream of the coolant injection hole (L=3.4D)
Fig. 13. Effects of the height of the BDSR on the film cooling performance for the configurations with the BDSR at either upstream (L = 0.9D) or downstream (L = 3.4D) of the
coolant injection hole.
 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456 455

0.35 BDSR (H=0.5D, L=-0.9D) 0.2 BDSR (H=0.5D, L=-0.9D)

BDSR (H=0.3D, L=3.4D) BDSR (H=0.3D, L=3.4D)
Baseline Baseline
0.18

Area averaged effectiveness

0.3

Standard deviation
0.16
0.25
0.14

0.2
0.12

0.15 0.1

0.1 0.08
0.40 0.60 0.80 1.00 1.20 0.40 0.60 0.80 1.00 1.20
Mass flux ratio, M Mass flux ratio, M

(a) area-averaged film cooling effectiveness (b) Corresponding standard deviation values
Fig. 14. The area-averaged film cooling effectiveness and the corresponding standard deviation values as a function of the mass flux ratio.

RR
defined as gareaav eraged ¼ A1 gxz dA, where A is the area of the inter- on the ground from being blown away by oncoming airflow, the
est over the test plate (i.e., the region of 1 6 X/D 6 15 and 1.5 6 Z/ BDSR concept was developed to generate a series of complex vor-
D 6 1.5 for the present study), and gxz is the corresponding cooling tex structures in the wake behind the BDSR, which would be able
effectiveness at each location within the area of the interest. The to make the coolant stream flow stay more firmly on the surface
standard deviation values of the measured film cooling effective- of interest for improved film cooling effectiveness. The experimen-
ness within the interest of area were also calculated in the present tal study was conducted in a low-speed, open-circuit wind tunnel.
study. Obviously, a higher value of the area-averaged film cooling While a Pressure Sensitive Paint (PSP) technique was used to map
effectiveness, gareaav eraged , would indicate a better overall film cool- the adiabatic film cooling effectiveness on the surface of interest
ing performance. A lower standard deviation value of the cooling based on a mass-flux analog to traditional temperature-based
effectiveness would represent a more uniform film cooling cover- cooling effectiveness measurements, a high-resolution PIV system
age over the area of interest. was used to conduct detailed flow field measurements in both
Based on the experimental results described above, three repre- streamwise and cross-flow planes to quantify the dynamic mixing
sentative configurations were selected for the comparative study process between the coolant jet streams and the mainstream flows
in order to assess the overall film cooling performance. The three over the test plate. The effects of the coolant blowing ratio (i.e.,
selected configurations include (1) the test case with the BDSR coolant-to-mainstream mass flux ratio, M = 0.40, 0.85, 1.00, 1.25),
located at the upstream of the coolant injection with a higher BDSR the mounting location of the BDSRs in relation to the coolant injec-
height (i.e., the test case with H = 0.5D and L = 0.9D); (2).the test tion holes (i.e., L = 3.0D, 0.9D, 2.7D, 3.4D), and the height of the
case with the BDSR located at the downstream of the coolant injec- BDSRs (i.e., H = 0.3D, 0.5D) on the film cooling effectiveness over
tion with a smaller BDSR height (i.e., the test case with H = 0.3D the surface of interest were examined in great detail based on
and L = 3.4D); and (3) the baseline case without the BDSR. Fig. 14 PSP and PIV measurements.
shows the area-averaged film cooling effectiveness and the corre- The PSP measurement results revealed clearly that, with the
sponding standard deviations for the three selected configurations same amount of the coolant streams exhausted from the injection
as a function of the blowing ratio. As shown clearly in Fig. 14, the holes, the film cooling effectiveness over the surface of interest can
BDSR case with H = 0.5D and L = 0.9D was found to have the high- be enhanced greatly by mounting the BDSRs over the test plate,
est averaged film cooling effectiveness and lowest standard devia- especially for the test cases with relatively high coolant blowing
tion at most of the bowling ratios. In comparison with the baseline ratios (i.e., M P 0.85). More specifically, in comparison with those
case, an augmentation of 50–100% was found to be achieved in the of the baseline case without BDSRs, the area-averaged film cooling
area-averaged film cooling effectiveness for the BDSR case with effectiveness over the surface of interest was found to increase by
H = 0.5D and L = 0.9D at relatively high blowing ratios of 50–100% due to the existence of the BDSRs (i.e., the test case with
M > 0.85. The corresponding standard deviation values for the H = 0.5D, L = 0.9D), as the coolant blowing ratios being M P 0.85. In
BDSR case with H = 0.5D and L = 0.9D were also found to be much addition to the higher overall film cooling effectiveness, the cover-
smaller for most of the bowling ratios in comparison to the base- age of the coolant streams over the surface of the interest, espe-
line case. All these results indicate that the film cooling configura- cially along the spanwise direction, was also found to become
tion with BDSR can provide a better and more uniform film cooling much more uniform due to the existence of the BDSRs.
over the surface of interest, especially for the test conditions that The PIV measurements reveal clearly that, the flow structures
require high blowing ratios. generated in the wake behind a BDSR are highly three-
dimensional, featured by two pairs of large-scale vortex structures.
The first pair of vortices were found to be formed in the horizontal
4. Conclusion plane behind the dune, as reported by previous studies. This pair of
vortex structures could generate continuous suction along the
An experimental study was conducted to explore a novel film ridges of the BDSR to make the coolant streams to spread out more
cooling design with a Barchan-dune-shaped-ramp (BDSR) concept widely and uniformly over the surface of the test plate. The second
to improve the film cooling performance over the surface of a test pair of the vortices are the streamwise anti-CRVs generated behind
plate with coolant streams injected from discrete circular holes at the horns of the BDSR, which are in the vertical plane normal to the
an injection angle of 35°. As inspired by the unique shape of direction of the oncoming mainstream flow. The streamwise anti-
Barchan dunes commonly seen in deserts to prevent sand particles CRVs were found to counteract the detrimental effects of the CRVs
456 W. Zhou, H. Hu / International Journal of Heat and Mass Transfer 103 (2016) 443–456

