You are on page 1of 19

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

net/publication/281692112

Experimental investigation of 2D flexible plunging hydrofoil

Article  in  Journal of Flow Visualization and Image Processing · January 2015


DOI: 10.1615/JFlowVisImageProc.2015013527

CITATIONS READS
0 78

4 authors, including:

Robert L. Mitchell Leonardo Martin-Alarcon


Eaton, Peachtree City, GA, United States The University of Calgary
2 PUBLICATIONS   10 CITATIONS    5 PUBLICATIONS   5 CITATIONS   

SEE PROFILE SEE PROFILE

Fangjun Shu
New Mexico State University
37 PUBLICATIONS   211 CITATIONS   

SEE PROFILE

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

Hummingbird Flow Analysis View project

Experimental investigation of 2D flexible plunging hydrofoil View project

All content following this page was uploaded by Fangjun Shu on 28 September 2015.

The user has requested enhancement of the downloaded file.


Journal of Flow Visualization & Image Processing, 20(4), 243–260 (2013)

EXPERIMENTAL INVESTIGATION OF 2D
FLEXIBLE PLUNGING HYDROFOIL
Ruijun Tian, Robert Mitchell, Leonardo Martin-Alarcon, &
Fangjun Shu∗

New Mexico State University, Las Cruces, NM 88003



Address all correspondence to Fangjun Shu E-mail: fangjun@gmail.com

It has long been hypothesized that the flight of birds and insects greatly benefits from the flexibility
and morphing facility of their wings. A significant advantage that flapping flexible wing models have
over quasi-steady rigid wing models was a much higher lift generation capability. Both experimental
and computational studies have shown that the leading edge vortex (LEV) plays an important role
in this higher lift generation. In this study, we further explore the internal mechanisms behind the
production of these high lift forces. Two NACA0012 miniature wings, one flexible and one rigid, were
actively plunged at various frequencies in a viscous glycerol-water solution. Two-dimensional, phase-
locked particle image velocimetry (PIV) measurements were conducted to investigate the evolution of
vortices. Simultaneous measurements of lift and thrust forces were taken during plunging to reveal
the relationship between the force generation and the surrounding flow field. Results from the flexible
hydrofoil were compared directly to results from the rigid one to reveal the influence of flexibility.
These results can be used to benchmark future computational work.

KEY WORDS: flapping flight, wing flexibility, thrust force, PIV

1. INTRODUCTION

Modern aviation industry has not steered away from the use of fixed-wing vehicles
despite a plethora of technological improvements seen during this past century. In re-
cent years, the mechanisms of natural flexible flapping flight have received an increased
amount of attention due to a growing military and industrial interest to develop un-
manned micro air vehicles (MAVs). At low Reynolds number regimes, flapping flight
proves to be superior to fixed-wing flight in terms of lift efficiency, thrust force genera-
tion, and maneuverability. In 2000, Lai and Platzer (2009) performed experiments with a
plunging NACA 0012 airfoil with zero freestream velocities, verifying that pure plung-
ing motion could generate thrust at various plunging frequencies. Further investigations
on the physical mechanisms behind flapping-wing flight are needed in order to develop
highly efficient flapping-wing vehicles.
Various experimental and computational studies have elucidated the mechanisms be-
hind the large lift generation and high stability of flapping flight. Early studies performed

1065-3090/13/$35.00 ⃝
c 2013 by Begell House, Inc. 243
244 Tian et al.

