13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
Boundary Layers Transition Measurement over an Airfoil
by Using PIV with High Magnification
Kijung Kwon1, Byeonghee Chang2, Jangyeon Lee3, Seung O Park4
1: Aerodynamics Department, Korea Aerospace Research Institute, Daejeon, Korea, kjkwon@kari.re.kr
2: Aerodynamics Department, Korea Aerospace Research Institute, Daejeon, Korea, cbh@kari.re.kr
3: Aviation Technology Department, Korea Aerospace Research Institute, Daejeon, Korea, jylee@kari.re.kr
4: Department of Aerospace Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea,
sopark@kaist.ac.kr
Abstract Boundary layer velocity vectors over a 16% elliptic airfoil at a Reynolds number of 3×105 were
measured by using a two dimensional particle image velocimetry with high magnification. To see the inside
of the thin boundary layer in detail the magnification factor was increased up to about 60 pixels/mm by
inserting extension tube between the lens and CCD camera. Tests were performed with and without
boundary layer transition trip attached on the both surfaces at 10% chord for comparison between laminar
and turbulent boundary layers. Surface mount hot film sensor was also used to determine boundary layer
transition from laminar to turbulent or relaminarization from turbulent to laminar by inspecting intermittency
factors. The purpose of this research was to determine the boundary layer transition behavior as angle of
attack changes. The experimental results indicates that as angle of attack increases boundary layer transition
position moves toward leading edge on the suction side while laminar boundary layer remains unchanged on
the pressure side with smooth surface condition. On the contrary, turbulent boundary layer on the pressure
side of trip surface condition is relaminarized as angle of attack increases causing early laminar boundary
layer separation. This verifies that the results revealed in previous research concerning lift curve slope higher
than 2πα at low angles of attack and lower than 2πα at high angles of attack, which is quite different from the
case of conventional airfoils. From the comparison between velocity profile shapes and intermittency factors,
boundary layer shape factors can be a good measure in determining transition even though there exists
pressure gradient along the airfoil surface.
1. Introduction
Wings for unmanned vehicles are often operated at low Reynolds number conditions. At such
low Reynolds numbers, aerodynamic characteristics can be dominated by laminar flow phenomena
including separation bubbles around leading edge and/or mid-chord of the upper surface of the
airfoil (Ingen 1985). Flow behaviors inside a boundary layer can hardly be observed by
experimental means since the boundary layer formed on the surface of a scaled model in wind
tunnel testing is very thin below several millimeters. Gleyzes (1985) reported boundary layer
profiles measured by hot-wire anemometry and laser Doppler velocimetry (LDV), but their results
were confined to a very high angle of attack (α) with relatively thick boundary layer. Lang (2002)
successfully used particle image velocimetry (PIV) technique to investigate laminar to turbulent
transition in the boundary layer of an artificially generated separation bubble.
The primary objective of the previous research was to investigate aerodynamic characteristics of
a 16% thickness elliptic airfoil at the Reynolds number of 3×105(Kwon 2005). After close looking
at the aerodynamic characteristics such as lift, drag and pitching moment, velocity fields around and
behind the airfoil were also surveyed by using a PIV system to explain unusual aerodynamic
characteristics. As the angle of attack increases in the range from α = 0° to 4°, the boundary layer
separation point on the suction side moves toward the trailing edge, while the separation point on
the pressure side remains fixed. This asymmetric flow behavior alters surface pressure distribution,
which in turn increases lift in a similar manner as if the airfoil had a flap.
-1-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
This paper focuses on the boundary layer velocity profiles and boundary layer transition
positions over an airfoil, since boundary layer transition is directly related to boundary layer
separation. Measurements were taken in a subsonic wind tunnel by using PIV with high
magnification and surface mount hot film sensors. To see inside the thin boundary layer the
magnification factor was increases up to about 60 pixels/mm and this high magnification was
obtained by inserting an extension tube between the lens and CCD camera. Intermittency factors (γ),
which indicated the boundary layer was laminar or turbulent, were calculated by hot film sensor
signals. Measurements are taken for the airfoil with and without boundary layer transition trip.