formed in the coolant jet stream to generate downwash flows to [15] K.T. Christensen, F. Kazemifar, G. Blois, J.M. Barros, Z. Tang, A. Hamed, et al., in:
Quantitative studies of environmental flows at the micro- and macro-scales,
suck the separated coolant stream back to the surface of interest.
16th Int. Symp. Flow Vis. Okinawa, Japan, 2014.
As a result, the coolant jet stream would more likely to stay [16] B. Horvat, Barchan dunes Seminar 2, (n.d.) <http://www-f1.ijs.si/~rudi/sola/
attached to the surface of interest and provide a better coverage Sem4.pdf>.
over the test plate along the spanwise direction. In summary, the [17] P. Hersen, K. Andersen, H. Elbelrhiti, B. Andreotti, P. Claudin, S. Douady,
Corridors of barchan dunes: Stability and size selection, Phys. Rev. E. 69 (2004)
augmentation of the film cooling effectiveness due to the existence 011304, http://dx.doi.org/10.1103/PhysRevE.69.011304.
of the BDSR is believed to be closely related to the downwash [18] S. Ou, J.-C. Han, A.B. Mehendale, C.P. Lee, Unsteady wake over a linear turbine
effects induced by the streamwise anti-CRVs generated behind blade cascade with air and CO2 film injection: Part I—effect on heat transfer
coefficients, J. Turbomach. 116 (1994) 721, http://dx.doi.org/10.1115/
the BDSR along with the dynamic interactions among various com- 1.2929465.
plex vortex structures generated in the wake behind of the BDSR. [19] J. Han, S. Dutta, S. Ekkad, Gas turbine Heat Transfer and Cooling Technology,
It should be noted that, while the findings of the present study CRC Press, 2012.
[20] S. Baldauf, M. Scheurlen, A. Schulz, S. Wittig, Correlation of film-cooling
demonstrated that the proposed BDSR concept can be used to aug- effectiveness from thermographic measurements at engine like conditions, J.
ment the film cooling effectiveness over the surface of interest by Turbomach. 124 (2002) 686, http://dx.doi.org/10.1115/1.1504443.
generating a series of complex vortex structures over the test plate, [21] L.M. Wright, Z. Gao, T.A. Varvel, J.-C. Han, in: Assessment of Steady State PSP,
TSP, and IR Measurement Techniques for Flat Plate Film Cooling, ASME 2005
more extensive investigations are still needed to elucidate the Summer Heat Transf. Conf. Collocated with ASME 2005 Pacific Rim Tech. Conf.
associated underlying physics and examine the relevant parame- Exhib. Integr. Packag. MEMS, NEMS, Electron. Syst. Heat Transf., Vol. 3, ASME,
ters, including discharge coefficients, heat transfer coefficients, 2005, pp. 37–46, http://dx.doi.org/10.1115/HT2005-72363.
[22] J.H. Bell, E.T. Schairer, L.A. Hand, R.D. Mehta, SURFACE pressure measurements
the pressure losses, the compressibility of the gaseous flows, the
using luminescent coatings 1, Annu. Rev. Fluid Mech. 33 (2001) 155–206,
geometry of the BDSR and the shape of the coolant injection holes, http://dx.doi.org/10.1146/annurev.fluid.33.1.155.