on flapping wing MAVs (Koochesfahani, 1989; Dickinson and Götz, 1993) focused on
the flow features such as vortical patterns and velocity profiles in the wake. Subsequent
studies then focused on investigating the high lift and thrust force generation with flap-
ping flight. In 1999, Dickinson et al. (1999) carried out experiments with a pair of scaled
up drosophila wings and concluded that the high lift coefficient resulted from three dis-
tinct mechanisms: delayed stall, rotational circulation, and wake capture. Using digital
particle image velocimetry (DPIV) method and real-time force measurements, Birch and
Dickinson (2003) quantitatively determined the contribution of the wake capture to the
high lift generation in flapping motion. In 2009, an experimental parametric study of 2D
asymmetric hovering flapping motion was conducted by Jardin et al. (2009) for a NACA
0012 airfoil. Their time-resolved PIV results indicated that an angle of attack during
the upstroke, smaller than that for downstroke, could reduce the wake capture, thus con-
tributed to smooth development and close attachment of the LEV during the downstroke,
yielding a higher lift and drag generation.
Using high-performance computers, various numerical models have been proposed
and developed in order to reveal the role of LEVs and TEVs in flapping flight. Lewin
and Haj-Hariri (2003) performed numerical simulations of an airfoil at a series of flap-
ping frequencies and found out that the interactions between the LEV and the TEV were
significantly affected by the heaving frequency. Moreover, it was observed that the thrust
generation and power efficiency generated by the airfoil were greatly influenced by dif-
ferent types of vortex development around the model. Diverse numerical simulations
based on different models (Wang, 2000; Sun and Wu, 2003) have been used to study
the dynamic features and power efficiencies in flapping flight. These were qualitatively
analyzed by examining the development of LEV and TEV under various flapping pa-
rameters such as the Reynolds number, flapping frequency, and flapping amplitude.
There is a general consensus that simulating an insect wing using a rigid flat plate
underrepresents the flight characteristics of real insects, who benefit from their soft and
flexible wings. In 2006, DeLuca et al. (2006) experimentally compared the behavior
of rigid and flexible wing models in a wind tunnel and discovered that, for a larger
range of attack angles, flexible models yielded higher lift/drag ratios than rigid ones; this
suggested that flexible wings offer insects higher flight efficiency and maneuverability.
Aerodynamists have also paid attention to the influence of the flexibility on the fluid
flow. Experiments developed by Heathcote and Gurssul (2005) in 2004 showed that
flow features, such as the strength of vortices and time-averaged wake velocity profiles,
were affected by wing flexibility. Moreover, the thrust/input-power ratio was found to be
greater for flexible airfoils than for rigid ones. Additionally, it was confirmed that there
existed an optimal airfoil stiffness for a given plunging frequency and amplitude. More
numerical work about the flow features and force response for flapping flexible wing
models can be found in (Dong et al., 2007; Shyy et al., 2008; Yang et al., 2010).
The overall objective of this study was to investigate the 2D flow field development,
dynamic response, and power efficiency of flexible flapping hydrofoils that featured var-

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 245

ious degrees of stiffness. The specific aims were to use DPIV and high-resolution force
measurements to: 1) characterize the flow field around a flexible NACA 0012 hydrofoil
subjected to active plunging and passive pitching, 2) study the effects of Reynolds num-
bers and wing flexibility, and 3) measure the time history of lift and thrust generation for
each case. The obtained results can be used as a benchmark for validating computational
models featuring similar parameters.

2. EXPERIMENTAL SETUP AND METHODS

The experimental approach was designed to quantitatively examine the flow induced by
plunging hydrofoils. Both flexible and rigid NACA0012 hydrofoils were used. Hydro-
dynamic force and resulting flow structures were measured over a range of plunging
frequencies. To ensure a two-dimensional flow, the experiments employed hydrofoils
that spanned the entire width of the tank.

2.1 Hydrofoils

The flexible hydrofoil was made with synthetic silicone resin (Sylgard 184 elastomer)
using an aluminum mold. The cured resin was optically clear, which allowed velocity
measurements on both sides of the hydrofoil to be conducted simultaneously. To pre-
vent the formation of bubbles in the hydrofoil, the resin was cured in a vacuum chamber
(20 kPa) with a controlled temperature of 23◦ C for 12 hours. The Young and Shear mod-
uli for the resin were measured to be 0.43 MPa and 0.14 MPa, respectively. Note that
these material properties vary depending on curing conditions such as temperature. A
rigid hydrofoil of the same shape, made of aluminium, was used for comparison pur-
poses. Both hydrofoils had a NACA0012 airfoil cross-section profile; however, the two
hydrofoils had different chord lengths: 5 cm for the flexible hydrofoil and 7 cm for the
rigid one.

2.2 Plunging Mechanism

The hydrofoils were actuated by the crank-shaft system shown in Fig. 1(a). The system
was driven by an 8000-count servo motor (Pittman servo motor by AMETEK PMC) and
controlled by a Galil motion system (DMC-4060 from GALIL company).
An aluminum circular disk, a crank-arm and a plunging rod were used to convert
the circular motion of the disk into the linear motion of the plunging hydrofoil; the
motion of the rod was restricted to one degree-of-freedom by using a linear bearing. The
hydrofoil was connected to the driving shaft by using couplers. The plunging amplitude
was adjusted by choosing the appropriate crank-arm connection point on the rotatory
disk. The force sensor can be inserted between the hydrofoil and the driving shaft to
measure the hydrodynamic forces.

Volume 20, Number 4, 2013


246 Tian et al.

(a)

20
S real
Rsin(θ)-1.43
10
S (mm)

-10

-20

0 1 2 3 4 5 6 7
θ

(b) (c)

FIG. 1: The sketches for (a) the crank-shaft system, (b) the motion of the driving crank-
shaft with respect to that of the rotatory disk, and (c) the hydrofoil displacement

As indicated in Fig. 1(b), the origin of the hydrofoil’s plunging motion was defined
by its location when the center of the driving disk and its connecting point with the crank-
arm were horizontally aligned. When the disk rotated√ at a constant speed ω, the hydrofoil
plunged in a quasi-sinusoidal mode as S = H − L2 − R2 cos2 (ωt) + R sin(ωt),
where S is the displacement of the hydrofoil, H = 125.5 mm is the distance from the

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 247

center of the rotating disk to the origin, L = 127 mm is the crank-arm length, R =
18.9 mm is the distance from the center of rotation disk to its connection spot with the
crank-arm, and θ is the disk rotation angle. As shown in Fig. 1(c), the actual hydrofoil
displacement with respect to disk rotation angle was compared to a sinusoidal mode in
the same figure.

