Você está na página 1de 8

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/259079469

Aerodynamic Characteristics of a Flexible Wing


MAV
Article January 2012

CITATIONS

READS

192

1 author:
Roy Tan Zhenyang
Nanyang Technological University
2 PUBLICATIONS 0 CITATIONS
SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,


letting you access and read them immediately.

Available from: Roy Tan Zhenyang


Retrieved on: 12 September 2016

Proceedings of the URECA@NTU 2011-12

Aerodynamic Characteristics of a Flexible Wing MAV


Tan Zhenyang
School of Mechanical and Aerospace
Engineering

Assistant Professor Yongki Go Tiauw Hiong


School of Mechanical and Aerospace
Engineering
Dr Adnan Maqsood

Abstract - The aerodynamic characteristics of


the flexible-wing Micro Air Vehicle (MAV) is
the main focus of this research study. Flexiblewing MAV is renowned for improved lift
characteristics and significantly smooth flight
in gusty conditions than its rigid-wing
counterpart. Various configurations were
tested in the wind tunnel to determine the ways
of further enhancing lift and increasing
stability during flight. The effects of
underlying skeleton are quantized and
compared with the rigid-wing and the flexible
wing without battens. The baseline geometric
dimensions for all MAVs include square crosssection of side 15 cm each with thickness 3mm
and aspect ratio of 1. Results show that the
orientation and location of skeleton layout
significantly govern the lift characteristics and
flight stability. Experiments were conducted at
speeds ranging from 4m/s to 14m/s but only
speed of 6m/s was discussed in this paper. The
aerodynamic assessment has demonstrated
significant improvement of lift characteristics
and flight stability for flexible wings over rigid
wings counterpart in low Reynolds number
flight envelope.

designs, and miniaturized components are


enabling many rapid advances.
The need to better understand fundamental
aspects of flight for small vehicles has
spawned a surge in high quality research in the
area of micro air vehicle, or MAV research.
These aircraft have a set of constraints which
are, in many ways, considerably different from
that of traditional aircraft and are often best
addressed by a multidisciplinary approach.
Fast-response non-linear controls, nanostructures, integrated propulsion and lift
mechanisms, highly flexible structures and
low Reynolds aerodynamics are just a few of
the important considerations which may be
combined in the execution of MAV research.
This study focuses on using flexible wing
membranes and varying underlying skeleton
frames to determine the aerodynamic
characteristics of a flexible wing of MAV.
One advantage of using flexible wings is that
typical MAVs operate at low Reynolds
numbers. The range of Reynolds numbers at
which MAVs fly is similar to that of an insect
or bird (103 - 105). Thus, some researchers
think that understanding bird flight or insect
flight will be useful to designing MAVs. A
new trend in the MAV community is to take
inspiration from flying insects or birds to
achieve unprecedented flight capabilities.
However, such low Reynolds numbers
decrease the lift-to-drag ratio that result in
poor aerodynamic performance. Moreover,
because of the size constraints, MAV designs
belong to the class of low aspect ratio wings.
These factors significantly degrade the
aerodynamic efficiency of MAVs.
Another advantage is that previous studies
have shown that flexible-wing MAVs perform
better than the conventional rigid-wing MAVs
in terms of improved aerodynamic efficiency

Keywords Flexible wing, MAV, Windtunnel testing


1 INTRODUCTION
Micro Air Vehicles (MAVs) is a class of
unmanned aerial vehicles (UAV) that has a
size restriction and may be autonomous.
Modern craft can be as small as 15 cm, with
insect-sized aircraft reportedly expected in the
future. Driven by applications ranging from
civil search-and-rescue missions to military
surveillance missions, there is a rising level of
interest and investment in better vehicle

Proceedings of the URECA@NTU 2011-12

and gust suppression. The wind gradients vary


significantly at low altitudes, with speeds
reaching up to 6m/s. Rigid wing MAVs are
known for poor gust performance thereby
decreasing there maneuver potential and
steady flight in such environments.
Flexible wing MAVs [1-3] can alleviate gust
effects by adaptive washout mechanism. The
adaptive washout in flexible wings is
produced by the deformation of the wing skin.
The flexible wing shape change is a function
of airspeed and angle of attack. When there is
an increase in airspeed, due to a head-on wind
gust, there will be a change in the wing shape.
This change brings a decrease in lift
efficiency. However, the overall lift remains
same due to the increase in airspeed. When the
wind gust dies, the flexible membrane will
revert back to its original shape, thereby
maintaining the same lift. Ideally, the flexible
wing MAV will fly smoothly in a turbulent
atmosphere by maintaining constant lift.
On the other hand, studies have shown that
flexible wings typically have lower lift-to-drag
ratio than the conventional rigid wings. This is
attributed to membrane deformation that
results in lift reduction and increased drag.
In this work, experiments are performed on
various configurations of flexible-wing MAV.
Specifically, wind-tunnel testing of low-aspect
ratio wings at low Reynolds number is
conducted. The test free-stream velocity
ranges from 4m/s to 14 m/s. The effect of
batten arrangements for various frame
configurations on lift characteristics and
moments are quantized and discussed.