Boundary layer trip technique is generally used in wind tunnel tests to simulate full scale or high
Reynolds number flow in low Reynolds number airfoil test condition in a laboratory to match the
transition positions and/or to eliminate laminar separation or separation bubbles (Gopalarathnam
2003). The main purpose of using boundary layer trip in this study is, however, to help understand
the effects of transition on aerodynamic characteristics.
2. Experimentation
Tests were conducted in a low-speed wind tunnel at Korea Aerospace Research Institute (KARI),
of which the test section is 1 m wide, 0.75 m high, and 2 m long. The maximum achievable wind
speed is 110 m/s and the turbulence intensities for empty test section condition are 0.12% in stream
wise direction and 0.15% in vertical and lateral directions. The cross sectional shape of the airfoil
was an ellipse whose major axis, the chord length (c), was 200 mm and the minor axis 32 mm. The
span of the model was 990 mm yielding the aspect ratio of 4.95. The model was constructed of
aluminum and its surface was coated with thin black vinyl to reduce scattering of laser light for PIV.
The Reynolds number based on the chord was 3×105 with the free stream velocity (U0) of about 25
m/s. The angle of attack was varied from -16° to 16° with the increment of 1°.
Tests were performed with and without boundary layer transition trip attached on the surface of
10% chord. Circular trip dots were used for the trip and the dot-Reynolds number was about 435.
The dots having a circular disk shape were made of Vinyl tape.
A two dimensional PIV system was used for the flow field measurements around the airfoil
surface to investigate boundary layer velocity profiles. Figure 1 shows the schematic of the PIV
measurement set up.
Fig. 1. Schematic of the PIV measurement
The laser light source was 200 mJ dual head pulsed Nd:YAG laser (Quantel) and its wave length
was 532 nm in the air. Laser guide arm from DANTEC was used to guide the laser from source to
-2-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
sheet beam optics. The sheet beam had an expanding angle of about 30˚ yielding the width of the
sheet to be about 300 mm at the measurement location. The thickness of the sheet was around 1mm.
The laser sheet plane was parallel to the wind direction to measure the stream wise and vertical
components of velocity. Optical glasses of 12.7 mm thickness were mounted on the ceiling and the
side walls of the test section. The flow was seeded with DEHS (Di-Ethyl-Hexyl-Sebacat, C26H50O4)
particles of nominal diameter of 1 µm generated from Laskin nozzle (TSI, Model 9307). A CCD
camera of 2k × 2k resolution (Kodak MEGAPLUS ES 4.0) was positioned 0.6 m apart from the
laser sheet. Two images were captured with a time delay of 4 µs and processed by a two-frame
cross-correlation method. Two image pairs were captured in 1 second, and 64 image pairs were
averaged for one measurement condition. The CCD camera and the laser source were synchronized
by a delay generator (BNC, Model 555), and for the main trigger signal, the trigger output from the
1st laser source was used. The interrogation window size was 32 pixel wide and 8 pixel high. Since
the particle movements were dominant in the stream wise direction close to the surface, rectangular
integration window was used for the boundary layer measurement. This interrogation window
corresponds to 0.53 by 0.133 mm in the x and y directions respectively. Since the interrogation
windows were overlapped by 50% in the cross-correlation calculation, the distances between the
two neighboring velocity vectors were 0.265 mm and 0.067 mm in the x and y directions,
respectively. Final velocity vectors were obtained by transforming from Cartesian coordinates to
airfoil surface coordinates, i.e. parallel and normal to the surface.
There are several sources of uncertainty in velocity measurements using PIV. One of the main
sources is related to PIV system hardware and the other comes from the analysis method of post
processing. The present analysis is concerned only with uncertainties in the PIV system effects and
follows the AIAA standard (AIAA 1995). The time delay between pulses is measured by photo
detector with 4 GHz oscilloscope. The frequency response of the photo detector was 1.2 GHz, and
the accuracy of the delay generator was 1.4 ns. As a result of the time delay measurement between
the two laser pulses, the uncertainty of the time was estimated to be around 10 ns. The major
uncertainty for distance came from the refraction caused by the optical glass window of the test
section side wall. By theoretical calculation using Snell’s law as well as direct measurement, the
maximum uncertainty at the edge of the images caused by the refraction was found to be 2.6 µm
that corresponded to 1% error for displacement measurement. The maximum overall uncertainty for
velocity vectors were around 0.75 m/s.