2.3 Flow Setup

To diminish the three-dimensional effects, the hydrofoils were installed in a fish tank
with one end close to, but not touching-the tank wall and another end close to an ad-
justable divider plate placed in the tank (as shown in Fig. 2). The dimensions of the fish
tank were 50 cm × 25 cm × 30 cm. A thin slot was cut in the divider plate for the driv-
ing shaft to go through. The spanwise dimensions for the flexible and rigid hydrofoils
were 10.2 cm and 9.8 cm, respectively. The working fluid used in these experiments was
a glycerin–water mixture with a volume ratio of 1.5:1.0, the fluid had a viscosity and
density of 5.41 × 10−6 m2 /s and 1.155 × 103 kg/m3 , respectively. The fluid was cho-
sen because of its relatively high viscosity; therefore, the plunging frequency was high
enough to generate hydrofoil deformation and still, the flow Reynolds number remained
relatively low.
After assembly, the leading edge of the hydrofoil was 15 cm from both the bottom
and left walls of the tank. This allowed enough free space for the vortex structures to
develop while reducing the far field effects created by the tank walls.

FIG. 2: Experimental setup

Volume 20, Number 4, 2013


248 Tian et al.

With the fish-tank setup mentioned above, the Reynolds number was recognized as
the main characteristic parameter for this unsteady problem. The Reynolds number was
defined as Re = U C/ν, where U is the characteristic velocity, C is the chord length,
and ν is the kinematic viscosity of the fluid. Due to the lack of incoming flow, the max-
imum plunging velocity U = 2πf R was chosen as the characteristic velocity. The other
important dimensionless parameter, the Strouhal number, was defined as St = f C/U ,
indicating a constant value for St = C/(2πR) independent of the variation of flapping
frequency. In this study, the Strouhal number was 0.42 and 0.59 for the flexible and rigid
hydrofoil, respectively.
During the PIV experiments four frequencies, i.e., 1 Hz, 2 Hz, 3 Hz, and 4 Hz, were
used to see the effects of different degrees of deformation on the vortex development
around the flexible hydrofoil. The maximum was set to be 4 Hz because the hydrofoil
would break at higher plunging frequencies. Accordingly, the Reynolds numbers were
1103, 2207, 3310, and 4413, respectively. To compare the results and maintain the same
Reynolds number, the plunging frequency for the rigid hydrofoils were thus set to be
0.7 Hz, 1.4 Hz, 2.1 Hz, and 2.8 Hz, respectively.

2.4 Particle Image Velocimetry (PIV) Setup

Flow fields around the plunging hydrofoils were measured using a 2D PIV system. The
system was composed of a double-pulse Nd–YAG laser (Quantel Inc.), a MotionPro X5
CMOS camera (Integrated Design Tools, Inc.) with a resolution of 2352 × 1728 pixels,
a timing hub (Integrated Design Tools, Inc.), optics to generate the laser sheet, and a
computer with image acquisition and processing software. The laser and optics were
set beneath the tank, as shown in Fig. 3, which allowed the laser sheet to illuminate
the region of interest through the center of the spanwise hydrofoil dimension. The laser
sheet was generated with the combination of a 1000 mm focal length spherical lens
and a −25.4 mm focal length cylindrical lens. In the flexible hydrofoil experiments the
camera covered a region of 212 mm × 156 mm, while for the rigid one it was 278 mm
× 205 mm. The calibration uncertainty was less than 0.5%.
The PIV camera was set normal to the laser sheet, at a height corresponding to the
center of plunging motion. The flow was seeded with glass microspheres with a mean
diameter of 11 µm. To ensure high quality PIV results, the number density of the parti-
cles was controlled so that there were 8 to 15 particles in a 32 × 32-pixel interrogation
window. The time intervals between laser pulses were adjusted to ensure the maximum
particle displacement was less than but close to 6 pixels.
Phase-locked PIV measurements were conducted at 32 evenly distributed phases in
each plunging cycle. To achieve this, the Galil motion controller software divided the
plunging motion into 32 phases and initiated triggering signals based on stepper mo-
tor position. For each phase, 100 pairs of images were recorded. Velocity fields were
evaluated with the fast Fourier transform (FFT)-based cross-correlation method using

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 249

FIG. 3: PIV setup

NASA PIVPROC software. PIV images were processed through a two-pass correlation
algorithm: 64 × 64 pixel interrogation windows followed by 32 × 32 pixel interroga-
tion windows with 50% overlap. The uncertainty of the ensemble-averaged velocity was
0.4% with 95% confidence level.