2 EXPERIMENTAL SETUP AND


PROCEDURES

frame on the lift and pitching moment. Each


configuration of underlying skeleton frame
differs in terms of number and orientation of
battens. The type of batten in each
configuration is crucial as it affects the
deformation pattern of the wing membrane.
The following Fig. 1 shows the layout of the
nine models, each with three different
configurations, that is, (i) with both trailing
edge and side edges; (ii) without trailing edge;
(iii) without both trailing edge and side edges.
(i)

(ii)

(iii)

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Model 7

Model 8

2.1 MODEL FABRICATION


The primary evaluation and analysis is based
on the data obtained from the experiments
conducted using the wind tunnel. The baseline
wing geometry consists of a square frame of
15 x 15 cm, resulting in aspect ratio of 1 for
the
wings.A
total
of
twenty-seven
configurations were fabricated to determine
the effects of different underlying skeleton

Model 9

Rigid Model

Figure 1: Wing frame design configurations

Proceedings of the URECA@NTU 2011-12

The wing skeleton frames are fabricated from


3mm thick aluminum plates. The plates are
machined by using CNC Laser cutting
machine instead of manually. This is because
the plates are too thin to experience the cracks
that may result from the larger induced stress
during manual machining. There is a central
rectangular part that is present in all
configurations as it is required to facilitate the
mounting of the electronics and battery in an
actual prototype. The frames are all single
pieced, thus, no glue or epoxy is required as
opposed to using individual carbon fibre rods
to fabricate the frames. (The original plan was
to use carbon fibre rods with epoxy to
construct the frames. However, it was aborted
in the process as the drag by the carbon frames
was too high.) The rigid wing (which is the
reference wing) is also made from similar
3mm thick aluminium sheet and its side
dimensions are identical to the flexible wing
frames. All the twenty-eight wing frames are
rigid and the membranes of the wings are
made of the same material. The membrane
used in this study is elastic latex obtained from
sixteen-inch large balloons. The materials
used are shown in Fig. 2.

Figure 2: Materials used in the experiment, starting


from left: 16-inch balloons and 3mm thick aluminium
sheet

For the elastic latex, the extension or


relaxation is kept constant as it is not the main
focus of this study. Since the square
dimension of the wing is 15 cm, the material is
cut in square dimension close to 15cm and
subsequently pasted onto the wing area using
epoxy or superglue.

open loop wind-tunnel. The test section is 2 m


long and internal cross-section is 0.72 m by
0.78 m. It uses a six component internal
balance to measure all forces and moments.
The model positioning system is of quadrant
type and is equipped with a sting model
support. The data acquisition system is based
on National Instruments (NI) platform and
Lab-View based software to graphically view
and record the data. It is known as Data
Acquisition, Reduction and Control System
(DARCS). The wind tunnel test section is
shown in Fig. 3.

Figure 3: Open-circuit wind tunnel test section with


sting balance

After conducting the experiments, further


calibration was required as the forces
registered by the load cell were not the true lift
and drag but were actually forces in the z-axis
and x-axis respectively as shown in Fig. 4. In
addition, it was discovered that the load cell
was poorly calibrated and the force values
registered were inaccurate.

Figure 4: Load cell axes

2.2 WING TUNNEL FACILITY


The wind-tunnel experimentation is carried at
Nanyang Technological University (NTU).
The facility is a low speed, low turbulence

Thus, a calibration was conducted to account


for the errors due to poor calibration of the
load cell. Weights varying from 0.1 kg to
1.5kg were loaded in the vertical direction
(lift-axis) and the corresponding force values

Proceedings of the URECA@NTU 2011-12

registered in the x and z-axes were noted. This


process was repeated for all the angles of
attack used in the experiments. The apparatus
setup is shown in Fig. 5.
Figure 6: Three configurations of model 5 and rigid
model

Figure 5: Calibration in the lift axis

3 RESULTS AND DISCUSSION


The experimental results presented in this
paper had all been corrected for wind-tunnel
blockage effects (solid blockage, wake
blockage and streamline blockage) from
Barlow[4]. As the angle of attack increased,
the correction factor for blockage effect also
increased. However, the maximum blockage
correction for most of the scenarios was less
than 10%. Support interference due to the
mechanical supports had little effects on the
aerodynamic forces and moments. Also, no
hysteresis was observed for all the wings
during the experiments.