In addition to the velocity profile measurement, intermittency factors were measured by using
surface mount hot film sensors. The sensor was attached on the airfoil model surface and details are
shown in Fig. 2.
Flow
7 mm
Hot-Film
15 mm
Lead Wire
Solder
Fig. 2. Surface mount hot-film sensor (DANTEC 55R47)
-3-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
Signals were conditioned by constant temperature anemometer, StreamLine(DANTEC 90N10)
and acquired with 4 kHz sampling rate by NI PCI 6036E. In general this sensor is used to measure
surface shear stress after calibration, but in this study the raw sensor signal was enough to see flow
regime. Digital signals were processed statistically to calculate intermittency factors, which are the
measure of laminar or turbulent. In calculating intermittency factor, there are various methods of
distinguishing the turbulence state.(Canepa 2002). The ‘Contol on window residence time’ method
was used in this study.
3. Results
Boundary layer velocity profiles normalized by the free stream velocity vs. normal distance (y)
from the airfoil surface normalized by the airfoil chord are shown in Fig. 3a, b on the suction and
pressure sides, respectively at various chord wise positions (x). The velocity profiles were obtained
by linearly interpolating the velocity components of neighboring points along the direction normal
to the airfoil surface, since the velocity vectors extracted from the PIV measurements were arranged
in a rectangular wind-coordinate system.
0.020
Smooth
Trip
x/c=0.1
0.2
0.3
0.4
0.7
0.8
0.9
0.015
α=0˚
y/c
0.010
0.005
0.000
0.000
0.020
0.020
0.015
0.015
Smooth
Trip
x/c=0.1
α=2˚
0.010
0.005
0.000
0.000
0.020
0.000
0.000
0.020
0.015
0.015
α=4˚
0.010
0.005
0.005
0.000
0.020
0.000
0.000
0.020
0.000
0.015
a
0.7
0.8
0.9
α=2˚
α=4˚
0.015
α=6˚
0.010
α=6˚
y/c
y/c
0.010
0.4
y/c
y/c
0.010
0.3
y/c
0.005
y/c
y/c
y/c
0.010
0.2
0.005
0.005
0.000
0.000
V/Uo
V/Uo
b
V/Uo
V/Uo
Fig. 3a, b. Velocity profiles. a on the suction pressure side. b on the suction side: ♦, smooth; ◇, trip
Boundary layer separation phenomenon can be easily determined by the velocity profiles. The
velocity profiles for α = 0° and 2° on the suction side (Fig. 3a) indicate that the flow separates at
-4-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
around x/c = 0.8 for the smooth airfoil while no separation occurs for the tripped airfoil. When the
angle of attack increases further, no signature of separation is seen in the velocity profiles up to x/c
= 0.9 on the suction side for both the smooth and tripped airfoils. On the other hand, the flow on the
pressure side of the smooth airfoil in Fig. 3b separates between x/c = 0.8 and 0.9 at all angles of
attack. For the case of tripped airfoil, however, the flow on the pressure side is seen to separate only
at higher angles of attack. As is the case for the aerodynamic coefficients, velocity profiles for both
the smooth and the tripped airfoil are nearly identical each other at α = 6°.
The primary problem of determining transition position with velocity profiles on an airfoil
surface is that well-known velocity profiles for laminar or turbulent is acceptable for flows over a
flat plate, which means no pressure gradient along stream-wise direction. Since there exists a
pressure gradient over an airfoil, we can roughly distinguish whether the boundary layer is laminar
or turbulent. In Fig. 3a, between x/c = 0.2 and 0.3 when angle of attack is 0° boundary layer of trip
condition changes from laminar to turbulent while that of smooth condition remains laminar on the
suction side. Laminar boundary layer separation of smooth condition and attached turbulent
boundary layer of trip condition are easily observed between x/c = 0.8 and 0.9. When angle of
attack is 6°, boundary layer profiles are all turbulent irrespective of the surface condition. On the
pressure side, however, turbulent boundary layer caused from transition trip dot is relaminarized as
angle of attack increases(Fig. 3b).
