Aerospace Science and Technology 10 (2006) 111119

www.elsevier.com/locate/aescte

The thrust and lift of an ornithopters membrane wings

with simple flapping motion
Che-Shu Lin, Chyanbin Hwu, Wen-Bin Young
Department of Aeronautics and Astronautics, National Cheng Kung University, No. 1, DarShei Rd., Tainan, 70101, Taiwan, ROC
Received 22 August 2005; received in revised form 7 October 2005; accepted 7 October 2005
Available online 1 December 2005

Abstract
Human beings flying with the help of aircrafts of various kinds have been able to fly for about one century. Although the flapping wings of
animals served as an inspiration to pioneers of human flight, we dont really understand how they work. In this study, we employ the concept of
four-bar linkage to design a flapping mechanism which simulates a flapping motion of a bird. Wind tunnel tests were performed to measure the
lift and thrust of the mechanical membrane flapping wing under different frequency, speed, and angle of attack. It is observed that the flexibility
of the wing structure will affect the thrust and lift force due to its deformation at high flapping frequency. The lift force will increase with the
increase of the flapping frequency under the corresponding flying speed. For the same flapping frequency, the flying speed can be increased by
decrease of the angle of attack with the trade of loosing some lift force. An angle of attack is necessary in a simple flapping motion in order to
derive a lift force. The flapping motion generates the thrust to acquire the flying speed. The flying speed and angle of attack combine to generate
the lift force for flying.
2005 Elsevier SAS. All rights reserved.
Keywords: Ornithopter; Flapping wing; Membrane wing; Flexible wing

1. Introduction
It has been an interesting subject in biology to study the flying mechanism and characteristics of flapping flight of birds [3
8,11]. Flapping flight is more complicated than the fixed-wing
flight. For an aircraft with fixed wings, only forward motion
is necessary to sustain the body with the induced aerodynamic
lift. For the flapping flight, the wing not only has a forward
motion but also the up and down flapping. In the down stroke
flapping, the wing is fully extended and produces both lift and
thrust at the same time. During the up stroke, some part of the
wing is folded to reduce the moment of inertia and the drags
of the wings. The wings are also twisted during the flapping to
vary the angle of attack for various flying motions. During the
hovering, fast forward motion, or slow forward motion, different wing strokes and attack angles are employed. The flapping
* Corresponding author. Tel.: (+886) 6 2757575 ext. 63672; fax: (+886) 6
2389940.
E-mail address: youngwb@mail.ncku.edu.tw (W.-B. Young).

1270-9638/$ see front matter 2005 Elsevier SAS. All rights reserved.
doi:10.1016/j.ast.2005.10.003

wing flight is also believed to have better maneuverability compared to the fixed wing flight.
In biological flight, the wings not only move forward relative
to the air, they also couple the motions of flapping up and down,
plunge and sweeping [9,10]. In order to attain efficient lift during the flapping motion, the wing will undergo a twist motion
to change the angle of incident in the down and up strokes. In
general, the down stroke will produce the most lift and thrust
with fully extended wings. At the up stroke, additional lift can
also be generated by twisting the wing to change the attack angle, especially during the take-off and hovering flight [2]. For
small insect flight, the wings flap at high frequency and the unsteady flow mechanism dominates [12]. During the up stroke,
the wing surfaces press together at the end for a period of time.
As the wings separate for the next down stroke motion, they rotate around the trailing edges first to form a V shape before they
begin to move away. This leads to the large circulation and lift
on the wing without the negatives of vortex shedding.
The flapping-wing aircraft configurations (ornithopter) received great attention to the researchers for a long time. One
application of the ornithopter, being a micro air vehicle (MAV),

112

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

has recently attracted much interest. The primary purpose of the

MAV currently is surveillance, which requires a slow flight and
excellent maneuverability. Early models of rubber powered ornithopters were successful built by Alphonse Penaud [1]. This
has been the prototype for many ornithopter models later built
as toys. Despite of the flapping wing design, the aerodynamic
behaviors of these ornithopters are quite different to those of
birds in the cruising flight. The flapping wing of Penauds ornithopter is constructed by a stiff leading-edge spar and some
flexible ribs attached to the spar. The wing skeleton is covered
with a thin sheet of flexible material. This design results in a
light flexible membrane wing that may twist and bend during
the flapping motion. This study constructs mechanical membrane flapping wings using carbon fiber reinforced epoxy as the
skeleton. A four-bar linkage driven by an electric motor was
used as the flapping mechanism. The wing surface was formed
by covering the skeleton with thin plastic film. Wind tunnel tests
were performed to measure the lift and thrust of the mechanical
membrane flapping wing under different frequency, speed, and
angle of attack.
2. Wing design and kinematics

4 =

b sin
2 ,
d sin

and
= 2 4 ,

(7)

= 4 ,

(8)

= 2 + ,


1 b
= sin
sin 2 ,
s
 2
2
2
1 c + s d
= cos
,
2cs


c
= sin1
sin ,
d
4 = 360 ( + ),

(9)

OA + AB + BD > OD,

(1)

OA + OD + BD > AB,

(2)

OA + AB BD < OD,

(3)

AB OA + BD > OD.