After applying the necessary wind tunnel


corrections, the experimental data is tabulated.
The objective is to find the best flexible wing
configuration within each model that gives
rise to the best overall lift over a range of
angle of attack from 0 to 18 and over a range
of speeds from 4m/s to 14m/s. At the same
time, the leading edge moment was also
recorded for further analysis. The lift
coefficients of the three wing configurations
and the rigid wing at wind speed of 6m/s are
plotted in Fig. 7. It can be clearly seen that the
flexible model 5 wings generally have higher
lift than the rigid wing. There is no significant
difference in the amount of lift generated by
the three configurations of model 5 over the
angle of attack 0 to 14. However, after 14,
the lift for model 5(i) and 5(ii) are
significantly higher than that for model 5(iii).
It may be due to the flapping of the flexible
wing at the trailing edge which more irregular
camber and earlier separated flow.

3.1 EFFECTS OF ABSENCE OF


TRAILING EDGE AND SIDE EDGES
In order to observe the effects associated with
trailing edge and side edges, model 5 was
chosen as the model to be studied. As shown
in Figure 1 previously, model 5(i) has both
trailing edge and side edges; model 5(ii) has
side edges only; and model 5(iii) has no
trailing edge and side edges. Fig. 6 shows all
the three different configurations for model 5.
Figure 7: Coefficient of lift comparison for model 5(i),
(ii), (iii) and rigid model at wind speed of 6m/s

Next, the leading edge moment coefficients of


the three wing configurations and the rigid
wing at wind speed of 6m/s are plotted in Fig.
8. It can also be clearly seen that the flexible
model 5 wings generally have larger leading

Proceedings of the URECA@NTU 2011-12

edge moment than the rigid wing. Lift and


leading edge moment are interrelated. When a
higher lift is generated, the pitching moment
will be larger (meaning more negative), thus
the leading edge moment is also larger. This
means model 5 wings are more generally more
stable than rigid wings. There is no significant
difference in the amount of leading edge
moment on the the three configurations of
model 5 over the angle of attack 0 to 18.

highest lift. This is because the flexible


membrane is able to deform more than the other
wings, thus providing the wing with additional
camber which in turn results in higher lift.
Another observation is that the lift produced by
model 2(i) and 3(i) is very close for all the
measured angles of attack. Therefore, it can be
inferred that the additional two battens in model
3(i) do not have significant effect on the lift
produced. For leading edge moment, model 1(i)
is also the largest among these wings, implying it
is the more stable. Both models 2(i) and 3(i) have
very similar moments for all the tested angles of
attack, implying that the additional two battens on
model 3(i) do not have any significant effect on
the moment as well.

Figure 8: Coefficient of leading edge moment


comparison for model 5(i), (ii), (iii) and rigid model at
wind speed of 6m/s

3.2 EFFECT OF SKELETON


STRUCTURE (AXIAL BATTENS)
To study the effect of skeleton structure,
specifically axial battens, models 1, 2 and 3 are
chosen for this study. These results are also
compared with that of rigid model and the lift
comparison and leading edge moment is plotted
in Fig. 9 and Fig 10 respectively.

Figure 9: Coefficient of lift comparison for model 1(i),


model 2(i), model 3(i) and rigid model at wind speed of
6m/s

It is observed that model 1(i), representing the


flexible wing without any battens, exhibits the

Figure 10: Coefficient of leading edge moment


comparison for model 1(i), model 2(i), model 3(i) and
rigid model at wind speed of 6m/s

3.3 EFFECT OF SKELETON


STRUCTURE (DIAGONAL
BATTENS)
To study the effect of skeleton structure,
specifically diagonal battens, models 4, 5, 6, 7, 8
and 9 are chosen for this study. These results are
also compared with that of rigid model and model
1(i) and the lift comparison and leading edge
moment is plotted in Fig. 11, 12 and Fig 13, 14
respectively. In Fig 11, models 4(i), 6(i) and 8(i)
are grouped together as the only difference in
these 3 models is the longer diagonal batten.
Therefore, the results can be compared and
accounted for in a systematic manner.
Firstly, by comparing models 1(i) and 6(i), it can
be observed that the additional shorter batten has
decreased the overall lift of the flexible wing
significantly. Further addition of a longer batten