in order to explore/optimize design paradigms for better film cool- [23] B. Johnson, W. Tian, K. Zhang, H. Hu, An experimental study of density ratio
ing protection of turbine blades from harsh environments. effects on the film cooling injection from discrete holes by using PIV and PSP
techniques, Int. J. Heat Mass Transf. 76 (2014) 337–349, http://dx.doi.org/
10.1016/j.ijheatmasstransfer.2014.04.028.
[24] E.R.G. Eckert, Similarity analysis of model experiments for film cooling in gas
Acknowledgments turbines, Wärme- Und Stoffübertragung 27 (1992) 217–223, http://dx.doi.org/
10.1007/BF01589919.
The technical assistances of Dr. Blake Johnson, Mr. Bill Richard, [25] J.N. Shadid, E.R.G. Eckert, The mass transfer analogy to heat transfer in fluids
with temperature-dependent properties, J. Turbomach. 113 (1991) 27, http://
and Mr. Hao Guo of Iowa State University are greatly appreciated.
dx.doi.org/10.1115/1.2927734.
Support from the National Science Foundation – United States with [26] D. Charbonnier, P. Ott, M. Jonsson, F. Cottier, T. Köbke, Experimental and
Grant numbers of IRES-1064235 and CBET-1435590 is gratefully Numerical Study of the Thermal Performance of a Film Cooled Turbine
acknowledged. Platform, Proc. Asme Turbo Expo 2009 Vol 3, Pts A B (2009) 1027–1038.
[27] Z. Yang, H. Hu, Study of trailing-edge cooling using pressure sensitive paint
technique, J. Propuls. Power. 27 (2011) 700–709, http://dx.doi.org/10.2514/1.
B34070.
References [28] B. Johnson, H. Hu, Measurement uncertainty analysis in determining adiabatic
film cooling effectiveness by using pressure sensitive paint technique, J.
[1] D.G. Bogard, K.A. Thole, Gas Turbine Film Cooling, J. Propuls. Power. 22 (2006) Turbomach. 138 (2016) 121004, http://dx.doi.org/10.1115/1.4033506.
249–270, http://dx.doi.org/10.2514/1.18034. [29] M. Sajben, Uncertainty estimates for pressure sensitive paint measurements,
[2] D.G. Bogard, in: Airfoil film cooling, Gas Turbine Handb., National Energy AIAA J. 31 (1993) 2105–2110, http://dx.doi.org/10.2514/3.11897.
Technology Laboratory, 2006. Section 4.2.2.1. [30] S.J. Kline, F.A. Mcclintock, Describing uncertainties in single sample
[3] T.F. Fric, A. Roshko, Vortical structure in the wake of a transverse jet, J. Fluid experiments, J. Mech. Eng. 75 (1953).
Mech. 279 (1994) 1–47, http://dx.doi.org/10.1017/S0022112094003800. [31] H. Hu, T. Saga, T. Kobayashi, N. Taniguchi, A study on a lobed jet mixing flow by
[4] R.J. Goldstein, E.R.G. Eckert, F. Burggraf, Effects of hole geometry and density on using stereoscopic particle image velocimetry technique, Phys. Fluids. 13
three-dimensional film cooling, Int. J. Heat Mass Transf. 17 (1974) 595–607, (2001) 3425, http://dx.doi.org/10.1063/1.1409537.
http://dx.doi.org/10.1016/0017-9310(74)90007-6. [32] D.L. Schmidt, B. Sen, D.G. Bogard, Film cooling with compound angle holes:
[5] K.M. Zaman, D. Rigby, J. Heidmann, Inclined jet in crossflow interacting with a adiabatic effectiveness, J. Turbomach. 118 (1996) 807, http://dx.doi.org/