Smooth
0.08
0.07
0.06
y/c
0.07
0.18
0.08
0.2
Trip
0.22
0.24
0.26
0.05
0.8
0.07
0.28
0.82
0.84
0.86
0.88
0.9
0.82
0.84
0.86
0.88
0.9
0.8
0.82
0.84
0.86
0.88
0.28 0.010.8
0.28
0.82
0.84
0.86
0.88
0.9
0.82
0.84
0.86
0.88
0.9
0.82
0.84
0.86
0.88
0.9
0.06
0.07
a
0.18
0.2
0.22
0.24
0.26
0.05
0.8
0.28
x/c
0.11
Smooth
0.03
0.1
0.02
y/c
0.09
0.18
0.2
Trip
0.22
0.24
0.26
0.01
0.28
0.03
0.1
0.02
0.09
b
0.18
0.2
0.22
0.24
0.26
x/c
-0.09
-0.03
Smooth
-0.04
-0.1
-0.11
y/c
-0.05
-0.030.18
0.2
0.22
0.24
0.26
0.28 -0.090.8
Trip
-0.04
-0.1
-0.11
-0.05
0.18
c
0.2
0.22
0.24
0.26
0.28
0.28
0.8
x/c
Fig. 4a-c. Boundary layer velocity vectors. a suction side at α = 0°, b suction side at α = 6°, c
pressure side at α = 6°
-5-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
Figure 4a-c. shows detail velocity vectors at the particularly selected area whose positions are
covered by dotted rectangular boxes in Fig. 3a, b. In Fig. 4a, it can be observed that the boundary
layer transition does not occur just behind trip dots which are attached at x/c = 0.1 but occurs at
around x/c = 0.26.
From the boundary layer profiles, shape factors (H) were calculated and are shown in Fig. 5. It is
well known that boundary layer shape factor for laminar flow on a flat plate is about 2.5 and for
turbulent flow 1.3. When angle of attack is 0° with trip condition, shape factors are close to 1.3 and
almost constant except at x/c = 0.1. At smooth condition, however, shape factors lie between 2 and
2.5 before separation point (indicated by S in Fig. 5). The reason that the shape factors does not
coincide with 2.5 is caused from the pressure gradient along the airfoil surface. On the suction side
of the smooth condition, it can be observed that the transition point (indicated by T in Fig. 5) moves
upward as angle of attack increases. At angle of attack 6° on the pressure side of trip condition,
shape factor increases along the airfoil surface between x/c = 0.4 and 0.8 indicating
relaminarization.
On the suction side with trip condition for all angle of attacks, shape factors are very high at x/c
= 0.1. This tells that local separation occurs just behind transition trip dots. Here we can see that
turbulent boundary layer does not separate. This means flow separation near trailing edge is directly
related to transition on the airfoil surface and consequently wake asymmetry which was observed in
the previous research should be explained in connection with boundary layer transition.
10
10
α = 0°
8
6
6
S
4
H
2
0.2
8
0.4
0.6
α = 4°
x/c
0.8
0
1.0
100.0
0.2
0.4
0.6
α = 6°
x/c
8
6
0.8
1.0
6
4
2
0
0.0
S
4
2
0
100.0
α = 2°
8
0.2
0.4
0.6
S
4
T
2
0.8
0
1.00.0
x/c
Fig. 5. Boundary layer shape factors:
side;
, trip on suction side;
relaminarization
S
T
0.2
R
0.4
0.6
0.8
1.0
x/c
, smooth on suction side;
, smooth on pressure
, trip on pressure side; S, separation; T, transition; R,
The intermittency is defined to be the fraction of time in which the flow is turbulent at a certain
position. Although the determination of the intermittency has not a well-defined procedure, these
values measured from hot film sensors indicate transition positions more clearly than velocity
profiles. Fortunately, even though there exists pressure gradient over the airfoil, transition positions
indicated by velocity profile shapes i.e. shape factors show good agreements with those by
intermittency factors. Figure 6a, b, shows intermittency along the airfoil surface as angle of attack
changes. Intermittency increases from 0 to 1 indicating transition as angle of attack increases on
-6-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
suction side of the smooth condition in Fig. 6a. Angle of attack between 4° and 5°, transition
position abruptly changes. In Fig. 6b, relaminarization can be observed as mentioned in the velocity
profile results. Even though there attached transition dots on the airfoil surface, increased pressure
on the pressure side suppresses transition to turbulent.