Proceedings of the URECA@NTU 2011-12

at either location in models 4(i) and 8(i) causes


the lift of the flexible wing to decrease more. This
is because the membrane is able to deform to a
lesser extent after addition of another batten. As
for models 5(i) and 7(i), the addition of a
diagonal batten has decreased lift for both wings
but model 5(i), with a longer batten, has
experienced a smaller decrease in lift. Next,
comparing models 5(i) and 7(i) with model 9(i),
the addition of each others batten has caused the
lift to decrease significantly. An interesting thing
to note is that even though the batten in model
7(i) is smaller than the batten in model 5(i), it has
resulted in a larger decrease in lift. Therefore, it
can be inferred that the location of the batten is
also crucial in affecting the overall lift of the
flexible wing.

Figure 11: Coefficient of lift comparison for models


1(i), 4(i), 6(i), 8(i) and rigid model at wind speed of
6m/s

diagonal battens do not have much effect on


the leading edge moment.

Figure 12: Coefficient of leading edge moment


comparison for model 1(i), 4(i), 6(i), 8(i) and rigid
model at wind speed of 6m/s

Figure 13: Coefficient of leading edge moment


comparison for model 1(i), 5(i), 7(i), 9(i) and rigid
model at wind speed of 6m/s

4 CONCLUSION AND FUTURE WORK

Figure 12: Coefficient of lift comparison for model 1(i),


5(i), 7(i), 9(i) and rigid model at wind speed of 6m/s

The leading edge moments of all the flexible


wings are significantly higher than of the rigid
wing. Besides that, the magnitudes of the
moment of the flexible wings are very close to
each other, implying that the addition of

Open-loop wind tunnel testing was conducted


on the various configurations of flexible-wing
used in MAVs. The parameters that were
varied in this study are speed of airflow in
wind tunnel testing and the batten
arrangement. During the analysis of results, it
was observed that all the flexible wings show
higher lift than the rigid wing with model 1(i)
(without any battens) generating the most lift
among all the models. Next, all flexible wings
have higher leading edge moment than the
rigid model, thus implying that flexible wings
are more stable in flight. Also, it is noted that
besides the amount of area covered by the
battens, the location of the battens also play an
important role in determining the amount of
lift generated by the flexible wing. Next, it is
observed that at low angles of attack, the

Proceedings of the URECA@NTU 2011-12

aerodynamic characteristics of the flexible


wings are almost similar with very little
differences. Thus, for future work, the
experiments can be conducted at much higher
angles of attack so that the stall characteristics
of the models can be investigated. Also, more
configurations can be tested further to find the
optimum configuration for lift and flight
stability.
ACKNOWLEDGMENT
We wish to acknowledge the funding support
for this project from Nanyang Technological
University under the Undergraduate Research
Experience
on
Campus
(URECA)
programme.
REFERENCES

[1]

Ifju, P., D.A. Jenkins, S. Ettinger, Y. Lian, W.


Shyy, and M.R. Waszak, Flexible-Wing based
Micro Air Vehicles, in 40th AIAA Aerospace
Sciences Meeting and Exhibit. 2002, AIAA:
Reno, Nevada, USA.

[2]

Ifju, P., B. Stanford, M. Sytsma, and R.


Albertani, Analysis of Flexible Wing Micro Air
Vehicle, in 25th AIAA Aerodynamic
Measurement Technology and Ground Testing
Conference. 2006, AIAA: San Francisco,
California, USA.

[3]

Deluca, A.M., M.F. Reeder, M.V. Ol, J.


Freeman, I. Bautista, and M. Simonich,
Experimental Investigation into the
Aerodynamic Properties of a Flexible and
Rigid Wing Micro Air Vehicle, in 24th AIAA
Aerodynamic Measurement Technology and
Ground Testing Conference. 2004: Portland,
Oregon, USA.

[4]

Barlow, J.B., Low-speed wind tunnel testing. 3


ed. 1999: Wiley.

[5]

Z.Zeng, M. Adnan, T.H. Go, Flexible Wing


Aerodynamics at Low Reynolds Numbers, in
28th International Congress of the
Aeronautical Sciences, ICAS 2012

Você também pode gostar