2.5 Force Measurement


In addition to PIV flow measurements, lift and thrust forces were acquired for each
plunging frequency. The forces were measured using a six-degree-of-freedom, silicon
strain gauge based force/torque sensor (Nano-17, ATI Industrial Automation). The force
measurement uncertainty was 1/160 N. Before each experiment, the force sensor was
validated using standard laboratory weights. The discrepancy for the force sensor was
below 0.5% for all the validations.
Within the ranges of the force limitation the flexible hydrofoil was plunged with
frequencies of 1.5 Hz, 2 Hz, 2.5 Hz, and 3.5 Hz to uncover its dynamic properties. Cor-
respondingly the rigid foil was flapped with frequencies of 1.06 Hz, 1.41 Hz, 1.77 Hz,
and 2.47 Hz to ensure the Reynolds number matches for both cases.
As shown prior in Fig. 1(a), the sensor was installed in the mounting hardware con-
necting the hydrofoil to the plunging rod; this arrangement provided a secure connection
while reducing unwanted hydrodynamic forces on the mounting bracket faces normal to
the plunging motion.

Volume 20, Number 4, 2013


250 Tian et al.

In every experiment, the plunging motion was started from the beginning of the
downstroke. Measurements were started after ten full plunging strokes to allow the full
development of the surrounding flow field. The dynamic force was recorded at a sam-
pling rate of 1000 Hz. The raw data were filtered using low-pass Butterworth digital filter
to attenuate the high-frequency electrical noises. Though the plunging motion was peri-
odic, minor differences were detected for each cycle. Thus the forces for 50 successive
periods were ensemble-averaged to obtain the dynamic forces for one entire plunging
cycle. The uncertainty bars with 95% confidence in the force results session were calcu-
lated based on the maximum deviations of the force measured in 50 cycles.

3. RESULTS AND DISCUSSION


Both flexible and rigid hydrofoils were evaluated to compare the effects of the flexibility
on vortex structure and hydrodynamic forces. As expected, the results varied as a func-
tion of hydrofoil rigidity. Specifically, the flexible hydrofoil produced thrust while the
rigid hydrofoil’s thrust generation was undetectable. Furthermore, although the vortex
structures appear similar, the flexible hydrofoil structure’s vortex center shifted down-
stream by a significant amount. These observations point towards increased efficiency
for the flexible hydrofoil.

3.1 Pitching Angle


In our experiment with the flexible hydrofoil, the model was actively plunged and pas-
sively pitched due to the flexibility and the interaction between the structure and the
fluid. Especially at higher frequencies, the fluid–structure interaction was strong enough
to generate a noticeable pitching angle that sometimes can be more than 20 degrees.
Figure 4 shows the pitching angle of the model in one period of plunging, starting
from the beginning of the downstroke. The pitching angle was measured as the angle
between the horizontal axis and the chord line of the deformed hydrofoil. The pitching
angle was defined positive when the trailing edge locates lower than the leading edge
and negative on the other hand. The hollow squares in Fig. 4 represent the experimental
results for 32 phases and the solid line is the sinusoidal curve fit of the data. Notice
that there was a phase lag between the pitching and plunging. For this case, the lag was
approximately 0.25 period, indicating that the pitching angle was related to the plunging
velocity of the hydrofoil.

3.2 PIV Results


The plunging cycle was evenly divided into 32 phases; PIV image pairs were taken at
each phase in order to visualize the evolution of the flow surrounding the hydrofoils.
Phase-averaged PIV results were presented in this section to reveal the flow field devel-
opment induced by the plunging hydrofoil.

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 251

20
20
Experimental Data
Sinusoidal Fitting
Plunging

Plunging Displacement (mm)


Pitching Angle (DEGREE)
10
10

0 0

-10
-10

-20
-20
0 0.2 0.4 0.6 0.8 1
t/T

FIG. 4: Pitching angle

In Figs. 5 and 6, the PIV results for the rigid hydrofoil at f = 2.8 Hz and the flexible
hydrofoil at f = 4 Hz were selected as examples to show the general development of
the surrounding vorticity field. Beginning with the phase where the driving rod of the
hydrofoil reached the top, 8 out of 32 successive phases were exhibited to show the
vortical structure evolution. For each sub-figure, t∗ represents the nondimensional time
defined as t∗ = t/T , where t is the time instant and T is the plunging period.
The 8 subimages from a to h in Fig. 5 represented the development of the flow field
around the rigid hydrofoil, while those in Fig. 6 were results for the flexible hydrofoil.
Because the laser sheet was projected from the bottom of the tank, the top side of the
rigid hydrofoil could not be illuminated, and thus, no PIV data was measured in this
region.
At the beginning of each stroke (down or up), both hydrofoils shed old LEVs and
TEVs and generated new vortices near both the leading and trailing edges because of
the reverse of plunging direction; nevertheless, the evolution of these vortexes was very
different in each case. For the flexible model, the LEV was stronger and attached to the
leading edge until the end of the stroke. This was mainly because of the passive pitching.
The LEV formed in the rigid hydrofoil was relatively weaker and detached before the end
of the stroke. However, the shed LEV pairs were not so quickly dissipated and remained
in front of the hydrofoil forming a forward jet flow, which yielded a drag force for the
model.
The development of the TEVs was also different because of the deformation. For
the flexible hydrofoil, the newly generated TEV in each stroke was stronger and kept at-