(4)

If the angular speed of the driving linkage is 2 , the angular

velocities of the connecting and following linkages can be expressed as
3 =

b sin
2 ,
c sin

Fig. 1. The four-bar linkage design for the wing mechanism.

(10)
(11)
(12)
(13)

where 4 is the angular speed of the following linkage, 2 and

4 are the position angles of the driving and following linkages, and the other lengths and angles are defined in Fig. 1. The
corresponding angular accelerations can be derived by differentiating with respect to time to have
b22 cos + c32 cos + d42
,
c sin
b22 cos + c32 + d42 cos
4 =
.
d sin

3 =
A four-bar linkage design was selected as the mechanism
for the wing flapping as shown in Fig. 1. The linkage OA is the
driving linkage and the linkage BD is connected to the wing
structure. The design of the four-bar linkage is a kind of the
Crank and Rocker mechanism where the driving linkage can
rotate a revolution completely and the following linkage undergoes a rocking motion. The Crank and Rocker mechanism
follows some constraints for the member lengths as

(6)

(14)
(15)

3. Experimental
The selected lengths for each linkages are OA = 8.5 mm,
AB = 31.5 mm, BD = 19 mm, and OD = 30.5 mm. The resulting flapping angle for the wing is shown in Fig. 2. The upward
flapping angle is about 39 degree and the downward flapping is
about 18 degree. In the early stage of the design, several tests
were performed to investigate the effect of the flapping upward
and downward strokes on the lift force. It was found that large
upward angle stroke or larger downward angle stroke do not affect the lift force. In this design, we have a larger upward angle
stroke.

(5)

Fig. 2. Flapping angle with respect to the position angle of the driving linkage.

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

(a)

113

(b)

Fig. 3. (a) Angular flapping velocity and (b) angular flapping acceleration with respect to the position angle of the driving linkage.

Fig. 4. Assembly of the mechanism, motor and the battery set.

Fig. 3 shows the corresponding angular velocity and acceleration of the following linkage. The rotational speed of the
motor is assumed to be constant in the analysis. From the velocity and acceleration, it can be assured that the mechanism
can operate smoothly without obvious jumps or jerks. The linkages were milled from a titanium alloy plate. Together with the
driving motor, the entire assembly of the flapping unit is shown
in Fig. 4.
It is expected that the wing area will affect the lift force. In
order to study the effect of the wing area on the lift force of a
flapping wing, two different wings were constructed for the test
of the lift force and thrust. The wing aspect ratio is about 2. The
two wings with span in 60 cm and 40 cm will be referred as
type A and B respectively. Types A and B were constructed by
epoxy reinforced carbon fiber composite frames and covered
with a PVC plastic film as shown in Fig. 5. The weights and
surface areas of the wings are 17.64 g, 0.054 m2 for type A
wing and 27.05 g, 0.102 m2 for type B wing.

In the wind tunnel test of the flapping wing, four different attack angles were used, 0, 5, 10, 15 degrees. The angle of attack
is formed by attached the wing to the driving mechanism with
a specific angle. The angle of attack is defined corresponding
to the wind-axes reference frame. The different wind velocities
used for the test are 0 to 20 km/hr in a step increase of 5 km/hr.
Load cells, MA-SM2, were mounted on the base of the driving
unit to measure the lift force or thrust on the wind tunnel test.
The capacity of the load cell is 3 kg with an excitation voltage
10 V. Standard weights were used to calibrate the load cell before tests. The signal from the load cell was recorded directly
by a Tektronix oscilloscope, TDS 1002. The recording duration
is 4 seconds with 2500 digital points that can be downloaded
to a personal computer after the test. Before a test, the initial
weight of the entire assembly above the load cell was recorded.
As the wings start to flap, the oscilloscope will record the load
variation during the preset duration. The lift during the flapping
motion can be derived by subtracting the measured load from
the initial weight. The average lift can be calculated by average

114

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

(a)

(b)

Fig. 5. Different wing constructions: (a) type A, (b) type B.

Fig. 6. Setup of the flapping wings for lift and thrust tests.

the lift along the measured time span. For the measurement of
thrust, the data can be collected by direct reading from the load
cell. The lift and thrust were measured with respect to the windaxes reference frame. The setups of the flapping wings for lift
and thrust tests are shown in Fig. 6.