The intermittency factors measured by hot film sensors show a good agreement on the whole
with the results of velocity profiles in determination of turbulent transition. It can be said that
boundary layer shape factor calculated from velocity profiles can be a good measure of transition
although there is small discrepancy between shape factors of flat plate and of airfoil surface.
1
1
0.8
0.8
α increases
0.6
α increases
γ
γ
0.6
0.4
0.4
Smooth
Suction Side
0.2
Trip
Pressure Side
0.2
0
0
0.2
0.4
a
0.6
x/c
0.8
1
0.2
b
0.4
0.6
0.8
1
x/c
Fig. 6a, b. Intermittency factors. a suction side of smooth condition, b pressure side of trip
condition: ◆, α = 0°; ■, α = 1°; ▲, α = 2°; ○, α = 3°; *, α = 4°; ●, α = 5°; □, α = 6°
4. Concluding Remarks
Boundary layer transition on an elliptic airfoil was studied both by PIV and by surface mount hot
film. From PIV measurement, velocity profiles were measured and shape factors were calculated.
From surface mount hot film sensor measurement, intermittency factors were computed. The
comparison between velocity profiles and intermittency factors shows a good agreement in
distinguishing boundary layer transitions. The experimental results verifies that the wake
asymmetry which causes the unusual aerodynamic characteristics such as high lift curve slope and
high drag coefficient in low angle of attack range of an elliptic airfoil at can be explained in
connection with the boundary layer transition. At low Reynolds numbers, laminar boundary layer
on an elliptic airfoil separates near trailing edge because of blunt trailing edge shape. As angle of
attack increases, however, boundary layer on a suction side changes to turbulent delaying boundary
layer separation while that on a pressure side remains laminar. It can be concluded that the wake
asymmetry in a low angle of attack range which was reveled in the previous research are caused
from the different flow regime i.e. laminar or turbulent between suction and pressure sides as angle
of attack increases.
Acknowledgements
This study was supported by the Korea Research Council of Public Science and Technology and
partial support also came from the BK-21 project from the Ministry of Education.
References
AIAA (1999) Assessment of experimental uncertainty with application to wind tunnel testing.
AIAA Standard S-071A-1999.
-7-
13th Int Symp on Applications of Laser Techniques to Fluid Mechanics
Lisbon, Portugal, 26-29 June, 2006
#1191
Canepa, E., Ubaldi, M., and Zunino P. (2002) Experiments in the application of intermittency
detection techniques to hot-film signals in transitional boundary layers. VKI RP 2002-47, 16th
Symposium on Measuring Techniques in Transonic and Supersonic Flow in Cascade and
Turbomachines, Cambridge, United Kingdom.
Gleyzes, C., Cousteix, J., and Bonnet, J. L. (1985) Theoretical and experimental study of low
Reynolds number transitional separation bubbles. Proceedings of the Conference on Low
Reynolds Number Airfoil Aerodynamics, Edited by Mueller, T. J., Notre Dame, Indiana, USA,
pp. 137-152.
Ingen J. L. V., and Boermans, L. M. M. (1985) Research on laminar separation bubbles at Delft
University of Technology in relation to low Reynolds number airfoil aerodynamics.
Proceedings of the Conference on Low Reynolds Number Airfoil Aerodynamics, Edited by
Mueller, T. J., Notre Dame, Indiana, USA, pp. 89-124.
Kwon, K., and Park, S. O (2005) Aerodynamic characteristics of an elliptic airfoil at low Reynolds
number. Journal of Aircraft, Vol. 42, No 6, pp. 1642-1644.
Lang, M., Rist, U., and Wagner, S. (2002) Investigations on disturbance amplification in a laminar
separation bubble by means of LDA and PIV. 11th International Symposium on Applications
of Laser Techniques to Fluid Mechanics, Lisbon, Portugal
-8-