Volume 20, Number 4, 2013


252 Tian et al.

Vort 1/s Vort 1/s


100 100
78 78
150 56 150 56
33 33
Ymm

Ymm
11 11
-11 -11
100 -33 100 -33
-56 -56
-78 -78
-100 -100
50 50
50 100 150 200 50 100 150 200
Xmm Xmm

(a) t∗ = 0 (b) t∗ = 0.125

Vort 1/s Vort 1/s


100 100
78 78
150 56 150 56
33 33
Ymm

Ymm
11 11
-11 -11
100 -33 100 -33
-56 -56
-78 -78
-100 -100
50 50
50 100 150 200 50 100 150 200
Xmm Xmm

(c) t∗ = 0.25 (d) t∗ = 0.375

Vort 1/s Vort 1/s


100 100
78 78
150 56 150 56
33 33
Ymm

Ymm

11 11
-11 -11
100 -33 100 -33
-56 -56
-78 -78
-100 -100
50 50
50 100 150 200 50 100 150 200
Xmm Xmm

(e) t∗ = 0.5 (f) t∗ = 0.625

Vort 1/s Vort 1/s


100 100
78 78
150 56 150 56
33 33
Ymm

Ymm

11 11
-11 -11
100 -33 100 -33
-56 -56
-78 -78
-100 -100
50 50
50 100 150 200 50 100 150 200
Xmm Xmm

(g) t∗ = 0.75 (h) t∗ = 0.875

FIG. 5: Vorticity contour plot at different phases in one cycle (rigid hydrofoil at
f = 2.8 Hz)

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 253

Vort 1/s Vort 1/s


100.0 100.0
77.8 77.8
100 55.6 100 55.6
33.3 33.3
Ymm

Ymm
11.1 11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(a) t∗ = 0 (b) t∗ = 0.125

Vort 1/s
Vort 1/s
100.0
100.0
77.8
100 100 77.8
55.6
55.6
33.3
33.3
Ymm

Ymm
11.1
11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(c) t∗ = 0.25 (d) t∗ = 0.375

Vort 1/s Vort 1/s


100.0 100.0
100 77.8 100 77.8
55.6 55.6
33.3 33.3
Ymm

Ymm

11.1 11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(e) t∗ = 0.5 (f) t∗ = 0.625

Vort 1/s Vort 1/s


100.0 100.0
100 77.8 100 77.8
55.6 55.6
33.3 33.3
Ymm

Ymm

11.1 11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(g) t∗ = 0.75 (h) t∗ = 0.875

FIG. 6: Vorticity contour plots at different phases in one cycle (flexible hydrofoil at f =
4 Hz)

Volume 20, Number 4, 2013


254 Tian et al.

tached to the trailing edge until the reverse pitching; the two detached vortices generated
during the previous entire stroke were still strong in the wake. The vortices induced a
rearward jet with high strength to produce a large thrust force for the flexible hydrofoil.
For the rigid hydrofoil, the TEVs were weaker and quickly detached, so the rearward jet
was weaker if existed, resulting in a lesser thrust force.
The selected flow field results for the flexible hydrofoil at two different frequencies,
f = 1 Hz and f = 3 Hz, were presented for comparison. As shown in Fig. 7, the sur-
rounding vorticity contour in the low frequency case showed similar features as that for
the rigid hydrofoil because the low plunging frequency caused little deformation in the
model. As a comparison, when the frequency was increased to 3 Hz, the flexible model
pitched significantly, especially at the midstroke as indicated in Figs. 8(b) and 8(d). A
comparison of the results at the two frequencies further reveals the effects of the flexi-
bility in the development of the vortex around the model.

3.3 Force Measurement


In MAV-related research, force (thrust and lift) generation and the energy efficiency are
important topics. In this study, hydrodynamic forces were measured for both the rigid
and flexible hydrofoils plunging at various frequencies.