4. Results and discussions

The lift forces for different flapping frequencies and wind
speeds are shown in Fig. 7 using the wing type A. The figure also shows the lift forces for different angles of attack. It

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

(a)

115

(b)

Fig. 7. The lift forces under different flapping frequencies and wind speeds at (a) angle of attack = 0, (b) angle of attack = 15 degree using the type A wing.

(a)

(b)

Fig. 8. The lift forces under different flapping frequencies and wind speeds at an angle of attack = 15 degree using (a) the type A wing and (b) the type B wing.

(a)

(b)

Fig. 9. The lift forces under different attack angles and wind speeds using (a) type A wing at 5 Hz and (b) type B wing at 7 Hz.

is shown that there is no lift force without the relative wind

speed. Although the flapping angles are different in upstroke
and down stroke, the upward force and downward force are
about the same during the entire flapping cycle and tend to
cancel each other in this situation. Increase of the flapping frequency does not induce any lift force also. With a relative wind
speed, a lift force can be generated and higher lift force is induced in a higher angle of attack. It is demonstrated that the
relative wind speed and the angle of attack are the major factors for the lift force of an ornithopter under a simple flapping
motion. For 15 degree of angle of attack, the lift force increases
with the increase of flapping frequency up to about 5 Hz. Notice
that for zero angle of attack, the lift will increase with the wind
speed in Fig. 7. Actually, in installation of the wing mechanism,
the leading edge of the wing is set to parallel to the wind direction. However, due to the weight of the wing itself, slight sag

of the trailing edge will occur, leading to certain lift readings in

the test.
The lift force is also shown to increase with respect to the
wind speed and flapping frequency in the wind tunnel tests as
in Fig. 8. The angle of attack for the wing under tests is set to
15 degree. Most of the data for type A wing is measured at the
frequency below 5 Hz. Due to the flexibility of type A wing, the
second mode deformation is induced for a higher flapping frequency. Under this situation of second mode deformation, the
lift force is saturated or even decreases. The interesting thing
is that the lift force for type B wing is about the same as type
A wing for the same flapping frequency. Notice that the wing
span for type B wing is about 40 cm while it is 60 cm for type
A wing. The lift force is not directly proportional to the wing
area for a flapping wing under the same flapping frequency and
wind speed. The possible cause will be that the flexibilities of

116

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

(a)

(b)

Fig. 10. The thrusts under different flapping frequencies and wind speeds at an angle of attack = 15 degree using (a) the type A wing and (b) the type B wing.

(a)

(b)

Fig. 11. The lift forces generated by different flapping frequencies and angles of attack using the type A wing.

(a)

(b)

Fig. 12. The lift forces generated by different flapping frequencies and angles of attack using the type B wing.

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

117

(a)

(b)
Fig. 13. Schematic diagram of the lift and drag forces near the root of the wing for the up and down strokes. (a) Angle of attack = 0; (b) angle of attack > 0.

the two wings are different. Large deformation along the wing
span will disturb the air flow, causing the reduction of the lift
force. At a higher frequency, the deformation of type A wing
becomes larger and the lift force curve with respect to the flapping frequency starts to level off as in Fig. 8.
The effect of wind speed on the lift force is shown in Fig. 9.
The lift force increases substantially at the high wind speed.
That means flying at higher speed is a more effective way in obtaining the required lift force rather than increasing the flapping
frequency. In the experimental wind tunnel tests, the wind speed
and flapping frequency are kept as separated independent variables. However, in the actual flying, the flying speed that can be
reached depends on the flapping frequency of the wings. In order to increase the flying speed, the flapping frequency must be
increased also.
The flying speed depends on the thrust of the flapping wings.
During the flying motion, the generated thrust can overcome the

drags and accelerate the ornithopter. In the wind tunnel tests, the
wind speed is set by the wind tunnel. However, the actual flying
speed that an ornithopter can reach depends on the thrust and
the drag. The drag force is a function of the flying speed and
the angle of attack. Fig. 10 shows the net thrust in a wind tunnel test under different flapping frequency and angle of attack =
15 degree. The curve with diamond symbols is corresponding
to thrust force for flapping with no wind. The thrust force becomes saturated at higher frequency for type A wing as shown
in Fig. 10(a). This is suspected to be caused by the large deformation of the wing shape during high frequency flapping. The
data shows that the net thrust may become negative in some
wind speeds and frequencies. The negative net thrust means that
the flapping motion at the specified frequency can not reach the
preset speed under this angle of attack. Since the type B wing
is smaller, most of the net thrust data for type B wing is at the
negative region. The largest obtainable speed is about 10 km/hr

118

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

(a)

(b)
Fig. 14. Schematic diagram of the lift and drag forces near the tip of the wing for the up and down strokes. (a) Angle of attack = 0; (b) angle of attack > 0.