vortex generator, J. Propuls. Power. 26 (2010) 947–954. 10.1115/1.2840938.
[6] A.F. Shinn, S. Pratap Vanka, Large eddy simulations of film-cooling flows with a [33] A.K. Sinha, D.G. Bogard, M.E. Crawford, Film-cooling effectiveness downstream
micro-ramp vortex generator, J. Turbomach. 135 (2013) 011004. of a single row of holes with variable density ratio, J. Turbomach. 113 (1991)
[7] B. An, J. Liu, C. Zhang, S. Zhou, Film cooling of cylindrical hole with a 442, http://dx.doi.org/10.1115/1.2927894.
downstream short crescent-shaped block, J. Heat Transfer 135 (2013) 031702, [34] D. Pedersen, Film cooling with large density differences between the
http://dx.doi.org/10.1115/1.4007879. mainstream and the secondary fluid measured by the heat-mass transfer
[8] S. Na, T. Shih, Increasing adiabatic film-cooling effectiveness by using an analogy, J. Heat Transfer 99 (1977) 620–627.
upstream ramp, J. Heat Transfer 129 (2007) 464–471. [35] Y. Zheng, A. Rinoshika, Multi-scale vortical structure analysis on large eddy
[9] E. Sakai, T. Takahashi, Y. Agata, Experimental study on effects of internal ribs simulation of dune wake flow, J. Vis. 18 (2015) 95–109, http://dx.doi.org/
and rear bumps on film cooling effectiveness, J. Turbomach. 135 (2013) 10.1007/s12650-014-0227-0.
031025, http://dx.doi.org/10.1115/1.4007546. [36] M. Omidyeganeh, U. Piomelli, K.T. Christensen, J.L. Best, in: Large-eddy
[10] R. Goldstein, Film cooling, Adv. Heat Transf. 7 (1) (1971) 321–379. simulation of flow over barchan dunes, Mar. River Dune Dyn. – MARID IV,
[11] R.S. Bunker, A review of shaped hole turbine film-cooling technology, J. Heat Bruges, 2013, pp. 191–198.
Transfer 127 (2005) 441, http://dx.doi.org/10.1115/1.1860562. [37] D.J. Needham, N. Riley, J.H.B. Smith, A jet in crossflow, J. Fluid Mech. 188
[12] B. Andreotti, P. Claudin, S. Douady, Selection of dune shapes and velocities Part (1988) 159, http://dx.doi.org/10.1017/S0022112088000680.
1: Dynamics of sand, wind and barchans, Eur. Phys. J. B – Condens. Matter. 28 [38] M. Ilak, P. Schlatter, S. Bagheri, D.S. Henningson, Bifurcation and stability
(2002) 321–339, http://dx.doi.org/10.1140/epjb/e2002-00236-4. analysis of a jet in cross-flow: onset of global instability at a low velocity ratio,
[13] E. Parteli, O. Durán, H. Herrmann, Minimal size of a barchan dune, Phys. Rev. E. J. Fluid Mech. 696 (2012) 94–121, http://dx.doi.org/10.1017/jfm.2012.10.
75 (2007) 011301, http://dx.doi.org/10.1103/PhysRevE.75.011301. [39] D.R. Getsinger, L. Gevorkyan, O.I. Smith, A.R. Karagozian, Structural and
[14] J.A. Palmer, R. Mejia-Alvarez, J.L. Best, K.T. Christensen, Particle-image stability characteristics of jets in crossflow, J. Fluid Mech. 760 (2014) 342–367,
velocimetry measurements of flow over interacting barchan dunes, Exp. http://dx.doi.org/10.1017/jfm.2014.605.
Fluids. 52 (2011) 809–829, http://dx.doi.org/10.1007/s00348-011-1104-4.

View publication stats