Vort 1/s Vort 1/s


50.0 50.0
100 38.9 100 38.9
27.8 27.8
16.7 16.7
Ymm

Ymm

5.6 5.6
-5.6 -5.6
50 -16.7 50 -16.7
-27.8 -27.8
-38.9 -38.9
-50.0 -50.0
50 100 150 200 50 100 150 200
Xmm Xmm

(a) t∗ = 0 (b) t∗ = 0.25

Vort 1/s Vort 1/s


50.0 50.0
100 38.9 100 38.9
27.8 27.8
16.7 16.7
Ymm

Ymm

5.6 5.6
-5.6 -5.6
50 -16.7 50 -16.7
-27.8 -27.8
-38.9 -38.9
-50.0 -50.0
50 100 150 200 50 100 150 200
Xmm Xmm

(c) t∗ = 0.5 (d) t∗ = 0.75


FIG. 7: Vorticity contour plot at different phases in one period (flexible hydrofoil at f =
1 Hz)

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 255

Vort 1/s Vort 1/s


100.0 100.0
100 77.8 100 77.8
55.6 55.6
33.3 33.3
Ymm

Ymm
11.1 11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(a) t∗ = 0 (b) t∗ = 0.25

Vort 1/s Vort 1/s


100.0 100.0
100 77.8 100 77.8
55.6 55.6
33.3 33.3
Ymm

Ymm
11.1 11.1
-11.1 -11.1
50 -33.3 50 -33.3
-55.6 -55.6
-77.8 -77.8
-100.0 -100.0
50 100 150 200 50 100 150 200
Xmm Xmm

(c) t∗ = 0.5 (d) t∗ = 0.75


FIG. 8: Vorticity contour plot at different phases in one period (flexible hydrofoil at f =
3 Hz)

The two plots in Fig. 9 were ensemble-averaged nondimensional hydrodynamic


lift force measured for (a) the rigid and (b) the flexible hydrofoil. The lift force was
nondimensionalized by F = (ρ(2πf R)2 )/2BC, where ρ is the fluid density, f is the
plunging frequency, R is the plunging amplitude, B and C are the spanwise and chord
length of the hydrofoil, respectively. The hydrodynamic lift (plunging) forces were plot-
ted with the buoyancy, gravity, inertial force and added mass effects subtracted. Here
we used the theoretical results from Sedov (1965) to derive the added mass effects as
madd = (π/4)ρC 2 B, where ρ is the density of the fluid.
Both the rigid and flexible hydrofoils generated a roughly sinusoidal lift force with
approximately zero mean value in one period. The results for the rigid hydrofoil also
indicated that the nondimensional lift force did not vary much under different plunging
frequencies, which meant the effects of the plunging frequency on the lift/plunging force
were negligible. The reason for the frequency independence would be that the effective
area of the surface perpendicular to the lift force direction was maintained the same
under different frequencies.
A comparison between the two plots also indicated that under lower frequencies the
flexibility did not significantly affect the magnitude of lift/plunging force because the
deformation was not significant at low frequencies. However, at a high Reynolds number

Volume 20, Number 4, 2013


256 Tian et al.

(a) (b)

FIG. 9: The nondimensional lift force in one period for (a) the rigid hydrofoil and (b)
the flexible hydrofoil

around 4000, i.e., f = 3.5 Hz for flexible model and f = 2.47 Hz for rigid one, the flexible
model experienced considerably lower lift/plunging forces. The relatively high degree of
deformation yielded high pitching angles. Thus, the effective area of the surface for the
plunging force, the projection of the hydrofoil plate on the horizontal plane, became
smaller, resulting in a decreased lift force coefficient at high flapping frequencies.
Nondimensionalized in the same way as the lift/plunging forces, the non-dimensional
thrust forces for the flexible hydrofoil are presented in Fig. 10. The thrust forces are ap-
proximately one order of magnitude less than the lift forces. It should be pointed out that
when the model was plunged, some unexpected vibration was induced in the horizontal

FIG. 10: The nondimensional thrust force in one period for flexible hydrofoil at different
frequencies

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 257

direction perpendicular to the leading edge because of imperfections in the mechanical


system. Fortunately, this vibration was very small in both the magnitude (less than 5%
of the chord length) and duration (less than 3% of the period) and therefore had lit-
tle effects on lift/plunging force measurement. However, because the thrust forces were
much lower, the vibration leads to an increased uncertainty in the thrust measurement.
Disregarding the effects of the vibration, at each plunging frequency two peaks were
observed in Fig. 10 corresponding to downstroke and upstroke, respectively. Because of
the sensitivity of the force sensor, the uncertainty is relatively high even after phase av-
eraging, especially under lower frequencies where the absolute values of the thrust force
were comparably low. Qualitatively the nondimensional thrust force is symmetric in the
down- and upstrokes and reached its maximum at f = 2.5 Hz in the present experiments,
which indicated that the optimal frequency resulting in a maximum non-dimensional
thrust force was around f = 2.5 Hz.
Thrust forces for the rigid hydrofoil were not presented because of the high measure-
ment uncertainty caused by vibration. The thrust force for the rigid hydrofoil should be
very low because, if one considered the rigid hydrofoil as a straight plate, no thrust force
would be generated during plunging due to its symmetric profile. The rigid hydrofoil
used in this study has a NACA0012 profile; though it is not symmetric with respect to
the vertical center line, the overall thrust force should be close to 0.
In Fig. 11, the time-averaged horizontal velocity profiles in front of the leading edge
and behind the trailing edge offered a visual insight in thrust generation. For both the
flexible and rigid hydrofoils, two jets were generated during plunging: one was forward
and the other one was backward. At higher frequencies, the flexible hydrofoil produced
a stronger backward jet and a weaker forward jet, indicating the generation of thrust
force. Though the backward jet generated by the rigid hydrofoil was also stronger than