at this angle attack. Notice that the data (zero net thrust force)
derived in this study is based on the measurement of the flapping mechanism only. The actual drag in the ornithopter must
include the entire body and the tail structures.
The flying speeds for different flapping frequencies can be
constructed based on Fig. 10. For each flapping frequency, the
wind speed approaching the zero net thrust force can be considered as the corresponding flying speed. The relations are shown
in Fig. 11(a) for different angles of attack for type A wing. With
specified frequency and wind speed, the corresponding lift force

can be derived from the previous experimental data. The lift

force for different frequencies and angles of attack is shown in
Fig. 11(b). For the same frequency, the flying speed is lower
for high angle of attack, but the lift force is higher. The lift
force and flying speed are shown to increase with the flapping
frequency. From the experiments, it was observed that the maximum flapping frequency was limited by the wing structure. As
the frequency is high enough to induce the second mode wing
deformation, the wing performance will decrease. It should be
noticed that the limitation of the driving system is not consid-

C.-S. Lin et al. / Aerospace Science and Technology 10 (2006) 111119

ered in the above discussion. The corresponding flying speed

and lift force for type B wing are shown in Fig. 12.
5. Conclusions
This study constructs mechanical membrane flapping wings
using carbon fiber reinforced epoxy as the skeleton. A fourbar linkage driven by an electric motor was used as the flapping mechanism. The wing surface was formed by covering
the composite skeleton with thin plastic film. Wind tunnel tests
were performed to measure the lift and thrust of the mechanical membrane flapping wing under different frequency, speed,
and angle of attack. It is observed that the flexibility of the wing
structure will affect the thrust and lift force due to its deformation at high flapping frequency. The lift force will increase with
the increase of the flapping frequency under the corresponding
flying speed. For the same flapping frequency, the flying speed
can be increased by decrease of the angle of attack with the
trade of loosing some lift force as shown in Figs. 11 and 12. The
mechanism of the thrust and lift forces generated by the simple
flapping motion are also shown in Figs. 13 and 14 schematically. For the wing cross section near the root, the speed of the
wing is low. The corresponding speed is most due to the flying speed, resulting in more lift force due to the aerodynamic
effect. Notice that the trailing edge of the wing section will deform up and down during the duty flapping cycle because of its
flexibility. For wing section far from the root, the flapping speed
is higher and comparable to the wind speed. For zero angle of
attack, the lift forces during the up and down strokes will cancel each other. An angle of attack is necessary in this situation
in order to derive a lift force. The flapping motion generates the
thrust to acquire the flying speed. The flying speed and angle of
attack combines to generate the lift force for flying.

119

Acknowledgements
The authors would like to thank for the financial support
from National Science Council in Republic of China under the
contract number of NSC 93-2212-E006-067.
References
[1] J.D. Delaurier, An ornithopter wing design, Canadian Aeronautics Space
J. 40 (1) (1994) 1018.
[2] T. Hedrick, B. Tobalske, A. Biewener, Estimate of circulation and gait
change based on a three-dimensional kinematic analysis of flight in cockatiels (Nymphius hollandicus) and ringed turtle-doves (Streptopelia risoria), J. Exp. Biol. 205 (2002) 13891409.
[3] C.J. Pennycuick, A wind-tunnel study of gliding flight in the pigeon
Columbia livia, J. Exp. Biol. 49 (1968) 509526.
[4] C.J. Pennycuick, Predicting wingbeat frequency and wavelength of birds,
J. Exp. Biol. 165 (1990) 171185.
[5] J.M.V. Rayner, Bounding and undulating flight in birds, J. Theoret.
Biol. 117 (1985) 4777.
[6] J.M.V. Rayner, Aerodynamic corrections for the flight of birds and bats in
wind tunnels, J. Zool. Lond. 234 (1994) 537563.
[7] J.M.V. Rayner, Mathematical modeling of the avian flight power curve,
Math. Meth. Appl. Sci. 24 (2001).
[8] W. Shyy, M. Berg, D. Ljungqvist, Flapping and flexible wings for biological and micro air vehicles, Progr. Aerospace Sci. 35 (1999) 455505.
[9] G.R. Spedding, J.M.V. Rayner, C.J. Pennycuick, Momentum and energy
in the wake of a pigeon (Columbia) in slow flight, J. Exp. Biol. 111 (1984)
81102.
[10] G.R. Spedding, The wake of a kestrel (Falco tinnunculus) in flapping
flight, J. Exp. Biol. 127 (1987) 5978.
[11] V.A. Tucker, G. Heine, Aerodynamics of gliding flight in a Harris hawk,
Parabuteo unicinctus, J. Exp. Biol. 165 (1990) 469489.
[12] T. Weis-Fogh, Quick estimates of flight fitness in hovering animals, including novel mechanisms for lift production, J. Exp. Biol. 59 (1973) 169230.