Um/s Um/s Um/s Um/s


-0.3 0 0.3 -0.3 0 0.3 -0.4 0 -0.4 0 0.4
150 150 200 200
1Hz
2Hz
0.7Hz
3Hz 1.4Hz
4Hz 2.1Hz
2.8Hz

100 100 150 150


Ymm

Ymm

Ymm

Ymm

100 100
50 50

50 50
50 100 150 200 50 100 150 200
Xmm Xmm

(a) (b)
FIG. 11: The flow jet generated by (a) the flexible hydrofoil and (b) the rigid hydrofoil

Volume 20, Number 4, 2013


258 Tian et al.

the forward jet, the reversed flows in the backward jet near the trailing edge counteracted
the momentum flux and produced little thrust.
In Table 1, the mean thrust force coefficient indicated by CT AV E was defined as
CT AV E = FT AV E /F , where FT AV E was the average of the thrust force in one entire
stroke and F was the characteristic force defined in previous section. The thrust coef-
ficient indicated by CT was calculated as CT = (4RFT AV E )/Einput and Einput =
∫T
0 Fdrive V dt, where R was the plunging amplitude, Fdrive was the real-time readings
of the force sensor in the vertical direction, V was the instantaneous plunging velocity
approached by the sinusoidal function, and T was the time duration of∑one period of flap-
ping motion. The integration was performed with the summation of N i=1 (T /N )Fdrivei
× Vi .
The thrust coefficient indicated the efficiency of the hydrofoil in generating a thrust
force. For the current flexible hydrofoil, the value of efficiency was monotonously in-
creasing until the maximum frequency f = 3.5 Hz. The time course of the thrust force
plot in Fig. 10 indicated that the high power efficiency at high plunging frequencies
mainly came from the decreased driving force from the motor. Due to the existence
of large deformation and pitching angles, a lesser driving force was needed to actively
plunge the model, yielding a less input power to the system.
The thrust coefficients in Table 1 were consistent with the vorticity contour analysis.
Within a certain range, the thrust coefficient increased with plunging frequency. In this
experiment, the maximum thrust coefficient was generated around f = 2.5 Hz. There-
after, the thrust coefficients decreased with the plunging frequency, indicating that an
optimal plunging frequency exist for flexible foils. This frequency should be dependent
of the foil geometry, flexibility, and Reynolds number.

4. CONCLUSIONS
In the present experiment, an integrated flow setup was built to study the evolution of
vortical flow generated by plunging hydrofoils. Both rigid and flexible hydrofoils were
investigated. Aerodynamic thrust and plunging forces were measured and analyzed for
both hydrofoils. The hydrofoils were plunged at various frequencies to address its in-
fluences in hydrofoil deformation, flow field evolution, and force generation. The max-
imum plunging frequency was 4 Hz because of the higher risks of breaking the flexible

TABLE 1: Mean thrust force coefficients and power


efficiency for flexible hydrofoil
Frequency (Hz) 1.5 2.0 2.5 3.5
Re 1655 2207 2758 3861
CT AV E 0.203 0.470 0.516 0.372
CT 0.032 0.079 0.088 0.134

Journal of Flow Visualization & Image Processing


Experimental Investigation of 2D Flexible Plunging Hydrofoil 259

hydrofoil at higher frequencies; high order deformation of the hydrofoil was therefore
not achieved in this study.
The effect of the flexibility in the vortex development and force generation was dis-
cussed by comparing the results obtained for the two models. It was found that the vortic-
ity generated by the flexible hydrofoil was much stronger and slower in detaching from
the leading and trailing edges of the model, especially at greater plunging frequencies.
Furthermore, the jet induced in front of the leading edge was much weaker than the one
in the trailing edge; a thrust force was generated as a result of this. The rigid hydrofoil
generated LEVs and TEVs of similar strength and size; slight differences were caused
by its nonsymmetric geometry. The induced jets in the rigid model were also similar in
strength, and thus generated small amounts of thrust.
The thrust and lift force measurements of the two hydrofoils were acquired at var-
ious frequencies. It was found that, within the range of our experiment frequency, the
nondimensional lift/plunging force in the rigid hydrofoil model was independent of the
plunging frequency. For the flexible hydrofoil, the high degrees of deformation greatly
decreased the plunging force generation.
The results also showed that the flexible hydrofoil was more efficient in thrust force
generation at higher plunging frequencies. Additionally, there existed an optimal fre-
quency for maximum thrust force coefficient, which was related to the hydrofoil geome-
try, stiffness, and fluid properties. Compared to Lai’s report (Lai and Platzer, 2009), PIV
measurements in current research offered flow field information for the pure plunging
hydrofoil and explained the mechanism for thrust force generation. On the other hand,
the force generation for Jardin’s model (Jardin et al., 2009) was calculated from the flow
field using the momentum equation, where the integration for the pressure term might
induce relatively large uncertainties for the lift and drag forces. Therefore, the direct
force measurements in current research provided accurate dynamic force measurements
for the plunging hydrofoil. In addition, the effects of flexibility on the flow field and
dynamic dynamic force were studied in detail for the purely plunging hydrofoil.
There are some limitations in the present study. The first was about PIV measure-
ments for the flow field surrounding the rigid hydrofoil. Due to the blockage of laser
sheet, the flow field measurements above the rigid hydrofoil was not available. Fortu-
nately, it was reasonable to assume that the flow field above the hydrofoil was periodic
and developed in the same manner as that in the lower side because of the symmetric
hydrofoil and plunging motion. Another limitation was that the imperfect plunging mo-
tion slightly influenced the thrust force measurement, which resulted in a relatively high
uncertainty in thrust force.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the support from Army Research Laboratory through
Army High Performance Computing Research Center (AHPCRC).

Volume 20, Number 4, 2013


260 Tian et al.

REFERENCES
Birch, J. M. and Dickinson, M. H., The influence of wingake interactions on the production of
aerodynamic forces in flapping flight, J. Exp. Biol., vol. 206, pp. 2257–2272, 2003.
DeLuca, A. M., Reeder, M. F., Freeman, J., and Ol, M. V., Flexible- and rigid-wing micro air
vehicle: Lift and drag comparison, J. Aircraft, vol. 43, no. 2, pp. 572–575, 2006.
Dickinson, M. and Götz, K., Unsteady aerodynamic performance of model wings at low
Reynolds numbers, J. Exp. Biol., vol. 174, pp. 45–64, 1993.
Dickinson, M., Lehmann, F.-O., and Sane, S., Wing rotation and the aerodynamic basis of insect
flight, Science, vol. 284, pp. 1954–1960, 1999.
Dong, H., Harff, M., and Mittal, R., Vortex structures and aerodynamic performance of finite-
aspect-ratio flapping wings in hovering motion, 37th AIAA Fluid Dynamics Conference and
Exhibit, 2007.
Heathcote, S. and Gursul, I., Flexible flapping airfoil propulsion at low Reynolds numbers,
AIAA-2005-1405, 2005.
Jardin, T., David, L., and Farcy, A., Characterization of vortical structures and loads based on
time-resolved PIV for asymmetric hovering flapping flight, Exp. Fluids, vol. 46, pp. 847–857,
2009.
Koochesfahani, M. M., Vortical patterns in the wake of an oscillating airfoil, AIAA J., vol. 27,
no. 9, pp. 1200–1205, 1989.
Lai, J. C. S. and Platzer, M. F., Characteristics of a plunging airfoil at zero freestream velocity,
AIAA J., vol. 39, no. 3, pp. 531–534, 2009.
Lewin, G. C. and Haj-Hariri, H., Modelling thrust generation of a two-dimensional heaving air-
foil in a viscous flow, J. Fluid Mech., vol. 492, pp. 339–362, 2003.
Sedov, L. I., Two-Dimensional Problems in Hydrodynamics and Aerodynamics, Moscow-
Leningrad: A division of John Wiley Sons Inc., 1965.
Shyy, W., Lian, Y., Tang, J., Liu, H., Trizila, P., and Standford, B., Computational aerodynamics
of low Reynolds number plunging, pitching and flexible wings for MAV applications, 46th
AIAA Aerospace Science Meeting and Exhibit, 2008.
Sun, M. and Wu, J. H., Aerodynamic force generation and power requirement in forward flight
in a fruit fly with modeled wing motion, J. Exp. Biol., vol. 206, pp. 3065–3083, 2003.
Wang, Z. J., Vortex shedding and frequency selection in flapping flight, J. Fluid Mech., vol. 410,
pp. 323–341, 2000.
Yang, T., Wei, M., and Zhao, H., Numerical study of flexible flapping wing propulsion, AIAA J.,
vol. 48, no. 12, pp. 2909–2915, 2010.

Journal of Flow Visualization & Image Processing

View publication stats

You might also like